Image Frame

Four Days of Facebook

This past September 12-15, HubSpot teamed up with Facebook to deliver a four-day series of live video content. During the week we covered all the latest trends in social media and how you can better use Facebook to grow your business.

Watch the videos here:

You can also read each video transcript by expanding the transcript section under each video!

Day 1 - How Social Has Changed and What Marketers Should Do

September 12, 11-11:45am ET

This panel of experts is here to help you understand the major trends in social media and how you can adapt your marketing strategy to take advantage of them. Speakers include Douglas Weiss, Product Growth Team Manager at Facebook, Daria Marmer, Product Manager, Social at HubSpot, Ryan Bonnici, Senior Director of Social/PR at HubSpot and moderator MK Getler, Learning and Development at HubSpot.




  • MK Getler:                              All right, crew. You know what to do. Let's get excited for the first day of The Four Days of Facebook. Excellent. I am so glad that you are as excited as I am to start this very first day of these four days of Facebook.

                                                          So, the interesting thing about this event is it is actually a global event. But, we're doing a global event while we're sitting here in Cambridge, which is really interesting. So, this digital event, we're trying to reach as far as we can.

                                                          So, those of you that are watching on Facebook right now. In the comments below, please toss in what area you're calling in from, what country you're calling in from, city, state, whatever it may be. Want to see how far we have reached all of you.

                                                          So, this event. We're gonna be streaming all week about social media, and how social media is impacting the small and medium business space.

                                                          We're gonna be starting to talk about some of the new tactics, some of the new trends in video, and of course this event is also named The Four Days of Facebook, so we're gonna be talking a lot about how Facebook is really transforming their platform to enable small and medium-sized business owners to transform their business. Things like video, messenger.

                                                          I'm also privileged and honored to be hosting the very first day with our three panelists. So, I'm very, very excited to be able to talk and pick their brains. Each one comes from a unique walk of life here on Facebook and in social media. So, I know you didn't come to this first 45-minutes worth of The Four Days of Facebook to talk to me at all.

                                                          You came to talk to these folks here to my left. So, I'm gonna turn things over to them, and let them introduce themselves.

    Daria Marmer:                      Hi guys. I'm Daria. I am the PM for all things social here at HubSpot. Previously, I worked as a PM at eBay, and also at LinkedIn. At some point, I even tried to reinvent cooking, and reinvent the cookbook on the iPad. And if I'm not conducting some sort of product orchestra, I'm probably in labor.

    Doug Weiss:                           I definitely can't say that.

    MK Getler:                              Actually, yeah. We technically have four panelists with us right now. I don't know if you can tell.

    Doug Weiss:                           Doug Weiss, I'm on the product partnerships team at Facebook. So, I actually manage a team that looks at all the different partnerships we can make to help build out the capabilities for some of our ad products that are designed for our non-digital-native clients. So, types of businesses that are in B2B, that are in automotive, that are in professional services, that are in retail. Really, companies that traditionally have not been as present on Facebook, but that we really think can leverage Facebook [00:05:00] in a lot of great ways. Especially with the partnerships that we're building.

                                                          Previous to Facebook, I spent a lot of my career either in management strategy consulting or working at startups. And if I'm not working, I do like to do a lot of travel, which is a lot easier when it's just me.

    Ryan Bonnici:                       Awesome. Hi, you guys. My name's Ryan Bonnici, I'm on the marketing team here at HubSpot, focusing on social, and PR, and campaigns. So, a lot of what I'm doing, and my team is doing is trying to work out how your wondrous algorithm works, which hopefully we get to have a little chat afterwards, I think.

                                                          But, in my previous life, I was head of marketing at SalesForce. And have worked in a lot of the marketing automation in tech space. So, love the space, love the social media platforms. And, yeah, I'm kind of excited to talk more about it all.

    MK Getler:                              I'm excited just to watch you guys geek out about one another. And totally just like nerd out about all things Facebook. So, let's just dive right in. So, Daria, we'll kick things off with you.

                                                          In the last few years, Facebook has really been like on a legitimate roll. Instagram has been like, like somewhat of a success. Kinda sorta. No, it's been a grand slam.

                                                          Messenger is now just like blowing up left, right, and center. So, like, what is your take on Facebook right now?

    Daria Marmer:                      I mean, obviously Facebook's blowing up. But, the thing that we don't talk about is that Facebook has been successful because they failed so frequently too.

                                                          So, like, how many of you guys remember that Facebook actually launched a Groupon competitor? Virtual currency? I think a phone in there at some point?

                                                          So, I think the whole point is that Facebook's done really, really well, because they've been able to iterate so quickly. So, being able to say, "Hey, let's go launch this feature. Let's go launch this product. Let's see how it goes. Let's measure it." I know that you guys are so analytics heavy on the way that you approach product. And if it's not working, pull the plug. And use those resources on something else.

                                                          And so, that allows you guys to iterate really, really quickly and do very well. But, frankly, if I knew the secret to their success, then I'd be on a beach somewhere. I mean, just saying.

    MK Getler:                              We'd all be basking in the Mediterranean sun right now, if we knew what the secrets were. Of course. Excellent.

                                                          So, if I'm understanding you correctly, really Facebook's success has just been like move on, next play, and let's double down on the things that are working really well for our business. Did I get that right, Doug?

    Doug Weiss:                           Yeah. I think that's exactly right. I think one of the things that I've been struck by since my, you know, when I first started at Facebook is this ability to want to test things out even though we're now touching almost two billion people a day, month rather. Where even though, you know, obviously we've had a lot of success, as you were saying that, you know, we've had just as many failures, and it's this ability to, again, test things, try things out and really respond to the data.

                                                          And one of the things I think Facebook does really fast is this whole idea of moving fast. And that's one of the really great things that we love about working with HubSpot, is that you guys kind of match our speed. A lot of the partnerships that we do, we want to try to build out new products, we want to be able to leverage the capabilities that other companies have had. But, because of the speed, or because of other priorities, they haven't been able to really work with us and try to figure out what actually is bringing the most value for our mutual clients.

                                                          And that's a great thing that, you know, why we are here today, and here for four days, rather, to actually really kind of talk about this great partnership that we've really kind of doubled down on in the last year, year and a half, because it is such a great, I think, merging of core capabilities. And you guys seem to have a very similar type of culture of wanting to try a lot of things out. And really not be too invested in one thing that you're not willing to pull the plug if it's not working, or double down if it is.

    MK Getler:                              Really, really ... You glazed over the stat, just like mind-blowing statistic. Can you say that again, a little bit more slowly.

    Doug Weiss:                           Yeah, so, I believe, and actually someone in PR, legal's gonna kill me for not knowing the specific number. But, you know, I think it's like 1.93 billion people are now on Facebook every month, which is, you know, like it's kind of crazy where you just kinda get a little bit numb to these numbers working at Facebook, which is probably like a very arrogant, and douche-y thing to say, but I promise I don't mean it in that fashion.

                                                          Where just because of the many different ways that clients and consumers are being able to leverage our platform, we're been able to touch a lot of different, you know, businesses as well as consumers, kind of all around the world.

                                                          And again, you know, you mentioned earlier when we were first starting that this is a global event. And that's another reason that we really like working with you guys is that you guys are also expanding in a very fast fashion. Where you're able to also think about how you can have the same core capabilities, but then also make sure that you're able to be using it not only in the U.S. but are in kind of markets all around the world.

    Ryan Bonnici:                       I'm curious around that billion dollar kind of ... Sorry, billion number. Like, obviously, Messenger's growing incredibly fast. And there's a bunch of different products within the core Facebook app. But, you know, speaking of like the global like, I guess, spectrum of this event ...

    Doug Weiss:                           Yeah.

    Ryan Bonnici:                       How do you find like the different countries and regions are using Facebook? Like, I'm from like Australia, so I know in Asia obviously messaging is huge, but curious like which kind of apps are growing the fastest in which regions if you have any insight there?

    Doug Weiss:                           Well, it's really interesting, because I think, as you were mentioning before, one of the things that we are also trying to do at Facebook is also learn from some of the different regions that are not where we are at headquarters. Where I think as you had said, messaging is something that we have seen also, obviously, blow up in APAC and Southeast Asia, especially, where you see this kind of movement from people who have been spending a lot of time on social networks actually moving over to Messenger-type products.

                                                          And so, you know, that's something that we expect to really actually ... And you already see evolving kind of in a lot of different other regions in the world. And again, I think that's one of the strengths of Facebook is that we are very aware of not only what's happening in the U.S but really responding to the things that are happening in other markets.

                                                          But, I guess to answer to your question, the interesting thing is like yes there are definitely differences, but you know, this idea of wanting to connect with other people, and to connect with businesses, is really kind of not that different all around the world. That's something that we see increasingly. People are changing how they interact with businesses and how they interact with other people all around the world.

                                                          And again, this is why we love building out these partnerships, so that not only are we trying to do it ourselves, but we're actually trying to learn from the things that our partners are bringing to the table as well.

    Ryan Bonnici:                       Got it.

    MK Getler:                              So, I'm super curious. There has to be some sort of intersectionality between the fact that people are using Messenger and that basically 80 percent of all social media usage is done via mobile these days. I'm sure we haven't bored some of you to tears, some of you are still watching and not on your mobile phones.

                                                          Yeah, sort of. So, Daria, so what does this mobile future mean for social media, and for how marketers should react to that?

    Daria Marmer:                      Yeah, so if you think about mobile you have, like, this much screen real estate. As opposed to like this much screen real estate. So, people are inherently impatient about getting the information that they want super, super fast on mobile.

                                                          And so, like, over the past five years, we see some things like responsive web design come up, we see websites that are just really, really, really, really long. And if you've ever thought about that, that's because people just want to scroll [00:12:30] on their mobile devices. And they don't want to really click around, and certainly what they don't want to do is fat-finger anything.

                                                          And I think Messenger is really the next step to that. So that, you know, even if you think about from a marketing-use case, forms that you'd have to fill out that have like a bunch of different fields, and you have to fat-finger your email address. And is it Gmail or is it that you're really trying to type in?

                                                          Well, really what you want is a way to know that this person is interested in your content. That this is a way to reach them. And have an open conversation, and that's what Facebook Messenger really allows you to do. So, what if you just replaced that form that you have on your mobile web and actually just start a conversation with them?

                                                          And that's really what Facebook is really enabling small and medium ... Well, you know, any marketer really to be doing right now. So that's pretty exciting.

    Doug Weiss:                           Yeah, and one thing I'll just add on that is that in addition to Messenger, you know, one of the things that we've been launching actually this week, and so super exciting for me, is kind of an expansion of our, what we call our lead ads product.

    Daria Marmer:                      Mm-hmm (affirmative).

    Doug Weiss:                           Where instead again of I know similarly have very fat fingers that totally typo every form I fill out manually. We have created a product here at Facebook that allows advertisers to be able to leverage the information that they've already shared, that a consumer has already shared on their Facebook profile.

                                                          If you think about the vast majority of the forms that you're filling out on a website, or on a mobile device, they're largely fields of information that you've already shared with Facebook to share with your friends and family. Name, birthday, email address, phone number. These are things that instead of having to type in tens if not hundreds of times, you can actually just automatically, with our new leads ad product, be able to just confirm the information is correct and submit.

                                                          So, instead of having to go through and a, manually type in the information, where we've done research where the average consumer actually takes 40 percent longer to fill out the same form on a mobile device as a desktop. And as we see increasingly a movement to mobile, this ability to create a streamlined, facilitated interaction between a business and a consumer is only more important.

                                                          And so, as a consumer, it's a much better experience, and as an advertiser, you're ensuring you're getting kind of the full scope of people who are interested in your business. You don't have those people who are dropping off because they're frustrated with filling out a form. You don't have those people who potentially are typing in the wrong information, and so you're not able to reach out to them.

                                                          And so, this is why we are really excited, especially with our work with HubSpot because you guys have such a great footprint of businesses that really care about connecting with new customers. That really care about lead generation. And this new product that we've launched this week really makes it easy for a HubSpot user to, within their HubSpot platform, launch a lead ad campaign, and start being able to connect with new customers, and drive new business to their business.

    Daria Marmer:                      And I just gotta jump in.

    Doug Weiss:                           Yeah.

    Daria Marmer:                      Because this just launched yesterday, today.

    Doug Weiss:                           Yeah. I'm super excited.

    Daria Marmer:                      And if any of you want to try it out, you don't actually have to be a paying HubSpot customer, you can just sign up for HubSpot Marketing Free, and it's available for anybody to check out and try. So, give it a shot.

    Ryan Bonnici:                       What I think is really cool about like the new product. But also, just like Messenger, and like, the concept of like native content within the platform.

    Doug Weiss:                           Mm-hmm (affirmative).

    Ryan Bonnici:                       Is that I think it just highlights for marketers, and for sales people, and just for businesses that want to grow. Like, how important like the milliseconds are with like a consumer on mobile.

    Doug Weiss:                           Yeah.

    Ryan Bonnici:                       We just find like the moment you take someone out of platform, whether it's to a website, or to a video hosted on another social-media platform, your drop-off rate like is over like 50 percent, typically.

                                                          I mean, how is like Facebook thinking about, I guess like, pulling more native experiences in content within their products?

    Doug Weiss:                           Yeah. No. I think that's another great point of like, you know, in addition to the fact that you have some of the fields prefilled, you also don't have that lag. Where we similarly find that if you have an ad where you click on it, and then you have to load up a mobile site, you have a majority of your clients drop, or sorry, your consumers drop off.

                                                          And this is like, I think, you know, you can see this across the board is what we are increasingly trying to do on Facebook. Not only with our ads product, but in kind of all aspects of our products, is to make it an enjoyable experience for the consumer. Where they're able to access the engaging content. They're able to connect with the right businesses. They're able to connect with their friends and family in a way that is really easy. Where, you know, we recently launched what we called our Watch tab within the app, where we're hopefully enabling this community of creators, this community of content creators that already exist to make it easy for them to bring their content to Facebook.

                                                          One of the things we're super excited about is that, like, we already, you know, the pivot that we've seen to mobile, which has now happened over the last three years. And I think people will say is now kind of happened. We are now seeing a similar pivot with video. Where, you know, two, three years ago, you go through your feed, and the vast majority of it was static image. Now when you go through your feed, and obviously, especially when you go to our Watch tab, it is increasingly video, because as you see, kind of devices improve, as you see connectivity improve, you're now able to bring that great kind of live-action content to the consumer.

                                                          And so, we're trying to figure out how do we make it easy for those content creators to be able to bring their content to our platform, because we are very aware that we're not going to be able to do it ourselves.

                                                          So, we want to make sure that we are creating the right tools, both ourselves and again then with our partners, to make it easy as possible for that content to be brought to our platform. So, again, that consumers are having kind of the delightful experience that we want them to have on it.

    Daria Marmer:                      I mean, Ryan, like, what percentage of HubSpot's content right now on Facebook is video?

    Ryan Bonnici:                       I mean, it's gotta be, like, 95 percent of what we're producing.

    Daria Marmer:                      Yeah.

    Ryan Bonnici:                       I think like, it's obviously, you know, creating the most value for users, because like that's what generates the most reach, and most views, and most clicks, which is great. But I think, like, the creation of video is still something that most marketing teams just like aren't staffed for.

    Doug Weiss:                           Yep.

    Ryan Bonnici:                       And it's definitely like a new skill that we're seeing, which we could go into like so much detail on. One thing I'd like love to get your thoughts on, which I'm really like fascinated by is obviously Instagram Stories has taken off, and it's massive now. Right? Like, most people using that instead of Snapchat, especially in like the business space.

                                                          But, you know, that's been a feature on Facebook, and the Facebook core outfit for a while now, but no one uses it. Like, all of my followers on Instagram, I'm sure are like on Facebook. Like, why don't they use it there? Like, like what are the insights there for how people use platforms?

    Doug Weiss:                           So, I mean, I ...

    Ryan Bonnici:                       Doesn't make sense to me.

    Doug Weiss:                           It is interesting, because I think it just reflects how people, you know, people are often both users of Facebook and Instagram. And so, people go to different platforms for different things. And you know, anecdotally I've had the very same experience where like my Insta Story feed just blows up constantly, where I don't have enough time to go through everything. Well, on Facebook it's just not happened yet.

                                                          And I think that goes back to, you know, what Daria was saying earlier, in that we are honestly testing things out. And you guys are kinda seeing it live, as it happens, where we want ... Story is obviously very attractive and has gotten ... Had a lot of success on Instagram. So we decided, you know, maybe we should try it in Facebook as well.

                                                          You know, we're still figuring out if that is the right thing to be doing. And what can we do to potentially tweak it a little bit more than it is today? So that it does have a little bit more activity than you're seeing today. But I think, you know, one of the interesting things, especially for a company that has such a large footprint, a lot of the testing that we do, you guys see, because it's in the app.

                                                          You know, we obviously do some smaller scale testing before we launch it more broadly. But, I think one of the things that is really unique about Facebook, at least in the large companies that I've worked in, is this willingness to try things out before knowing 100 percent that it's gonna succeed.

                                                          So, you know, that was a very long answer to I think we're trying to figure it out as we speak.

    MK Getler:                              So, Ryan, you mentioned that 90 percent of HubSpot's content on Facebook is now turned into video.

    Ryan Bonnici:                       Yeah ...

    MK Getler:                              So, maybe we're not quite in the shift of like having our live feed happen yet on Facebook. But, like for someone who hasn't even started publishing video on Facebook for their business, what advice, what three tips can you give them to get started with video?

    Ryan Bonnici:                       Yeah. I mean, okay, I think the way I'd simplify it is that like video for mobile, and Facebook, and you would probably know this in more detail than me. But like, it's almost like the opposite of what you learn at video school. Right?

                                                          So, you learn to create stories that like build suspense, and like a story arc, and like there's this reveal at the end. Well, if you look at like anyone that's doing Facebook videos really, really well, like the first second, millisecond is almost like the end of the video. And then they like, like they capture your attention with like the cliffhanger at the end, which might be like a beautiful tasty cake or something, which I love.

                                                          And then they'll show you it being made. But, if they didn't show you that quickly at the beginning, you don't really know what's being made, so you would drop off.

                                                          So, I'd say, you know starting out with like the most important, the most sharable, the most controversial thing is probably one of the first things I'd recommend, which is I think where live video can be difficult, because that's typically not always able to be done.

                                                          So, that's important. I mean, playing with, I think, dimensions of screen is really important too. So, you know, kinda Snapchat introduced this whole like 9 x 16 format. And if you create a video, like, with your phone in a vertical format, if you were scrolling through Facebook, like, it takes up the entire real estate of someone's Facebook screen. So, you're not competing with, like, an ad above you, or just an organic wall post below you. Like, you've got their undivided attention. Now the content needs to capture their attention obviously.

                                                          And then I think like the third one, I'd say, is just like it doesn't need to be super perfectly produced, the content. Almost like we see, kind of like an inverse relationship with production and views and engagement, the more real, and natural the content is, people seem to respond really well to that, because it doesn't feel like it's branded content. It feels like it could be a friend.

    MK Getler:                              That seems kind of counter intuitive. So, you mean that the bigger the budget, the more you spend on making the video, the less impact it's gonna have?

    Ryan Bonnici:                       I mean, not necessarily. I think if you are spending that kind of budget, or putting so much production work in, you [need to think a lot, I think, about like how you make sure that it's really engaging. So, some brands that do it really well are like, I don't know if anyone's seen like Purple, who like sell those beds on Facebook. I see, like, retargeting for that all the time. But, their video, it's obviously really high production, but there's like very quirky elements in the video. It's set in kind of like ... I think it's like, not Snow White, but like there's like weird thematic elements of the video, or the whole Dollar Shave Club video as well. Right?

                                                          Like, that video looks like not super high-end, but it's like one shot. It's this guy moving through a warehouse. It's like lots of things happening. So, it's like being timed to like every second, it's delivering another punch to keep you engaged. So, I think it's just harder to do that with like highly produced content.

    MK Getler:                              Gotcha. So, basically you don't need to have a big budget to make great content.

    Ryan Bonnici:                       No, I don't think so. I mean, there's like lots of platforms out there now that we partner with a bunch of them like Animoto, Shakr that help you kind of create, like, quick videos that are templated. Adobe Spark even, there's a lot of great tools out there now.

                                                          And that's good, I think, for like 40 percent of your content, your daily stuff. And then you might want to do something a little bit more, like, Adobe After Effects for like the higher end sort of video content, perhaps.

    Daria Marmer:                      The important thing is to be human. Right?

    Ryan Bonnici:                       Yeah, the important thing is to be human. And just to like to have a good story. Right? Like it comes back to the tenants of inbound marketing. Like, know what your audience cares about, and like, tell them a really great, captivating story. Just, like, kinda give them the hook at the very beginning.

    MK Getler:                              Yeah, I think that that's a good point too. So, when thinking about human, and also kinda thinking about my personal interaction with the Messenger platform, that kinda feels a little personal to me to be on that platform. So, thinking more about how a business can insert themselves into this conversations that were originally intended for like one-to-one communication.

                                                          Daria, what insight do you have there for businesses?

    Daria Marmer:                      I mean, as human as you can be. You see things like chat bots right now insert a few seconds of typing between that. And obviously, it's a chat bot, it could, you know, it could respond immediately, like, there's really no need to have ...

    Ryan Bonnici:                       Your attention. Yeah.

    Daria Marmer:                      A few seconds.

    Ryan Bonnici:                       It's suspense. Yeah.

    Daria Marmer:                      But, it really is a second of like, "Oh, it feels like I'm having a conversation with somebody." As opposed to just having a, like, I don't know, like, a very bad conversation, where there's no person on the other end. And really thinking through all of the usage flows. I mean, I think one of the cool things that our marketing team did for this event ...

                                                          So, a lot of you guys registered through Facebook Messenger. Is that, like, there were gifs in the Messenger flow, which was so cool. And that's just one of those small things that it just made it so human to see, like, yeah, yes we're doing this in an automated fashion. Yes, it has to scale to thousands of people. But, no, we don't have to lose our sense of, like, being people doing it. Having a sense of humor while we're going through that entire process.

    Ryan Bonnici:                       And I think it's like being authentic to the platform you're in. Right?

    Daria Marmer:                      Yeah. Yeah.

    Ryan Bonnici:                       Like, Messenger is very much a one-to-one personal channel. It's going to evolve over time, absolutely. But, it's very different to email.

    Doug Weiss:                           Mm-hmm (affirmative).

    Ryan Bonnici:                       Like, at the moment, anyways.

    Doug Weiss:                           And I think just kinda reemphasizing the point of thinking about all of the different use cases and flows, because I think there are, again, I hope like PR's not gonna kill me for this, there are scenarios that like actually Messenger doesn't make sense.

                                                          Like, where it's actually a very important situation. Where there's like a lot of sensitivity. Where you need to know how to compliment Messenger with the other channels that you've been using from the day-one of your business. And think about like how it actually can work very well with a phone call. With an email. With you know, any of the other channels that a business may-

    Daria Marmer:                      In-person conversation.

    Doug Weiss:                           In-person, yeah. And think about how it can, yes, bring definite efficiencies. And definite increased reach. And definite more increase effectiveness for some scenarios, but maybe, at some point, knowing when to hand it off to something else. And really think of it as a full-portfolio communication strategy, as opposed to siloed different ways of connecting with a consumer.

    MK Getler:                              Fascinating. Okay, so, maybe I'm one of the only ones in the room right here who's like, "Uh, bots?" Okay, so what is the deal with these bots? We've talked about how we've leveraged bots to actually help with the registrants for this event. But like, high level, like, give me like the two second download of what a bot is.

    Doug Weiss:                           Is that to me?

    MK Getler:                              I'm just gonna toss that grenade to the masses and see who jumps on it.

    Doug Weiss:                           Yeah. I mean, I think that, you know, a bot, like, in its simplest form is like an automated way to respond to some type of thread on Messenger, where you have kind of ... Depending on what the user has said, and depending on what information you know about them, you can like, you know, efficiently and effectively, without a human actually typing in a response, give them a large majority of their information that they might be interested in.

                                                          You know, this is really helpful in scenarios where you have basically the same questions happening time and again, the same types of threads, the same type, again, patterns of conversation that really don't need the flexibility and the human factor, quite bluntly, of a human.

                                                          But rather, can actually predict what is the right answer, what is the thing that the user actually wants to understand. I think the one thing that we have seen is that there is obviously a very wide range of quality of bots on our platform. Of like really investing in making sure that it is a good experience for the user.

                                                          I think, you know, it's definitely important to bring efficiency to your platform, and to how you do, you know, lead generation, or you know, customer care. But, at the same time, you want to make sure that you're actually still bringing a good experience for the user.

                                                          If you, you know, use a bot and the user gets frustrated because it doesn't have the right information, and they're just kind of in this loop. I mean, you know, not so much on Messenger, but I can think about my own experiences calling a call center, where you know, the last time I, like, cannot think about a good experience I've had calling an airlines where like I actually just want to talk to someone to understand how delayed my flight is, and I'm just in this endless loop of like, you know, computer-generated talk tracks that actually isn't giving me the information that I need.

                                                          So, again, I think, you know, this idea of bringing efficiency is definitely important for businesses, especially as we see people moving to Messenger, but you want to make sure that you are, again, creating that great user experience, so that they don't get so frustrated that they may actually defect from your business in the end.

    Daria Marmer:                      I mean, and one of the cool things about bots is that you can incorporate a lot of information into how smart the bot is.

    Doug Weiss:                           Mm-hmm (affirmative).

    Daria Marmer:                      So, like, one of the things for marketers that I'm personally very excited about is the integration with CRM. And so, being able to say, like, when you start a conversation with the bot, to know that this is somebody who I have in my list of prospects, this is somebody who is brand new to my site. This is somebody who is a long-standing customer. Is all going to change the way that I'm going to interact with them on the site. And that's what's going to make that bot experience not feel like the interaction that you have with an airline.

    Doug Weiss:                           Yes.

    Daria Marmer:                      You know, if you think about some of the things like, "Oh, we know that you're ..." What an airline should be doing, like, "We know that your flight is [00:30:30] in ten minutes, and hey, we checked your GPS, say, we know that you're more than ten minutes away from the, uh, from the airport at this time. Like, would you like to rebook?" And make that super, super easy for the user. That's like the amazing use case, not putting people through the same rigamarole.

    MK Getler:                              So, Ryan, where is HubSpot evolving their usage of bots beyond just the registrants for this event?

    Ryan Bonnici:                       Yeah, I mean, the marketing team is doing a bunch of different things in the bot space. I think, like, the syncing to CRM space is super interesting, because Facebook obviously this is like such a wealth of data for marketers and businesses.

                                                          I mean, there's so many use cases I think, whether it's like creating a one-to-one connection for a sales rep, automating sales rep messages, like you would in an automated sales rep email thread.

                                                          I think it's like there's so many use cases. I think with like where you just need to be like careful though is like not over-complicating a bot and trying to create like a streamlined like process for where things can go. There's like a bot I was trialing just the other day, and it's basically like a news bot that looks at the ... Like, whether or not a news article is like, like lean to one political angle versus the other.

                                                          But, it's kind of like that's something I'd prefer to just go to a website and like throw in a news article, and see. Like, the UI is something that I'd like to see visual, like graphs of, so it doesn't really work super well. But then, for conversational stuff like registering for an event, getting like NPS survey back data from someone. Like, trying to book a meeting. Like, that quick sort of fast stuff, I think is great.

                                                          The best use case I've seen so far though, and it's not in a Facebook bot, but it was in a native-app bot was Equinox, the American gym. Like, when you sign up with them, they'll put you through this like bot that like asks you like about like how fit you want to get, like the days of the week that you work out best on. And it even asks you what I thought was amazing, like a motivational message to show up on your screen when you're snoozing your alarm in the morning.

                                                          So, like, it like flashes this thing that you wrote that's really like, you know, dreamy kind of about like what you want to achieve with yourself. And like, there's no way to look at that and like to snooze, because like the guilt just kills you.

                                                          So, like, that was like an interesting thing, whereas I would've never filled that out on a website if I had signed up for a gym and they were like, "Fill out this long survey." But the fact that I kind of didn't know where this survey was going kind of worked well for that. So, there's so many uses.

    MK Getler:                              I think I need a bot to counteract that bot if I snooze my alarm, then I ignore that first bot, then I need another bot [00:33:00] to say like, "No really, get up. Get out of bed."

                                                          So, for everybody who is watching remote, please leave your questions for our panelists in the comments. I actually have my phone, I'm not just looking pictures of my cute dog, I'm actually looking, and crawling through, to see if I can pull some questions from you in the audience.

                                                          So, another question that I have for your panelists is like ... What's next for Facebook and for HubSpot together, collaboratively? I'll kick that off to Daria, and then maybe we'll just go down the line, and have everybody take a stab at that question.

    Daria Marmer:                      Sure. So, what's next is there's a huge event coming up in Boston sometime soon, it's called Inbound. Just kinda small. You know, only like over 20,000 people gonna show up.

    MK Getler:                              Oh that's it. No big deal.

    Daria Marmer:                      You know, yeah. So, some of the big announcements are gonna be then, so I don't want to totally take away the thunder of that. But, clearly, a lot of the things that we've been talking about. We've launched recently Facebook video, we've done emojis in the social platform recently. Tagging people on Facebook. Something that we haven't spoken on this panel, but is hugely important for anything co-marketing related. These are all things that have come out in the last few months on HubSpot.

                                                          At Inbound, there might be some Insta some announcement. But, I'm not gonna say anything.

    MK Getler:                              Was that a teaser? You just drop a teaser on us? Yeah?

    Daria Marmer:                      Just ... Yeah.

                                                          And obviously, lead ads launched today, and there are gonna be some interesting ads announcements as well coming up at Inbound. So, yeah.

    Doug Weiss:                           Yeah, I mean, I know I've mentioned this already, but I think just because for us, it is such an exciting partnership that launched this week. Is this, you know, the lead ads launch that we had. For us, you know, one of the biggest things that we hear from our clients time and again, is not only the desire to connect with new customers, the desire to connect with existing customers, and get additional data on them, which is what our lead ads product so well.

                                                          You know, by creating a fully native ability to get first-party information on users, you're really able to not only get leads at a lower cost per lead, but really getting the high-quality information at scale that almost every business is looking for.

                                                          But the thing is, is like the lead is only the first step. And so, we're really excited with our partnership with HubSpot is that, that you know, by launching your ads from HubSpot, you're automatically ensuring that those leads are going directly into your HubSpot CRM. So you can do all the, you know, appropriate followup. So that not only are you generating leads, but you're actually generating leads that will convert, which is the thing that actually matters.

                                                          And so, you know, especially, you know, in conjunction with some of the work that we've already done that we launched, which allows you to really understand the ROI of these campaigns, and understand what is the real impact to the business, which is what we are very focused on at Facebook.

                                                          You know, if you look back five years, you would run a campaign on Facebook and you'd get like data on impressions and clicks, and you know, in the end, like, a business doesn't usually care about that. What you care about is how it's gonna impact your business. And we think, you know, lead ads is one step in that direction in that it helps you generate leads. But now with this new partnership that we've developed in HubSpot, we've made it really easy for businesses of all kinds, in a really lightweight fashion, not only to generate leads but to generate leads that are really gonna convert.

    Ryan Bonnici:                       Totally. I mean, we're using all of this stuff ourselves now as well. So, I think for anyone watching that's like wanting really tactical [00:36:30] and practical tips on how to do this, I'd say like follow and like our Facebook page, because we're gonna be sharing all of the learnings that we're having from the platform, and the optimization that we're seeing.

                                                          We're also gonna start rolling out all of the videos that we're doing as templates. So, people can download them via our Facebook page, and just use our templates for their videos.

                                                          So, similar to all of the great e-book template content that we've got, like buyer persona templates, like budget templates, we'll have that all for our Facebook videos, and like lead gen ad, kind of like, best practice templates. So, hopefully people won't [00:37:00] have to worry too much about like creating their own native video. We're gonna try and help out a bit there.

    MK Getler:                              Fantastic. All right. So, some people have been leaving comments, and questions for our panelists. So, Doug, this one is actually for you. How would someone use this in a B2B space where personal business and emails are different? I believe that this in this case is Messenger, and potentially bots. So blending the two email addresses from work and personal.

    Doug Weiss:                           Oh, okay. Yeah, I mean, I think, like it's no secret that Facebook is largely a social platform. You know, what we have seen is that we have been able to get, you know, with the lead ad products, you know, the email is gonna pull whatever email that they've shared, kind of on their Facebook profile. But, you do have the flexibility to really ask very specifically for their work email.

                                                          And so, you can make sure that then you're getting that right information, and then by also additionally having the personal email, you can then also, you know, capture all the other information that might be relevant [00:38:00] from their Facebook profile.

                                                          You then can easily match that up to potentially a profile that they already have in the HubSpot CRM, so that you're able, again, to leverage these two sources of great information, to create a much more fleshed out profile of the user, or of the customer.

                                                          In our mind, like, that's one of the things that, again, is exciting about our partnership with HubSpot, is that we ... Our strengths are in slightly different areas, where, you know, a mutual customer can really be able to leverage the benefits and the strengths of each of them to create the best understanding of a user, and the best understanding of a customer. And then, you know, leverage either Messenger or any of the other channels that they might ... That a business might use to really, again, cultivate that relationship and hopefully turn them into a more loyal customer.

    MK Getler:                              Fantastic. And I'll toss this one over to you, Ryan. So, as a startup, what's the best way to market a product? And maybe what tools would HubSpot and Facebook recommend for that startup?

    Ryan Bonnici:                       Good question. God, that's a hard one. I mean, I think it really depends on like what it is that you like sell, to be honest. Like, if you're selling a service, a product, an enterprise, or like, a small, mid-market product, it's super different.

                                                          I think if you're like in like CPG, so like, consumer-packaged goods, or fast-moving consumer goods, like, Instagram can be a great like channel to start to sell that product as well as Facebook. I think like there's obviously just so many variations, but without getting into the specifics, I just would say like really understanding your consumer, and realizing that like the intent on social is really different to the intent on your blog. Right?

                                                          Like, when someone goes to search on Google, like, how to do X, and then you've written a blog post about that to then move them through your funnel, there's like that direct response intent. Like, they're thinking and doing. Whereas like social is really about seeing and thinking. There's like less of an intent kind of like function there, I find.

                                                          It's not to say you can't like drive that through ads, and still get like really cost effective returns. [00:40:00] But, from the like native and organic Facebook in general, I'd say is like you're trying to go higher in the funnel to create content that like that person wants to engage with, even maybe when they're not in the like active-buying-of-your-product mode.

                                                          So, to put that into context. Right? Like, we obviously sell marketing-sales software, and we have a lot of free different marketing-sales software as well. But, like, our content that does the best on Facebook that still drives the leads and portals for us at the bottom of the funnel, is that content that's higher up. Like, you know, how to get through [00:40:30] like the workweek from hell, because like, our buyer persona, like, marketer sales people, like they work in an office, like they all sometimes have bad workweeks.

                                                          So, it's really I think just like matching what the person is happening within their life, and giving them the right content at the right time.

    Doug Weiss:                           If I could just add on to that. I think two additional things I would add. I think, one is also leveraging the information that you have, which I think goes back to the point that was being made of like, understanding what it is that you're trying to achieve with Facebook, and understanding who your customer is, and understanding what your product is. And then leveraging information that you already have on them. So, if you already have customers, use that to find more customers like that on Facebook. So, we have a great product called Lookalike Audiences where you can upload, and you know, HubSpot has already built this into the platform where from HubSpot you could automatically upload your current customers.

                                                          And since, again, we have as we were saying before, such a massive footprint in the billions of people, we can figure out on our backend what are [00:41:30] other people that look just like your most valuable customers, and make sure that your ads are reaching them, who are, again, those people who are most likely to convert.

                                                          And so, leveraging that information that you already have to be able to inform your Facebook campaigns will make it much more likely that you are connecting with the right people, and much more likely that you're actually then converting them into customers.

                                                          The second thing I will say, and you know, I think this is reflective in kind of how both companies work is also to test. Often, people will think one thing, and then when you actually launch campaigns, what you thought was actually the opposite of what actually drives the results that you want.

                                                          And so, making sure that you're testing everything. You know, testing different forms of creative, testing different ad formats, testing the different platforms, and figuring out what is the thing that really resonates, because it really is very different for every business. And so, making sure that you have the results and the formats that work for you is really the most important thing. And the only way you can do that is by testing it yourself.

    Daria Marmer:                      And I'm gonna jump on with one more thing, which is Facebook has a lot of reviews now on businesses. And if you're going to have a business page on Facebook, go to your top promoters, go to the people who are already delighted by your product, and have them write a good review, because for the people who are there who are like, "Oh, well, let me just check out this company, like, see if it's any good." That's gonna make a really big difference if they see somebody, especially if they see somebody who they know ...

    Doug Weiss:                           Mm-hmm (affirmative).

    Daria Marmer:                      Who's vouched for that product. And that can really help.

    Ryan Bonnici:                       The other thing is, okay, if you are a startup company that's like still trying to work out like product and market fit, and understanding and evolving your products. One tool that I've like seen a lot of like startups use, and it's I forget the name of it, but it's, I think like, the ad format that lets you kind of create a canvas within Facebook. What's that called?

    Doug Weiss:                           It's called Canvas.

    Ryan Bonnici:                       Oh yeah, sorry. I got it. I know it. So, like the Facebook Canvas ad sets, and so what some startups will be doing is like they'll test different messaging. So there are different taglines in these like, almost like Facebook micro-site ads, [00:43:30] and they'll like look at them, like, which, like micro-site tagline or colors convert into the most customers. And they'll use that data to then inform their designer to build their websites, build their tagline.

                                                          So, it's kind of nice, like, taking the marketing space out of subjectivity, and into a little bit more objectivity. So, yeah.

    MK Getler:                              You just gave some small startup some massive, massive growth hacks right there, guys. I want to thank each one of you for this incredibly, incredibly insightful conversation. You guys have educated me to no end, I just wish I had an idea with which to incubate all of these ideas.

                                                          Excellent. So, for those of you that have dialed in, thank you so much for joining us today. I just want to give you a heads up that this is day one of day four of our Four Days of Facebook. So, tomorrow, this is the day that I have really, I have come to find that many people are long awaiting. The Brian Halligan and Gary Vee conversation.

                                                          [00:44:30] So, dial in at 2 P.M. to hear Brian and Gary Vee hash it out. Again, thanks so, so much to my panelists today. Continue to leave your comments. If you have questions, we're going to be online and answering those questions even after we wrap today.

                                                          Excellent. Thank you guys so much today.

Meet the Speakers

  • Douglas Weiss

    Douglas Weiss

    Product Growth Manager, Facebook
  • Daria Marner

    Daria Marner

    Product Manager Social, HubSpot
  • Ryan Bonnici

    Ryan Bonnici

    Senior Director of Social/PR, HubSpot
  • MK Getler

    MK Getler

    Learning & Development @ HubSpot & Four Days of Facebook MC

Day 2 (Part 1) - Growing a Business in the New Age of Social

September 13, 2-2:30pm ET & 3-3:45pm ET

Part One: Both Brian Halligan and Gary Vaynerchuk have built and scaled businesses. If they were starting a business today, what role would social media marketing play? What networks would they focus on? What role would video and ads play? How would they address mobile consumers? How would they leverage AI? Hear their answers and more in this live session. 




Day 2 (Part 2) - The Impact of Social Influence

Part Two: The Impact of Social Influence in 2017. MIT’s Sinan Aral is a leading researcher and lecturer on social media and influence. He studies and talks about influence and the impact that technology has had on it. Sinan will be giving a TED-talk style session as part of the second day of Four Days of Facebook.




  • Brian Halligan:                     Okay. Welcome, everyone, to my interview today with Gary Vaynerchuk. Big Gary V. fan. Gary's been on our advisory board for a long time. We're honored to have him there. Been a keynote speaker twice at Inbound, written for our blog, he's CEO of Vayner Media, a YouTube star, social media icon. Every time I chat with Gary, I learn new stuff, so super excited to have him on.

                                                          Gary, welcome.

    Gary Vaynerchuk:               right now, and I'm ready to rock and roll. Let's give them some content.

    Brian Halligan:                     Hey, most importantly, Pats are playing the Jets on October 15th. How are we feeling about that outcome?

    Gary Vaynerchuk:               I'm feeling tremendous because I'm going for Owen 16. That is for the New York Jets this year. So I feel like that game will follow the pattern of last year, where the Patrons blew the Jets out after years of tough games. So I think we will be in a classic 41 to six Patriots victory, which under normal circumstances is devastating but very on-strategy for me this year.

    Brian Halligan:                     Our incentives are aligned.

    Gary Vaynerchuk:               Aligned.

    Brian Halligan:                     Gary, let's talk about Facebook. What do you think Zuckerberg and those folks, what have they done so well in the last, call it, five years? They're on fire.

    Gary Vaynerchuk:               Mark, Cheryl, the rest of that team, the product people, Carolyn, all of them, they have continuously put themselves out of business. What I mean by that, Brian, is that they have continued to iterate. They adjusted to mobile. They adjusted to video. They listen to the market, and they look at big data and make decisions on the product. They make some intuitive decisions. They make decisions that are unpopular.

                                                          If you take it all the way to Instagram, through Kevin Systrom and his team, they don't care what you and I think about them copying features from Snapchat the same way that, in the consumer packaged goods space, if somebody figures out that people like mint chocolate chip flavor, every competitor starts making mint chocolate. Features are not patents or patentable, or normally aren't.

                                                          So what they have done is everything I believe in, which is move quickly, especially considering their size, never rested on their laurels. What got them there is the exact thing that will make them lose in the future.

                                                          If you think the amount of things Facebook did from 2011 to 2015, through all this, in social media , think about what Facebook did as a business from 2011 to '15 on the product and what Twitter did from 2011 to 2015. Twitter did zero. Zero. Same fucking product 2011 that it was in '15. That led  to its vulnerabilities.

                                                          Facebook changed time and time again. Remember when everybody went up in arms that they created the Newsfeed? 'Cause you used to go to people's walls. I mean, the Newsfeed is their business. It's a trillion dollar business. People started pages to, like, one million people ...

                                                          They're run properly, which is you listen to your users, you adjust to that feedback, but you have vision to where you're going. They're about to do the same thing with original programming. It's going to work. There's too many eyeballs. There's too many technology shifts in the future.

                                                          I really think you should put yourself out of business before somebody else does it. That's what they do. They are not gonna allow somebody to come along and do to MySpace what they did to them. They're not gonna let somebody else do that to them.

    Brian Halligan:                     Sure. You brought up a couple other socia networks. So let's just say you're a 20-person company, you're running marketing for that company, you're strapped for time. You've got Facebook, you've got Instagram, you've got Snapchat, you've got Twitter, you've got LinkedIn. You're in a B2B market. You're trying to grow. Where do you spend your time?

    Gary Vaynerchuk:               B2B?

    Brian Halligan:                     Yeah.

    Gary Vaynerchuk:               You spend your time ... me, if I'm doing that, if I'm selling a SAS product, you always want to be a brand 'cause then people come to you.

                                                          One thing I admire about you guys, if I could take second to give you a commercial, you and your partners, I thought, from afar, when we didn't know each other, when we met at a restaurant in Midtown, I was like, "Oh, these guys actually understand content creation. They're producing events and value to their users. There's a halo effect to that, which is gonna be impactful in their business."

                                                          I don't think there's anybody remotely close, in my competitive landscape, who was more obsessed with giving away free information than I was. There's no seminars. But I just wanted to bring value, and if I was gonna be right, and I brought the most information, it would work out, and it has. It will continue to.

                                                          So first, I want every B2B company to think of themselves as a media company  so that people come to them. I would probably, believe it or not, probably start a podcast, a weekly show. If I'm selling a SAS product in the ad tech space, I'm probably starting a show called This Week in Advertising, brought to you by Ten Tech.

                                                          My first guest is probably somebody who is in the TV advertising space and has nothing to do with programmatic right? I'm trying to bring value to the overall market, not my product. I'm not doing an infomercial. You guys have done that well. You don't do infomercials for HubSpot. You're in the game of it, and it trickles. So I think you're building a brand. I think that's what I'm building. That's what I think everybody else should be doing.

                                                          So first and foremost, whether it's audio, video, or written word, I would create a medium, a media company, a newspaper magazine, a radio station or radio show, a television show. Make a CNBC show. Make a Wall Street Journal competitor. Make a radio show that competes with Dave Ramsey in your genre.

    Brian Halligan:                     Sure.

    Gary Vaynerchuk:               Then, I would take the hour of that, let's say it's a podcast, and I would hire people to cut the audio and make videos of images behind the sound. I would hire a copywriter to transcribe three or four of the pieces that were most interesting in the show and create blog, medium, and LinkedIn posts. That's what I've done.

                                                          You've watched me for a long time, not that you're watching every day. My intuition is you're aware that probably over the last year or so, year and a half, it's accelerated dramatically, and it's because I realized if I make a ... I have two core shows: the Ask Gary V. Show and Daily V. One is a show Q&A. One's a blog, and  it's documenting. That became the funnel for all my content.

                                                          I'm producing more social content than ever because I'm DJing the assets of the monster content. If you're a small, 20-person company, that's more efficient. You're creating more, and more impact, bringing more value to the system, and selling more stuff, ultimately.

    Brian Halligan:                     Okay. So your advice is be a media company. Let's talk about media. You've got audio these days, video, text. What's your advice on the balance? It feels to me the world has shifted big time to video. Is text dead?

    Gary Vaynerchuk:               No. The written word is forever. Audio is forever, and video is forever. So no, it is not dead. It's just not maybe in its hyper moment the way it was in 2003, in the same way, as you know, this is podcast's second go around, and it did a hell of a lot better this time than it did in-

    Brian Halligan:                     Yep.

    Gary Vaynerchuk:               Right? Remember? Odeo. Wine Library TV was called a video podcast 'cause podcasting was getting going in 2004, '05, '06. So no, text is not ... as a matter of fact, I would argue that written word ... Again, I always say don't listen to me. Watch what I'm doing. Why am I writing full, 25-sentence posts on Facebook? 'Cause it's working.

                                                          Listen. From cavemen  to the robots that will ultimately kill us, the written word, sound, and video are the platforms. They go through ebbs and flows. But people should triple down on what they're good at.

    Brian Halligan:                     Talk about that. Talk about that. 'Cause some people think work on your weaknesses. It sounds like you don't think that. It's like lean into your strengths.

    Gary Vaynerchuk:               I do not. I think that-

    Brian Halligan:                     What if you don't have a strength in video creation? You're a 20-person company. You got one marketer, and, jeez, that marketer grew up in PR and writing ... What do you do?

    Gary Vaynerchuk:               She or he should write their asses off, and they should hire a post-production video editor to take those written words and make them into video content. I believe it's ... I'll use music. We are in the era of DJ. Right? I don't know your music tastes, Brian. I know a lot of other people's. People think EDM's music. People don't. People think hip hop is music. People don't.

                                                          Here's what I know. Creating original songs and being able to DJ them with different beats and different genres are both equally phenomenally important. You can be DJ Khaled. You can be Drake. They both matter. Figure it out. But better your strengths, because when somebody who has 22 years in copywriting PR decides that he is now some phenomenal Instagram meme-maker, they get exposed.

                                                          Then they start shitting on all these new things because they're fearful of it instead of surrounding themselves. Right?

    Brian Halligan:                     Yeah.

    Gary Vaynerchuk:               So, look. I can literally talk an entire book out in one breath.

    Brian Halligan:                     Yes, you can.

    Gary Vaynerchuk:               But I can't write two sentences together coherently. So in January, when Crushing It comes out, it will be my fifth New York Times bestselling book. I find that laughable for a man who can't write three sentences together. It's because I created infrastructure, and Stephanie Land, my ghostwriter, and I am betting on my strengths of video and talking, gift of gab, and  I'm producing ... I'm the number one followed person on medium.

                                                          I can't write. I have Colin Campbell, who's a beast in transcribing me. I'm filming this now. He may listen ... he may want to write ... he'll pitch me, "Hey, let's do betting on your strengths." He'll transcribe everything here, but it'll be missing some stuff. He'll call me and interview me, as if it's you or [Fast 00:11:23] company, for four more questions that give more meat. Then I speak in first person in a blog post.

    Brian Halligan:                     Got it. Lean into your strengths. Got it. You are a master of-

    Gary Vaynerchuk:               Hold on real quick. I apologize.

    Brian Halligan:                     No problem.

    Gary Vaynerchuk:               Surround your weaknesses. Just 'cause you're awesome at video creation, doesn't mean you should punt the written word. So maybe you get somebody who's entry level and young, and maybe they're not as great as you want them to be. But everybody watching this right now should be producing content on all three key pillars: the written word, audio, and video.

                                                          But they, as the lead captain, need to be putting their energy on the one that they're best at.

    Brian Halligan:                     Got it. Social, created the content, leaned into my strengths, got other people to help me with my weaknesses. Where do I spend my energy, whether it's my personal energy, or interns, or people helping me out? Instagram, Facebook, Snapchat, still Twitter? Let's just say you're getting started today in marketing, and you're building a social media brand. Where are you spending your energy if you're that marketer?

    Gary Vaynerchuk:               To speak in general terms to bring the most value to that question,  and I'll go into nuances, you can't be alive in the game without a Facebook and an Instagram. I genuinely believe that. I think it would be highly detrimental. If you're a brand that is trying to reach 30-year-olds and under, I believe that then you would add Snapchat immediately into it.

                                                          If you want to build for the long, long term, I believe you can't live without Twitter 'cause it's the only place where you can still community manage, and you can build real affinity and real depth. You can also culture hack and find your moments. So if you're looking to build a 10-year business, Twitter becomes imperative. But Twitter's the least sales push of the platforms. So there's that.

                                                          I also think that, at this point, I just don't like the idea of not having either ... I'll be honest. Knowing your audience pretty well ... I think the same way there was a point where you and I decided that you had to have a website, there was a point in time where you and I decided that you needed to have a blog, and there was a point in time where we thought you had to have a social media account, I think I'm on the verge of believing that it is unacceptable for a modern-day business to not have a show.

                                                          I mean this. Either a call-in show ... As you can see by all the personalities, the interview show is the easiest. Right?

    Brian Halligan:                     Yeah.

    Gary Vaynerchuk:               What you're doing now is a piece of cake. You ask three questions, and you let the person ... Right? Everybody can do that. Everybody can become the B2B magazine 20/20 of their industry.

                                                          are watching right now are wasting tons of money on booths at trade shows that if they took that same money and hired a sound engineer and a crafty, entrepreneurial kid, could have the leading podcast in their sector and have everybody listening to it, which is all but 900 people, yet that means they can get all the talent to come and work for them, they can get customers.

                                                          So I think I'm at a place now where I would challenge everybody that, if you don't have minimally a monthly show, for me, if you really have ambition, a weekly show ... not everybody is narcissistic enough, at this point, to have somebody following them around like I do to do a blog. But everybody can interview.

    Brian Halligan:                     Got it. Gary, let's talk about Facebook itself. You have, I believe, mastered the medium. What advice do you give on leveraging Facebook as a platform, as a marketer today? I want to hear about that, and then we'll talk about messaging apps.

    Gary Vaynerchuk:               Great. So I think Facebook is a place where ... first of all, everybody needs to understand their ad product is grossly under-priced.

    Brian Halligan:                     Still?

    Gary Vaynerchuk:               Yeah, still, and I'll get into the nuances of why I believe that. If this was Google 2002, I'd be sitting here and be like, "Guys, AdWords is under-priced." Forget about long tail. Even the core words are probably under-priced because it's a game of alternatives if you spend ad dollars. Facebook's CPM's might be higher. It's just, what are you gonna do with that $1,000?

                                                          So if you're in the direct response business, Google still really matters, intent-based marketing really matters. But you can get some of those nuances. You can get 80% of that on Facebook, and you get 100% branding on Facebook. To make it simple, I think everybody has to think about Facebook as the birth child between television and direct mail. If you do that, you're in the right mindset of why it's special.

                                                         The reason I think I've done well is 'cause I've made that religious belief in the platform. I'm a practitioner, so I know how the ad product and the segmentation works. I know that I can run a piece of content against employees of companies. So if I wanna win in real estate, I can run ads against employees of Keller Williams. There's just so much you can do. It's crazy. People are so naïve.

                                                          I make content that brings people value. I think most of the content you see today is selfish, and I make selfless content. It's selfless in the short term. It's selfish in the macro. I'm playing for legacy and long-term wealth, not for sales and short-term dollars. So the content I'm making isn't to convert you into a funnel to transact on. It's to build the relationship and bring value over the long term.

                                                          I think it's why I do well in the I'm not there to kind of innuendo you into my services, like you know so many speakers do. I'm there to be historically correct so I can look back and be like, "See? I told you so," and trade on that.

                                                          So I would say selfless content. Do you know how many people that are watching this that should know more about Facebook and don't because they're headline readers? They're not real practitioners. They haven't spent real money on the platform. They don't really understand.

    Brian Halligan:                     Let me interrupt you for a sec.

    Gary Vaynerchuk:               Please.

    Brian Halligan:                     What if you don't have a lot of money? Can you still make Facebook work for you? Or is it, gosh, it's more pay to play, and it's just too hard to do it now?

    Gary Vaynerchuk:               If you don't have money, you have time and hustle, and it means you don't have money for Google or television or radio either. So it's the reality of business, right? To me, it's about what alternatives can you make? Yeah, I mean, look. To me, I think you start reaching out, becoming part of the community, asking people with big fan pages to share your content so you start building organically.

                                                          But to me, listen, I think you can do real damage for $1,000 a month on Facebook. I really do. I think you'll probably lose $5,000 for five months in a row figuring it out, but I do think that there's something there. So, yeah, look. I mean, you know this. Every medium is pay to play: blogging, content marketing, something you guys were iconic in.

                                                          When everybody started throwing up bloggers and WordPresses, was it pay to play? I guess. I mean, if you had money to bring awareness to your blog, you probably did better and got bigger faster. But could you be just great at it and eventually be found?  Of course. Merit has a spot in this game forever.

    Brian Halligan:                     Okay. Let me try something on you. It feels to me that over the last five years, the role of organic and paid is shifting a bit. So for example-

    Gary Vaynerchuk:               It has.

    Brian Halligan:                     Five years ago, you write a great blog post, and you can get found in Google. Now, if you do a Google search on your mobile phone, which is most of the places you're doing Google searches, the whole darn page is paid ads. You do a great post on Facebook, you really need to post it ...It feels to me the world is becoming a little trickier to be a small business in.

    Gary Vaynerchuk:               1,000%.

    Brian Halligan:                     Yes.

    Gary Vaynerchuk:               Yes. Now let's talk about a couple of ways. Number one, first of all, I wanna do this 'cause I want to bring value to people watching. Everybody talks about boosting 'cause it's just what we got used to, and I don't know how you're referring to it. I just want to make sure everybody understands I don't want to boost posts so I get more likes. I want to plan ads before we post the content so that people who I want to see it, see it. It's a huge nuance, Brian.

                                                          There's a lot of people boosting to get more likes so they feel good about themselves. I want to target 41 to 44-year-old men who love the Knicks and make a Patrick Ewing joke in the copy to make my transaction. Got it?

    Brian Halligan:                     Yeah.

    Gary Vaynerchuk:           Number two. That's right. Let me tell you where else this has happened: every other medium in the history of time. I don't know your history, just 'cause I don't know it, when it comes to email. But email was super fun for me at first because in 1996, I had 91% open rates. I don't have 91% open rates on anymore.

    Brian Halligan:                     I bet you don't.

    Gary Vaynerchuk:               Internet advertising in 1997/8, I used to be able to get people to put up a Wine Library banner for free 'cause they thought it was cool. Not so much anymore. Influencer marketing, super fun for me six years ago. $100 got me enormous awareness. Not so much anymore. People ask me for 50,000.

                                                          It is always supply and demand of being right. There's enormous opportunity in voice right now. Build Alexa skills for your business. You'll be one of the people that has it in your industry. People will download it. Guess what's gonna happen? When everybody has an Alexa skill in six years, it's gonna be harder.

                                                          So, absolutely. Why do you think I was yelling at everybody so loudly in 2009 to take advantage of social media? I knew it would get more expensive.

    Brian Halligan:                     Okay. I think we're about out of time. What I wanna do is thank you very much for your time, and your energy, and your passion you bring to this whole endeavor. It's inspiring.

    Gary Vaynerchuk:               Thank you, Brian.

    Brian Halligan:                     I want to invite everyone to come tomorrow. One of my old professors, Sinan Aral, from MIT, who's just a brilliant guy and the nicest guy on the planet, he is gonna be talking about his latest research and what's going on in social media.

                                                         So join us tomorrow. Thanks, everyone, for coming.

    Gary Vaynerchuk:               Thanks, guys.

    Brian Halligan:                     Thanks.

  • Speaker 1:             [ello, everyone. Thank you for tuning in to day two of the Four Days of Facebook. If you are watching live now, you'll be able to find all the days' sessions on our Facebook page. If you registered, you're going to get links in your email or by Facebook messenger, so keep an eye on those. We'll also be sending some recap emails with links and with other resources that you can use to get into some of our products to see the recaps to learn more. Today ...  Hopefully you attended the session before this with Gary V. and Brian. Hopefully, it was great.

                                        Today, we have Sinan Aral from MIT who's going to talk to us about social influence. Sinan is the David Austin professor of management and professor of information technology in Marketing MIT Sloan here in Cambridge. He is also a startup founder. His research focuses on social contagion, product virality, and measuring and managing how information diffuses in social networks such as Twitter and Facebook. He's writing a book called Height Machine around social technology and how it's transforming social business and health. Give us a virtual round of applause for Sinan.

                                        All right, thank you guys.

    Sinan Aral:            Thank you.

    Speaker 1:             Sinan, thank you very much-

    Sinan Aral:            Thank you.

    Speaker 1:             ... for being here. I'll let you take it away.

    Sinan Aral:            Thank you guys. Thank you so much. Thanks for the invitation. It's an honor to be here. It's always great to see a company that is partially born of MIT succeed and grow so rapidly. Let me give you a quick story about how I got invited here, just a lesson in social influence. Brian Halligan came to MIT, he walked up to my office, he came in, he closed the door behind him, and then he twisted my arm behind my back, and he said, "You will be speaking at the Four Days of Social Influence." That is how you conduct social influence. He is quite persuasive, so here I am.

                                        Today, I wanted to give you a little bit of the science, the rigor and the really deep scientific research that goes into how we can really understand social influence and social advertising in today's online digital world. I'm going to do a bit of a whirlwind tour of some of the major studies we've done with a number of the companies you all know and love to give you what I think are some very modern scientific insights into the way the world works.

                                        Straight off the bat, there's been an explosion of digital social signals in recent years from all of these different platforms and media, and these signals come from your peers, but also the crowd in the form of ratings, reviews, and so on. This has been an explosion in a very short period of time. If you think back, Facebook was founded in 2004, the iPhone in 2007, so if you go back 10 years, maybe 12, the only technologies we had to communicate with one another, phone, fax, and email.

                                        Now, think about the way you get information today. Think about Twitter, Facebook, Instagram. You know exactly what your parents had for dinner last night because of these social platforms, and we are constantly bombarded by trillions of tweets, posts, pokes, referrals, notifications, et cetera, and all of these messages are influencing what we think about products, candidates, what kind type of health activities we should be involved in, where we should go to dinner, where we should stay in a hotel, where we should spend our vacation, and the like. Everything about the way we make decisions has been hyper socialized in 10 short years, and that is a massive difference in the way human beings make decisions today compared to how they did just a few years ago.

                                        This obviously affects commerce. This is a UK Wired cover story. It says, "Commerce Gets Social: How your Networks Are Driving What You Buy." The implication is, your friend buys something or stays at a certain hotel or goes to a certain restaurant, they post that on Facebook, and that has social spillover effects to everyone in their network. They realize that, "Oh, there's social proof. Somebody that I know and trust ate at that restaurant, likes that restaurant," so it can create massive changes in the demand patterns in society based on socializing the choices that people have made in the past, but it's not just about commerce. It's also at the heart of social revolutions. These technologies are being used to organize and incentivize and influence people to get out into the square and protest to make strides towards democracy around the world.

                                        Not just about that, it's also massively influential over our politics, our elections. The way we think about what's going on in the world has been dramatically changed by these technologies, and it's not just business and politics, but it also has dramatic effects on our health. Many of us now wear devices that track our health, how far we run, how many steps we take, and so on. A lot of those devices socialize that information with our friends by posting to Facebook, "I just ran five miles this morning," or, "I took 10,000 steps today." What effect does that have on the health behaviors on our friends and how does that spill over in the network to other people that we know and that we communicate with.

                                        At MIT, I run the social analytics lab where we conduct research on these types of platforms and how they're affecting society, affecting our decision-making, politics, business, health, and so on, and we've worked with a number of companies to study this problem from many different angles.

                                        Today, I want to share a little bit under the hood of the science of social influence with you and to give you a little bit of a flavor of how this is radically transforming society. One of those corporate collaborators, Rob Cain, CIO of Coca-Cola, shared this slide with us, and this slide describes four decades of consumer engagement since the 1980s.

                                        He says, "Look, in the 1980s, it was all about a single message. We poured a lot of time and a lot of money into developing a single commercial that would air at the Super Bowl, and that was it. That was the one statement that we had for our consumers. In the 1990s, we focused on segmentation. We realized not everybody's the same, so we'll draw a boundary around 18 to 24-year-old gamers or soccer moms or whatever the category may be, and we'll tailor the message to these broad groups.

                                        "Then in the 2000s, it was much more about customization and personalization. Now, I have detailed data about your individual preferences, and I can tailor messages individually for you and to become a little more effective in my communication and my marketing, but since about 2010, it's really been about the socially-linked consumer, that these individuals are not making decisions or developing preferences in a vacuum. They consume all of these products and messages in a network of individuals who are also sharing their experiences about these messages and products, and that's how consumers really see the world today, and also citizens in terms of how they vote and other things like that."

                                        Really, we're in this era of the socially-linked consumer, and in my lab, we study these massive social networks to really understand how is information moving in these social networks? How is it changing decisions? We conduct not only observational studies with millions of participants, but we also run experiments at very large scale to really get a causal inference, and that'll be a big theme of this talk.

                                        When we talk about social influence in the context of marketing and Facebook and advertising  and so on, we're really talking about peer effects, how your behaviors affect the behaviors and perceptions of your peers. We're really talking about social influence when we talk about marketing on Facebook or places like Twitter and other things like that.

                                        Really, peer effects in social influence are important for at least two very important elements of social marketing. One is prediction, being able to predict someone's preferences, which is relevant to  the marketing action of targeting. What do I mean by that? I mean if I know what your friends prefer, I can build a much better understanding of what you prefer because your preferences are correlated, and I'm going to get into that, but it's also very important for inferring causal behavior change, which is very different than prediction. What I mean by that and the tactic that that's relevant for is peer-to-peer marketing or viral marketing, and that's really about how does your behavior cause a change in the behaviors, perceptions, or preferences of your friends. How does that spill over into a causal effect on your friends?

                                        That's relevant for viral marketing because in viral marketing, we give customers incentives to bring their friends to the product, like for instance, DirecTV has a advertisement that says, "Turn your friends into Benjamins," and the commercial is you're on the golf course or whatever, and everybody's, your friends are there, and their faces turn into Benjamin Franklin because every time you bring them to DirecTV, you get a hundred dollars, and Benjamin Franklin is on the hundred-dollar bill, so this is a causal effect on their decision to adopt the product. You have to go out and change their behavior, whereas targeting and prediction only relies on correlation. If I know what your friends like, I have a better picture of what you like.

                                        That's very different, and that will feed in to what I call the core problem of social advertising, which I'll get into in a minute, so keep those two things in mind, prediction for targeting, causal behavior change for viral marketing.

                                        Let's talk about social network targeting. What is social network targeting, why is it important, how does it work? Essentially, the way it works is through homophily. Homophily simply means that birds of a feather flock together. We tend to make friends with people who are like ourselves, and so if you know the preferences of someone's friends, you have a much better understanding of their preferences. If your friends like hiking, soccer, reading, and The Killers, the band The Killers, you are statistically more likely to like hiking, soccer, reading, and The Killers, and so I don't even have to have any personal data about you. If I understand what's happening in your network, I know a lot about you already.

                                        In this way, network connections can reveal your preferences, and marketers can use this data to better target messages to the right people. Let's say that I have a population on Facebook of two billion people. I have a budget, which limits the number of advertisements I can send. Who do I send messages to amongst those two billion? Understanding how network preferences are correlated and networks can have big effects on marketing success.

                                        Let me give you an example. My colleagues, while I was at NYU, conducted this study called the Network Neighbors Study. They worked with a large telecommunications company that wanted to market a new telecommunications service. They had long experience with targeted marketing. They had sophisticated segmentation models, data, experience, intuition. They had demographic data, geographic data, loyalty data, lots of experience and intuition about who would have an affinity for this new product, and they added one variable to their targeting process, which was whether the network neighbors of a potential target had already adopted the service or not. Just understanding how many of the friends of the potential target have adopted this service?

                                        Here's a hypothetical network. The network that they have, being a telecommunications company is who's calling whom, which phone numbers are calling which other numbers. Here, you have, the nodes are people, the lines are phone calls between people. The thickness of the lines is the number of calls those people exchange, and there's two types of people. The gold nodes are prior adopters, people who've already adopted this service. The blue nodes are non- adopters, people who have not adopted this service. There's really two types of non-adopter. There's a non-adopter that has network neighbors who are adopters. Those are circled in green, so all of these nodes that you see circled in green, they have a friend, at least one friend who has adopted this service. Then there are these two nodes on the other side or these three nodes who have no friends who are network neighbor adopters. They do have friends, but none of their social cohort has adopted this service yet.

                                        What they did was they took their sophisticated segmentation models, and they added, they have all these variables, demographics, geographics, loyalty data, past transactions, all this stuff, and they added one variable. Do you have a friend or friends who have adopted this in the past, and they ran a randomized experiment where they did the targeting with this variable, and they did the targeting without this variable, and this was the result.

                                        Typical of display ads, click-through rates, and conversion rates. Actually,  these are not click-through rates, these are conversion rates, which means these people went on to actually purchase the service. For the group without the additional feature of network neighbor information in their targeting model, you got about a .28% conversion rate in terms of people who saw the ad actually purchased the service. You add that one variable and redo the marketing, and you get nearly a five-X increase in the effectiveness of the advertisements.

                                      All of that information about what your friends do proves very effective in figuring out who to market to on Facebook or on Twitter in terms of network targeting. That's network targeting. We use prediction to understand your preferences by looking at the preferences of your friends.

                                        Another really important tactic is viral marketing. That's like DirecTV. I give incentives to current customers to bring their friends to the product. Very different than prediction and targeting, and the crucial element, the core problem of social advertising in knowing whether to spend money on viral marketing or network targeting is what's known as the reflection problem. You have peer effects and social influence. They're important for prediction and inferring behavior change, causal behavior change, and the core problem in knowing which of these tactics to engage in to be successful is what's known as the core problem of social advertising, and that's the reflection problem.

                                        The reflection problem is very simple to understand. It just says that we know that human behaviors tend to cluster in network space and in time. People who are connected tend to do the same things at approximately the same times as their neighbors. The question is, are they doing this because of their friend's influence or for alternative explanations, and there are many alternative explanations that could create these correlations in behaviors in network space and in time in the network.

                                        One really good alternative explanation is one that we've already talked about: homophily. Birds of a feather flock together. People are friends with others who have the same preferences, so they'll do the same things as approximately the same time because they have similar interest. May have nothing to do with one friend influencing the other.

                                        Another good alternative explanation is what's known as confounding factors. The famous sociologist Max Weber has this great quote, "  If you see a crowd of people all put up their umbrellas at the same time, you don't assume that social influence is responsible."

                                        If I'm looking at this picture, it's people in a field, and let's say that all of these umbrellas were closed, and then the first one in the bottom left corner opens up, and then the one next to it opens up and opens up, and there's sort of cascade of umbrellas opening in this field. You wouldn't look at that picture and say to yourself, "Oh, social influence. It's a cascade or a contagion of umbrellas opening."  You'd go, "Oh, that's probably some third factor that's affecting all of those people simultaneously, which is a rain shower passing from one side of the field to the other," but too often, we look at this social data and we go, "Oh, it's a contagion," when in fact, it's just correlated preferences or some third factor that's affecting people who are linked in greater amounts.

                                        You're saying, "Sinan, you're talking about the rain. What does that have to do with Facebook? I don't really understand." Well, if you're talking about confounding factors,  there are lots of things online that can mimic the rain. For instance, if homophily is true, which we know it is, and you have the same preferences as your friends do, then you are going to have correlated exposure to advertising because people are going to target you with similar advertisements because you fit a similar preference profile.

                                        Those correlated advertisements are like the rain, hitting people who are connected more than people that you're not connected to, so you'll have correlations in what advertising you see with your friends because you have similar preferences, and that advertising can be like the rain in this example creating or mimicking social influence where there is none. You really have to sort out scientifically is there influence or is this just correlation in order to understand should I do targeting or should I do viral marketing?

                                        Let me give you an example. We worked with Yahoo on this study where we collected nearly 30 million users of an instant messaging network, daily traffic of these instant messaging users over time. We also collected detailed demographic and geographic data for these same 30 million users and comprehensive detailed and precise data about their online behaviors and activities. 90 billion page views for these 30 million people.

                                        What Yahoo wanted to understand is how big of an effect is social influence in the diffusion of a new product that had launched called Yahoo go, which was a personalized news and weather service for your phone. You essentially got personalized weather, personalized stock tips and news you curated for your phone, and they wanted to know, are people influencing their friends to adopt this product or not? There's the adoption curve over time the first six months since it was launched, daily adoption rates.

                                        What we did was we first built  a naïve model that just says, "What's the likelihood of you adopting if your friend adopted?" which takes into account all of the correlated preferences as well as the social influence. Then we built a separate model which controlled for homophily and confounding factors and all this stuff that can mimic influence but isn't influence. Here's what we found.

                                        If you just used the naïve model that looks at what's the likelihood of you adopting, given that a friend or friends of yours have adopted, this is the influence curve that you get. The X axis is days from launch, 20 days out, 40 days out, 60 days out, and so on. The Y axis is influence, and in this graph, what it means is that first dot means that if you had a friend or friends who adopted at 20 days after the launch of the product, you were 16 times more likely to adopt the product yourself, which is a huge number.

                                        That shows dramatic social influence, and this is the influence curve that you get. It looks like there's big social influence at the beginning, and there is less social influence later on, but it's still significant, a three-X increase in your likelihood of adopting, but not a 16-X increase.

                                        Now, if I am your data scientist and you're the chief marketing officer of Yahoo, and I come to you with this data, you might smartly, as a chief marketing office, say, "Well, I know that when influence is high, I want viral marketing, and so I want a budget allocation where early on in the lifecycle of this product, I devote 80% of my budget to viral marketing and 20% to targeting, and then later on, when influence is less valuable, I reallocate and now do 80% to targeting and 20% to viral marketing. That way, I capture the effectiveness of both."

                                        If you were the chief marketing officer of the company and I was the chief data scientist, I give you this graph, and you made that what seems like a really smart decision, we should both be fired because the actual influence curve is this one. When you control for homophily and confounding factors, influence isn't nearly as big as you thought it was.

                                        In fact, in the study, we found that a naïve model overestimates influence by up to 700%, which is way off the mark. What you see of the real numbers is that it's a little bit more sensible. You're not 16 times more likely to adopt the product, you're three times more likely, which is still a huge number of 300%. It's pretty constant throughout the life cycle of the product, so what you should actually do is have a stable allocation between viral marketing and network targeting across the lifecycle of this product that is more in tune with about a 300% increase from social influence, not 1,600%.

                                        We like that study. That was nice, but we wanted to do something a little bit more rigorous, which was to run an experiment because experiments are the gold standard of causal inference. We ran a big study on Facebook, which included almost two million people, and we asked a slightly more sophisticated question, which was about what we call viral product design. Instead of slapping a viral marketing strategy on top of a product that already exist, can we engineer the product or service to be more viral from the beginning, and that's what we call viral product design. Here's how we did it.

                                        We worked with a movie studio that was launching movie app on Facebook. It was an app where you could read about movies, read about the directors and actors, you could buy tickets with your friends, you could rate the movies, you could read about ratings and so on. It was like Flixster for Facebook. It was an app at the time that Facebook had a lot of apps on Facebook.

                                        What we did with this application in collaboration with the studio was we took the application, and as people downloaded the application, we randomly assigned them to control and treatment groups. For the treatment groups, we randomly enabled certain viral features, and for the control groups, we disabled these features. On the right-hand side there, you see a treatment group user with three friends, and the lines to those friends are solid, indicating that those channels of communication have been turned on. The control group user in white also has three friends, but those lines are dash, indicating we randomly toggled those features off for those people, so they were in a position to not have access to any of those features. Then we just measured, how does the addition of these features inquiries the diffusion of this application in the network?

                                        Again, we have these three versions of the application. No features, one viral feature, and another viral feature, which I'll describe in a minute, and then we release these three different versions of the application into the wild, and we just watched as they diffused from person to person, and we measured the diffusion rates, the speed, the number of people who adopted and so on. What is the causal effect of adding this viral feature to the spread of the application amongst friends?

                                        We observed the adoption and use of the application by friends of the control and experimental groups, and here's what we found.

                                        The two viral features that we studied were personal invitations and broadcast notifications or a passive awareness campaign. Personal invitations, in that version of the experimental application, there were buttons that said, "Invite your friends." If you click on that, you get a list of all your Facebook friends. You could click who you want to send it to, write a personalized note, and hit send, and it would go to their inbox.

                                        In the passive awareness campaign, it was broadcast notifications, so every time you took what was defined as a key action on the app, you rated a movie, you friended somebody, you bought tickets, et cetera, it would send a message to all your Facebook friends that said, "Sinan just rated Terminator 2 four out of five stars on Facebook on this application. You might be interested in this application. Here's the link to download." Now, Facebook's terms of service and the way Facebook works has changed, obviously, since this application, but the insights are still invaluable. Let me show you what we found.

                                        We found that the personal invitations were about three times as effective as the broadcast notifications in terms of getting people to adopt. At 6% of the time upon receiving a personal invitation, people adopted the application, 2% of the time upon receiving a broadcast notification, people adopted the application. The personal invitations nearly doubled the global diffusion of this application in the network, which is a big number, but it was the passive awareness campaign that had a much larger effect.

                                        Now, this might seem counterintuitive because I've just told you that the invitations are three times as effective, but the broadcast notifications are creating more adoptions, but that's the difference between the influence of a given type of message and the reach of the message.

                                        What's happening is that there are many, many, many, many more broadcast notifications because they're automated, they go to all the friends. People have to choose to send the invitations. The invitations are more effective, but many fewer invitations are actually sent, so the broadcast notification app had more diffusion.

                                        Now, in terms of ... Now, you say to me, "Sinan, you tricked me. I mean, first, I thought invitations were the way to go, now I know that broadcast notifications were the way to go," but it's actually very nuanced because the personal invitations create a stickiness, which keeps people on the application longer than the broadcast notifications do, and that's because of what economist call local network externalities because I am much more interested in sticking with an app that I'm using with my best friend than using that app with a random person in my Facebook network from high school that I don't ever talk to that happened to get a broadcast notification from me. When I am with my friends on the app, that social engagement creates a stickiness, which creates a 17% sustained reduction and churn compared to the other version.

                                        When we published this study, the Harvard Business Review had a two-page expose on it, and they say, "Which of these feature you used depends on what you're trying to accomplish. If you want spectacular growth with a heterogeneous population and you don't mind or can afford a high degree of churn where people are dropping off the application, then broadcast notifications are the way to go, but if you don't mind having slower growth, but a sustained community of people who are going to stick with the application, then invitations are the way to go. Obviously, you can combine  the two, and this is sort of scientific evidence about what you're going to get from different types of messaging on these social platforms."

                                        That was also a fun study, but we wanted to dig deeper into the individual level and ask, can we identify individual people who are influential or susceptible to influence in their social networks? We took the same app that, the movie app that we were working with, and in that broadcast notification where you rate a movie and it's going to go to all of your friends, we randomly blocked some of those messages so that we could measure how influential any given person's messages are on their Facebook network.

                                        For instance, imagine I'm standing in the middle of my Facebook network, and all of my friends are circled around me. I rate a movie four out of five stars. It creates a message that would go to everyone. Now, the experimenter, MIT, blocks some of those messages, so the receipt of that message is randomized, whether you get that message or not is randomized. Then what we do, we take the people who got the message, randomly selected to receive the message, and compare their adoption rates and their engagement rates to the people randomly selected to have the message blocked. That gives us an individual measure of Sinan social influence over his network.

                                        Once I have that process down and that analytical framework down, now I can do that for any observable characteristic of individuals. For instance, I can ask, "How influential is Sinan?" Then I can ask, "Well, how influential is Sinan over the women in his network compared to the men? How influential is Sinan over 18 to 24 year olds or people over 50? How influential is Sinan for Europeans versus Americans?" and I can do that for every individual, and then I can aggregate that up to the population level and say, "Who's more influential about movies? Men on women or women on men, and so on and so forth.

                                        We randomize the receipt of these notifications. Only a randomly selected subset of neighbors received the peer messages, and we get to analyze the causal evidence of influence and susceptibility at the individual level. When I aggregate this up, I can show you results like this.

                                        Here is your susceptibility to influence as a function of your relationship status on Facebook, for instance. Remember, for every observable  characteristic, I can see how influence and susceptibility vary with that characteristic. If you're single, you're more susceptible to influence than if you don't report your relationship status on Facebook. That's the hold-out set that's not shown here. All of these are compared to each other and to that hold-out set. If you're in a relationship, you're even more susceptible to influence than if you're single. If you're engaged, you're even more susceptible to influence than if you're in a relationship, and if you're married, you're not susceptible to influence at all apparently, and if you label yourself it's complicated in terms of your relationship status, you'll pretty much do anything anyone tells you on Facebook, which is a really nice thing to know about who you should be marketing to.

                                        All jokes aside, when we aggregate this up to a population of 12 million people for which we have data, we can find some really interesting things about how influence ebbs and flows in the network. For instance, this is a heat map that shows your own influence on the Y axis and your own susceptibility on the X axis. White, hot regions are areas of high prevalence. Lots of people there. Dark regions, not a lot of people there, so less prevalence.

                                        What you see here, if you look all the way to the top of the Y axis at 2.5 and look all the way over to that dark corner, people who are highly influential tend not to be susceptible to influence. If you look at the bottom-right corner, people who are highly susceptible to influence tend not to be influential, and if you look at the gradient, there's this general trade-off between influence and susceptibility. The more influential you are, the less susceptible to influence you are in general, which is a very interesting finding about the patterns of influence and susceptibility in the Facebook network.

                                        Now, this might explain some really important societal trends, like for instance, why Kanye West on Twitter has 26 million followers, but only follows one person. Highly influential, not really willing to be influenced by anyone.

                                        We can also look at, for instance, your influence on the X axis against your friends' influence on the Y axis. What you see here in the middle, that white, hot dot in the middle, that's a group of people who are above average in influence who have friends who are above average in influence. Those are good people to target because if you change their perception, then they will influence their friends and their friends are also influential, and they will influence their friends.

                                        Now, if you take that white dot and you move your eye all the way to the left-hand axis right between 2 and 1.5, those people are just as influential as the people in the white dot. They just don't have friends who are influential, so they're much less likely to give you a sort of two-hop influence from receiving a message from them.

                                        That's targeting and viral marketing. What about social advertising? In Facebook, this is probably one of their biggest sources of revenue and one of the biggest marketing activities you can engage in on Facebook.

                                        The way it works is, essentially, that identity and content are intimately linked in social media. Every time we post on Facebook or Twitter, our name, our actual name on Facebook or some sort of moniker on Twitter or Reddit is shown along with the content that we're posting, and that's what makes social media social. We need to be able to attribute this content to somebody and relate to that person. That's what the social part of it is.

                                        Obviously, Facebook, this is a foundation of new social advertising in the sense that Google's friendorsements are when you do a search, it doesn't just show you, for instance, Google Nexus 7 16 gigabyte has 2,500 reviews and gets 4.5 out of 5 stars, but it also shows you that your friend Roger Willis gives  it five stars and says, "Great value." That's social advertising because it puts your friend's social proof in the advertisement for the product.

                                        Obviously, Facebook runs on this too, if my friend is a fan of Blockbuster, his name will appear in ads from Blockbuster for that social proof that will persuade me to be more trusting of this ad or to think that it would apply to me. A lot of this has to do with the fact that we share preferences with our friends, so when I see, "Oh, my friend John liked that, I generally like things that John likes, then I'm much more likely to be interested."

                                        We've been doing research on this kind of advertising since the mid-2000s, and let me just show you some of the most recent stuff, hot off the presses, literally finishing this study this week. We did a massive experiment with WeChat among 37 million WeChat users using advertisements in their Moments feed, which is very similar to Facebook's news feed, social advertisements in the news feed. Essentially, these are advertisements for cars and other types of things where you can either add a social cue, like the picture and name of one of your friends to the ad or not. This is a very simple experiment where we either included the social cues randomly or we excluded them randomly, and we can measure the effectiveness of social advertising  across multiple different industry verticals and product types.

                                        We had 71 different products across 25 product categories in this 37 million-person experiment, and we measured exactly how much lift you get from including the social proof in an ad for a car versus for an electronics product or for a food product or for a hotel chain.

                                        Here's what we found: dramatic variation in the effectiveness of social advertising across product categories. Those are a little bit hard to read, so let me highlight some of them for you. Some of the top high effectiveness social advertising categories: fashion, food, cars. Some of the less influential categories: electrical appliances, eCommerce platforms, financial services.

                                        Let me just say that all of them experienced a positive lift from social advertising. Some of them experienced a three-X lift over the bottom, so there's great variation, but they're all being lifted by social advertising.

                                        Then we wanted to think a little bit more theoretically about this, how does this work and why is it happening, so we looked, for instance, at status goods versus non-status goods, and we found that status goods were affected by social advertising much more than non-status goods, and then we looked at the difference between experience goods and search goods. Experience goods are goods that you have to experience or hear the story of the experience of your friend to really understand. Search goods are goods that you can basically understand everything about them by reading the product description. A digital camera might be a search good, a hotel might be an experience good. We found that social advertising was approximately similarly effective for experience and search goods, even though slightly more for experience goods than for search goods.

                                        That leads me to the future of social advertising. If we have these individual estimates of influence and susceptibility and we know that social proof lifts the effectiveness of advertising, can we be a little bit smarter about how we do social advertising? What about who should I show in a social ad and to whom for which products because you shouldn't just take a random person. I might be very influential over which statistics books you should read, but not very influential over fashion trends or something else. Choosing who to show in an advertisement to whom for which product is the next step in personalized social advertising.

                                        We ran an experiment on a social news aggregation website like Reddit to understand how the identify of an individual can affect people's social engagement with that individual. What we did was we randomly anonymized 5% of the posts that people were making on this, just erased their name, and tried to look at the effect of what it means to show their name on their likelihood of getting upvotes on their comments and also replies to their comments.

                                        What we found, essentially, was tremendous variation. Commenter's identity significantly changed turnout and positivity rates, upvotes, in both directions, negatively and positively for different people. What that means is, some people, when you show their identity get a boost in reply rates and upvotes to their comments when people know that it's Brian Halligan who's posting, and for some people, it's a real negative to show their name. When you show who it is, they get more downvotes and less replies just by showing who it is, like if Sinan posted, nobody wants to comment, nobody wants to vote or upvote. They want to downvote.

                                      Since there's this tremendous variation, you can imagine that these different people are influential, and either showing their name or not, it means that using these personalized identities can affect how effective social advertising is on these platforms. We went back to WeChat, and we looked at, well, what if we looked at the characteristics of the individuals who are being shown in those social ads. How does that affect the effectiveness of social advertising for status, non-status goods, and experience and search goods?

                                        What we found, we looked at two characteristics of the person you might put in an app. One is the status, which we measure by how many friends they have on the platform and how central they are in the network, which is a typical measure of status, and another one was expertise, which is measured by how many articles did they read about that product category in the last period of time? Are they somebody who's up on this category or not?

                                        This is the effect of the status of the person you show on the ad for the effectiveness of social advertising across status and non-status goods and experience and search goods. What you see is that the status of the person you show in the ad really boosts the effectiveness of ads for status goods, which makes perfect sense. It also boosts the effectiveness of ads for non-status goods, but at a lower rate than for status goods. The status of the person you show in the ad affects the effectiveness of ads for experience goods at a very high rate, and for search goods, not even statistically significantly at all. The status of the person you show in the ad does not affect the effectiveness of that ad for a search good, but it does for an experience good.

                                        We also looked at their expertise by how much they're up on that topic, and we found essentially that expertise affects status and non-  status goods and experience goods at approximately the same level, and expertise doesn't really affect search goods. Why? Because all of that information is easily researched by the individual themselves online. I don't need your expertise if it's something that I can just read for myself.

                                        These are the foundations of personalized social advertising because it tells you what are the characteristics of people that will determine if you put them in an ad for a particular person whether it will lift that ad's effectiveness and by how much.

                                        I'm ... How much time do I have? I think I'm going to skip this.

    Sinan Aral:            10, 15 minutes. Okay. You know what, I will do this since I have the extra time.

                                        I want to cover two more topics, quickly, in 10, 15 minutes. The first is social influence as it relates to ratings online. I had an experience where I was  living in New York and I went to a restaurant, and I wanted to rate this restaurant on Yelp. I went to this restaurant near NYU, it's called Dojo, and it was an average experience. The food was average, the service was average. Everything about the experience was average, so I wanted to rate this restaurant on Yelp. I logged in, there I am, I'm logged in. In the top-right corner, there's my profile. Here's where I would go to rate that application, I mean, that restaurant, one out of five stars.

                                   Given that I had an average experience, I want to give it a three. Right as I'm about to give the thing a three, I notice this bright red five-star rating from Shar H. who's waxing poetic about, "The prices here are amazing and ... that fresh and amazing sweet ginger dressing was so good," and I was like, "Well, she's got a point. The prices were pretty good for what I got, and that ginger dressing was really good," so I gave the place a four instead of a three.Then I stepped back, and I said to myself, "Well, that's not good," because if I'm being influenced by the rating of the person who came before me, then there can be dramatic herding effects in ratings that could affect business' strategy with how they deal with ratings.

                                        We did another experiment where we randomly manipulated ratings on a social news aggregation website. We randomly gave some people a thumbs up for a post and some people a thumbs down as the very first rating that was experienced on that site for the post, and then we just measured how does this affect the ratings of these randomly-selected posts up or down. The quality of the post or the identity of the poster cannot affect this result because we're randomly choosing who to upvote and downvote, so everyone is, on average, identical in the treatment and control group.

                                        If we find any effect on future ratings from this, it's completely due to this herding effect being influenced by the prior rating. Here's what we found: When we gave people a thumbs up on a post as the first rating, it increased the average rating of that post by 25%. That is a massive effect. What you see here are the distributions of ratings for the positively manipulated in red, negatively manipulated in blue, and the thin gray line is the control group, and you see the entire ratings distribution for the positively manipulated is shifted to the right. We're not counting our manipulation as part of that, obviously. There are thousands of votes here.

                                        All of this is due to herding and social influence in ratings, and this creates superstar effects in ratings. This positive herding snowballs into ratings stardom. You're 30% more likely to exceed a score of 10 if we gave you a randomly-assigned upvote at the beginning of the rating cycle, and that's no small  feat because the mean rating on this website is 1.9. With 30% greater likelihood, you shoot off into the stratosphere of ratings because of this path dependence.

                                        That has clear implications for businesses. Have the customers who had a positive experience rate, and rate early because that's going to affect everything that comes later, but it also has implications for these platforms. How do you police fraudulent ratings? Reddit has a black box machine-learning algorithm that seeks out fraudulent ratings, and when it finds one, it rips it out of the system, but what does that do to all of the legitimate ratings that were influenced by that fraudulent rating. Those aren't fraudulent. You can't rip them out, and the fraud is sort of seeped into the system.

                                        Six months after we published this study, Reddit had a blog post with a dramatic policy change. They said, "We're not going to show votes on any given content until at least three or four hundred votes had been accumulated it," and the blog post said, "We are doing this to prevent social-influence bias in our rating system," so this can affect both how platforms are designed and how marketers use these platforms, but they also have societal implications.

                                        When we published this study, it was the presidential election, and I was listening to NPR in my car, and I was always hearing polar results, like who's ahead or who's favorable ratings or unfavorable ratings and so on. Then I asked myself after doing this study, I said, "Are those polls predicting the results of the election or are they driving the results of the election?" Political scientists have proven that if your candidate is ahead, that can suppress voter turnout. You go, "Oh, well, I don't need to vote. The guy I want to win or the woman I want to win is already winning, so I don't need to vote," but what this shows is it could actually change your opinion on who to vote for based on what society thinks is a good or bad candidate.

                                        I'll end with one final study, which is the effect of social influence on our health. We did massive study with a large global fitness tracking company where we analyzed data from nearly 14 million runners, global running data over five years, half a billion runs that these people logged during those five years. These are really important statistics like how people run across the days of the week. You can find very interesting things out from statistics like this, like people run more on the weekend than on the weekday, people don't run while they're sleeping. These are very important statistics, but we can learn much more sophisticated things from this as well because we have the social network of this fitness tracking application, people who follow each other and are informed of each other's running activity on a daily basis. We know who runs, how they run, when, where, and how fast they run, who their friends are for a global network of runners over five years.

                                        What we're interested in, actually, in this particular study, and we're looking at habit formation, habit dissolution, and so on, but in this study, we're interested in understanding what I talked about at the beginning: how your running behavior affects the running behavior of your friends. If you run more, does that inspire your friends to run more? Maybe you have friends on Facebook that run, and then they post their run to Facebook, a picture of them all sweaty with a little map of where they ran that said, "Seven miles, I ran seven miles this morning." That's a social signal from a friend indicating their running behavior.

                                        How does that affect your running behavior, so we wanted to know, is exercise contagious as enabled by these digital platforms, but we couldn't really run a randomized experiment in this setting because it's hard to walk around with a cattle prod and prod some people to run more and randomly select some people to not get prodded to run more. We couldn't figure out how to make that work, so we needed to find some sort of natural exogenous analog to the randomized experiment, so what we used was the weather.

                                        This is a graph of running and precipitation in New York City over a year's time. You have the millimeters of rain on any given day in blue, and the running activity per capita in green, and what you see there is every time you have a spike in rain, you have a drop in running behavior, which is just visible to the naked eye. You might also experience this. When it's raining, you don't want run. Makes sense. When it's beautiful outside, you run more.

                                        We also had GPS-trace data for these people's runs, and on the left-hand side there, you see people running on a sunny Saturday in New York City. By the time the afternoon rolls around, Central Park lights up because lots of people are running there, couple of people even run in Jersey. I didn't know that happened, but Manhattan is awash in activity. Lots of people running, even over the bridges and so on and so forth. On a rainy Saturday, much less activity. The weather is a very good, natural randomized experiment for trying to understand how running behavior is contagious or not contagious, influencing your friends to run.

                                        What we did was we collected this weather data from about 40,000 weather stations across nearly 40 countries for which we have running data. This is the location of the 2,700 weather stations, the United States, and I just put up this map to show you that the density of our weather data maps very well to the actual runners that we have in the major cities and so on.

                                        The trick that we're using here is we essentially want to understand, let's say that there's a person, a runner in New York City who has a friend who lives in Phoenix, Arizona. The person experiencing rain in New York City, their friend is not experiencing rain in Phoenix, so what we want to do is ask a very tongue-in-cheek question, but very rigorously, which is, does rainy day in New York reduce running in Arizona, and if it does, that can only be due to social influence because the rain in New York is not happening in Phoenix, Arizona.

                                        What did we find? Massive social influence effects. When you run more, your friends run more. If you run an extra kilometer today, [00:53:30] that will inspire your friends to run three tenths of a kilometer more today. If you run faster, your friends will run faster. If you run for a longer time period, your friends will run for a longer time period, so massive social influence effects, which decay over time, so this has less of an effect one day out, that's T plus one, two days out, T plus two, those are the three graphs you see in each one. On the left, you see distance. On the right, you see pace. We also looked at minutes [00:54:00] run, calories burned, all with same pattern of effect.

                                        Then we wanted to ask, well, how does that influence vary across different kinds of partners. For instance, we might want to ask, do friends who are less active influence friends who are more active or is it the other way around,? Do couch potatoes influence marathon runners or do marathon runners influence couch potatoes? When we did this analysis, we were a little surprised. [00:54:30] There are all these ... I like in an audience that isn't online to ask people, "Well, what do you think? Is it couch potatoes influencing marathon runners or marathon runners influencing couch potatoes?" Get the answer in your mind and think about whether your couch potato friends or your marathon runner friends are influencing you more and have that in your mind. Let me show you the results.

                                        It turns out that the couch potatoes influence the marathon runners, but the marathon runners don't really influence the couch potatoes, and that's probably because [00:55:00] of social reference points. Festinger's 1959 theory of social comparisons is that we make social comparisons to others that guide our behavior, but if you're so far ahead of me that I can't really relate to you, you're going to have much less influence.

                                        In this case, what we know, there's two versions of Festinger's theory. One is that we make upward comparisons to people who are head of us, and one is that we make downward comparisons to people that [00:55:30] are below us. What we find here, if you notice, the one in this graph is where we have the same activity level, so people who are slightly more active than us do influence us, but people who are much less active than us influence us more. What that means is, just in terms of running behavior, social influence manifest when I look over my shoulder at the person who's catching up to me in the race than when I look to [00:56:00] the person who's way ahead of me in the race. We can find out these very detailed patterns of social influence in this way.

                                        Another one that we looked at, and this is the final result, is gender. Do men influence women or do women influence men and so on and so forth? What we found here was pretty interesting as well. By far, this contagion is most prevalent amongst men, men influencing other men. Men influence other men the most, [00:56:30] women influence other women and men, but women aren't really influenced by men's running habits. Now, this may be true in other things, but I can't generalize to other behaviors, only running, but this is important to understand the patterns of influence and the way they work over time.

                                        The takeaway message from this talk is that social influence is a big factor in many different kinds [00:57:00] of outcomes that occur in society from the way we vote to how we eat and exercise to how we consume products and in every other way. That has been dramatically accelerated in the last 10 years because prior to Facebook, Twitter, Instagram, Snapchat, and the rest, which now are the major sources of our news and of our information, we had much less intensive and automated machine-learning driven dissemination [00:57:30] of these social signals. Now that we live in a world that is hyper socialized and powered by what I call height machine, we really have to have a scientific understanding of how it works in order to more effectively build our businesses, defend our democracy, and improve our health. Thank you.

    Speaker 1:             Awesome. Thank [00:58:00] you very much, Sinan. That was enlightening. Thank you everyone for watching. Tomorrow, we will be back at 3:00 p.m. Myself and Crystal King from our academy team are going to show you how to build a buyer persona, and we're going to help you get inside of some of the tools we've talked about this week and actually do some of the tactics. Hopefully, a lot of what Sinan talked about today can influence how you're thinking about how to target those people, who you should target, how you should think about ads, how you should think about networks, but we will see you tomorrow. Thanks again. Thanks, Sinan.

    Sinan Aral:            Absolutely. [00:58:30] Thanks. Thanks, guys.

Meet the Speakers

  • Gary Vaynerchuk


    Vlogger, Podcaster, Investor, Author
  • Brian Halligan

    Brian Halligan

    Founder/CEO, HubSpot
  • Sinan Aral

    Sinan Aral

    David Austin Professor of Management at MIT Sloan

Day 3 - How Your Persona Uses Facebook and How to Reach Them

September 14, 3-3:45pm ET

Crystal King, Social Media Professor for HubSpot Academy, and Marcus Andrews, HubSpot Product Marketing Manager, will walk you through a persona mapping exercise to figure out who your target market is and how they use Facebook. From there we’ll work with you to walk through the steps of building a campaign.



  • Marcus Andrews:                Hello everyone, and welcome to day three of the Four Days of Facebook. It's hard to believe. It's going fast, it's going very fast. Hope you enjoyed yesterday. Yesterday was a big day. We had Gary Vaynerchuc, our CEO Brian Halligan, Sinan Aral. If you are interested, go back and watch those videos. If you missed it, the were pretty awesome. It was really an action packed day yesterday. I'm Marcus Andrews from the Product Marketing team here at HubSpot. Today, we're really going to roll up our sleeves and get our hands dirty in the product. With me is Crystal King from our Academy team. She's our Social Media professor so-

    Crystal King:                          Hello.

    Marcus Andrews:                ... between the two of us, hopeful, we can drop some knowledge and really impart some kind of tactical, how do you do this stuff sort of thing today.

                                                          Crystal is going to walk you through how you build a buyer persona. So, hopefully this will be fast, we've got about 45 minutes, but hopefully, at the end of today you'll know how to or have the beginning of a buyer persona that you can use to target your ads. Really inform a lot of your campaigns. I'm gonna show you how to get inside of HubSpot Marketing Free, or HubSpot Marketing and create your first lead ad. We're going to be able to take questions today so if you have questions, just drop them on the Facebook page. You can even test it out right now. I would love to know how well you can hear us, if everything looks good, and let us know where you're from. That way we can start to make sure everything's working and take a look and ... or, if you have questions now, feel free to drop them in. We'll be monitoring them and answering them live.

                                                          Great. Okay. So, Crystal is, like I mentioned, she's our Social Media Academy professor at HubSpot. She speaks and writes on social media and today she's gonna talk to us about buyer personas. Crystal, do you have a highlight of the Four Days of Facebook so far?

    Crystal King:                          Yes, yesterday. Sonat's talk was amazing. He talked about social influence and if you haven't seen it yet, you should definitely go back. I'm gonna re-watch it because there were some incredible facts in there that are super useful for marketers.

    Marcus Andrews:                Very cool. Totally agree.

    Crystal King:                          So today I'm gonna walk you through buyer personas. If you're a current HubSpot customer you probably are a little bit familiar with buyer personas. If not, don't worry, we're gonna walk you through a high level overview and I'm gonna take you deeper. A bit more of a deep dive into some data points that are gonna help you target your Facebook lead ads a little bit better. I'm gonna set you up for what Marcus is actually gonna show you. Let's talk about buyer personas and what are they?

                                                          Essentially, as you can see on the screen, they're a semi-fictional representation of your ideal customer. They're based upon research. Some select educated speculation. Very select. But all about your customer demographics, behavior patterns, motivations, the goals of the customer that you ideally want to have purchase your products. Why are buyer personas important, essentially, they are kind of the glue that holds your inbound marketing together. Understanding the customers that you are trying to sell to and what their interests are and who they are is going to help you not only target your advertising better, but it's gonna help you create better content, which will hopefully lead to more attracted ... you'll attract more people if ... You'll get more leads, and hopefully convert those people into customers.

                                                          Let's talk about how you're going ... What does that look like? What does a buyer persona look like after you've created it? Here's one that I'm not gonna dive too deep into yet. This is the demographics for Corporate Cathy. This is a sample buyer persona. She is essentially the ideal buyer for our imaginary product today. You can see there's a little bit of information about her. We know a few personal things about her. That she's got a couple of kids, she's been in her role for about five years, she's a manager with high expectations for her team. She's worried about generating enough leads for the sales team. How do you get all this to look like this? How do you create a buyer persona that's sort of similar?

                                                          You have to start with conducting persona research. You actually need to really do a deep dive into what your customers are thinking about. It needs to be based off of actual research. You can't just assume you know who your customers are. The good thing is is that there's a lot of great ways to do this research. You get started by interviewing your customers. That means talking to your current client base. That's the best place to start. You can talk to your prospects. I would recommend that you talk to co-workers who might be customer facing, that are dealing with customers themselves and what they're hearing. You can also talk to people ... I would say your customers that are both good customers and then also your customers who might be ones you would consider problematic or unhappy potentially because they might have some very valuable insights.

                                                          I would imagine amongst those customers, you're gonna find somebody who's really your ideal target. So, what kinds of questions did you ask?

    Marcus Andrews:                Any advice on setting those calls up or anything like that? Is it, you know, I don't know, you  call someone and say, hey, lets have a buyer persona interview? Probably not, right?

    Crystal King:                          No.

    Marcus Andrews:                ... flow of a normal talk?

    Crystal King:                          I would say in the flow of a normal conversation, would be a better way to do it. You can also do surveys of your customers. I think that's a good way to do it. Or, as part of the flow as you're maybe onboarding them, you can get a little bit more information. You can get this information a variety of ways. There's very specific types of questions you can ask. Some of it you'll know just as a matter of maybe while you're doing your qualification, so things like the job role and the title. What industries are they in, what is the buyer trying to accomplish, what their biggest challenges are. One other thing that's super important is how they learn information about their job. How do they learn and is that through magazines, is it through their colleagues, is it through an academy like HubSpot Academy.  Find out where they're getting their information.

    Marcus Andrews:                It probably needs to be standardized to some point, right?

    Crystal King:                          Yeah.

    Marcus Andrews:                Or, maybe you should ask the same series of questions.

    Crystal King:                          I would say that's definitely a good thing to do.  Because you want to create some synergy between the customer information that you're acquiring so that when you're creating your persona, you have a similar set of information. For example, then once you have all that data, you're gonna take it and put it into a template  much like you see here on the screen. That will have information that is pretty clear about this ideal buyer. What their background is, the key information about them. You want to understand their gender, age range, household income. These are some basic demographics. The buzzwords and mannerisms that might apply to them, which are also interesting but will help with targeted advertising later. What are they most challenged by?

                                                          Then also from a sales perspective, it's really important to understand what are their primary challenges in their workplace and what are their most common objections. Why wouldn't they want to buy your product? One thing that's also very helpful is to have some real quotes that are part of this buyer persona. Because this buyer persona is something that you are going to hopefully proliferate amongst your employees at the company so that everybody knows who the ultimate buyer is. As they're creating content, as their creating their sales pitches, as they're developing marketing material, everyone's marching to the beat of the same drum.

    Marcus Andrews:                I know it helps, it's nice is we kind of came on board, that's one of the first things you learn about is our buyer persona. We have a couple actually. But, it does just that. Like you said, it really helps you get in their head a little bit. Makes it easier for you to start creating content because you already know who you're building it for.

    Crystal King:                          Exactly. What do you do to take it a little bit further? You saw the demographics and the buyer persona, but there's something that's gonna help you do Facebook advertising a little bit more specifically.

    Marcus Andrews:                100%.

    Crystal King:                          The cool think about Facebook is that you can get really detailed. You can know a lot more about your customers than you might have imagined. Facebook's a platform that has all sorts of really personal information on people. How do you reach those people in Facebook, which is a site that most people are using for personal use. That means you need to know a little bit more about your customer than what you might have had in the original buyer persona. We're gonna look at what we call psychographics. Psychographics measure a customer's attitudes and their interests rather than general demographics. Let's look at some of the types of things that you might want to ask instead or you want to discover about your buyers as you're creating a deeper buyer persona.

                                                          Note that on this screen here, that the two columns don't necessarily correlate directly, but they're each important to know about your buyer persona. You want to understand things about them such as their aspirations, their motivations, their opinions on things. The lifestyle that they're leading and the hobbies that they might have. Some of the kinds of questions you might ask is, what brands do they identify with? Are they a PC or a MAC user, for example. What online sites do they visit? Are they interested in TMZ, or are they checking out Newsweek, for example. What types of books and magazines do they read? What are their hobbies?

                                                          You might want to understand if they have a particular political bent. If they exercise regularly, or if they're into certain sports teams. How tech savvy are they? It could be really important, especially if you have a technology product. Then who or what influences their product choices? And where do they prefer to shop? What does this actually mean? It means that you have to find this data and it's easier than you might think. That's the cool part about it. It's that you might be thinking how on earth do I find all of this? Because it's really personal, right?

    Marcus Andrews:                Mm-hmm -

    Crystal King:                          But what's really cool is you probably already have a lot of this information either in blog or website comments, sales or purchase data. You can have industry ports that might have some of this information. The really, I would say one of the best, places to get this information is actually your social media analytics. For example, Twitter and Facebook have these fairly deep insights. I would go into your business pages and look at your insights and you can get all sorts of persona attributes that you might not have really realized are there. I realize it's small on this screen here, but what you're seeing is actually the HubSpot Academy Twitter analytics and the Facebook analytics. The Twitter analytics, for example, show you that HubSpot Academy followers tend to like entertainment and humor. That might mean that I want to adjust my content to be a little bit more funny, right? Get a few more memes or jokes in there?

    Marcus Andrews:                Mm-hmm -

    Crystal King:                          There's a lot of really great information in these social insights that I highly recommend that you dig through, because what they're gonna do is actually help you target your keywords a lot better. Once you have all that data, what's that gonna look like now? Here's the demographics for Corporate Cathy. We talked about that this is the general demographics, all sorts of really basic things about her. She's in her late 30's to early 40's, her income is like 70K, she's got really high ambitions. We talked a little bit about this earlier, but what might it look like if you go a little bit further and get all of that psychographic data.

                                                          This is what Corporate Cathy looks like. Suddenly she becomes a little bit more human, right? You know that she loves her Apple products and she drives a Volkswagen and she likes to shop on Etsy and she does yoga. She's probably a little bit technologically inclined because she's reading Mashable and Fast Company and she has a Fitbit, right? There's a lot of information you can glean about her, so when you're trying to figure out how you're gonna target your Facebook ads, you can dip into these psychographics and your buyer personas and really actually find your ideal customer.

                                                          I'm actually gonna turn it over to Marcus now because he's gonna show you exactly how to talk about, how to map this type of psychographic information into your lead ads.

    Marcus Andrews:                Awesome. Thank you, Crystal. Yes, I think after listening to Sonat, yesterday too, he was connecting some dots about how some of this stuff works or he was telling us about homophole and how birds of a feather flock together and that sort of thing. If you do, you build this profile, this buyer persona of people and you're able to start to connect dots and look at this ... it's one person, but it can help you focus on an entire group of people. I know some of you growth hackers out there are maybe saying this could take me a while to build. There's some steps that go in here, but I promise you, take the time to build this out. Like an hour-

    Crystal King:                          It's worth it.

    Marcus Andrews:                ... you've got the walk-through from the expert. If you take the time to just kind of walk through this and build it, it's gonna save you a ton of time. You're gonna be able to use it right now, as I dive into this.

                                                          Okay. Let's jump into, well, before we jump into the product, let's talk about a couple things really quickly. We are gonna build a lead ad. A lead ad is a newer ad unit for Facebook. It's built specifically for SMB, B2B marketers and it's great, because it works to help convert a mobile audience on Instagram and Facebook into leads. It's something like more than 80% of all social media usage is coming from a mobile device now. Everybody  is spending time on Facebook and Instagram, but they're doing it from mobile. You need to be able to ... Right now, there's a lot of opportunity in that space because a lot of B2B marketers aren't great at it.

    Crystal King:                          Right.

    Marcus Andrews:                If you can figure it out, you can fish where the fish are, you can be one of the first people to kind of get there and convert this audience. Lead ads are great at doing it, you can see a little example here, because they offer a great CTA. You have this ad on the mobile app, it takes up the full screen and the CTA can be something like get a quote or subscribe or download or learn more and it will take you to a form inside of the app. The great thing about this ad unit is that the form is pre-populated with data that Facebook already knows about the user. They don't have to take the time to fill out the form, which is what kills conversion rates on mobile for a typical B2B market.

    Crystal King:                          Makes it super easy.

    Marcus Andrews:                Makes it super easy. If you, as the user, want to edit some stuff, you can do it. As the advertiser, you can ask a custom question, but we're seeing great results from it for our customers. We're seeing higher conversion rates than desktop in some places. The cost per acquisition is still pretty cheap right now, really.

    Crystal King:                          Oh, that's exciting.

    Marcus Andrews:                Yes. Okay. All right. A couple quick housekeeping things, we're gonna jump into the product, too. If you don't have HubSpot, or if you don't have HubSpot Marketing Free, take a second and go to, and sign up. You can do it right now, while I'm still talking, and it'll take you five minutes, and you'll be able to follow along with us. The other thing you're gonna want to do is you're gonna want to check ... If you don't have a Facebook ads account, it's perfectly fine. You don't need one, but if you do have one, it's good to have access to the account. You can check at

                                                          Then, there's two other things that you'll need to do to really make this effective. One is that you need to set a budget, right? It's really hard to figure out how much to spend on ads. Before you get in, know what products to use, you have to justify the budget for your boss, potentially, you have to figure out how much am I gonna spend on this, then the important question to ask is if I'm gonna spend on this, what am I gonna get in return? I don't think anybody should use ads if they don't use them to get for a positive ROI.

    Crystal King:                          Yeah, it's better to have an educated guess than a shot in the dark.

    Marcus Andrews:                100%. 100%. Then, the fourth thing is, you're gonna need a privacy policy for lead ads. We have a resource for that, too. So, let's take a look. Okay. Here we go. Okay. I just mentioned that you should set a butted for your ads. What we have ad is a way to build a budget for your ads based on ROI. You can come to this site, you can say I want to spend $2,000 or maybe I just want to spend $100. What I'd recommend for what we're doing here is a budget of $ [00:18:00] 30. If you could actually take $30 and plug it into lead ads today, you're gonna see some results and you're gonna start to get some data. It's a test that you can actually put into action.

                                                          If your budget is $100, take a quick guess at your CPC. Also, your target conversion rate, your average sales price, and your lead to customer rate. These are things you can kind of guess at if you don't have a great idea. What it's gonna do, it's gonna show you for your budget, this is what you can expect to get based on the inputs that you put in. You don't have to go in blind, you can build a little projection, you can take this to your boss and say, hey, I need $500 or I need $200 to run this test. I think it should get us four or five leads. You're gonna actually have a pretty accurate idea of what's happening there.

                                                          The other resource that I want to share with you quickly is the privacy policy blog post that our product manager of our ads product actually wrote. This is just gonna help you get set up with the very basic privacy policy. It's really easy to jump in here and check out this post and then get started. Okay. Let's look at the product.

                                                          Inside of HubSpot marketing, if you have our HubSpot paid tools, you're gonna find under content a new ads item in the nav. It's gonna take you to the lead ads creation dashboard. If you have HubSpot Marketing Three, like hopefully, a lot of you do, you probably could just set on up in the time that I've been talking here. You're gonna see and ads item in the nav. You click on this and it's gonna take you to everything we need to get started today.

                                                          All right. Once you've made it to this screen, it's time to create your first ad. To do that, what we're gonna do is click this button and we will jump into the creation flow. All right. First up is to connect your account, once you've done that, it's very easy to ... What I would recommend doing, and hopefully you're following along, if you can do this along with me that's great. If you want to just take some notes and get in here and then do it yourself a little bit later, go for it. Feel free to try and launch one today.

                                                          What you're seeing right now is we're inside of Facebook. We're inside of the Facebook ads manager. This is a native integration that we built with Facebook. HubSpot is the first to actually pilot  this integration with Facebook. What you're seeing is the form builder for lead ads. On the right is your preview of the ad, on your left is the form. The first step is to tell Facebook what information you want to collect. Email, and full name, you click more options, it's gonna show you what else Facebook can automatically populate into these forms. So, if you're B2B, check out this bottom bucket here. Maybe you want to capture job title and work email and not email. So you can click off of email, you can say I want their work email and Facebook will do it's best to give you their work email. Or, you can specify that in the form.

                                                           You can capture a lot of information here, but you want to be careful about how much information you do capture. The more, the lower the conversion rate, probably. Be careful here, if you can make these simple? Do it. If this is the first time someone is hearing about you through this ad, you may want to have a very simple form. The second part of this is the ability to ask a custom question, so let me grab this really quick, the ability to ask a custom question. In the custom question, you can ask like, what is ... and you can make this a short answer or a drop down and basically you can have it open-ended or it can be yes or no. You can delete the question. You can do multiple choice and have this as ... Lots of fun things you can do here. It can be a marketing budget, et cetera.

                                                          For the thank you screen you want to put in wherever, this is probably a landing page that you built or a landing page on your site where somebody can pick up the information that you promise them in the ad and then the privacy policy. You want to link to the privacy policy that you build. Okay? We just built the form, let's move onto the next part, which is the creative.

                                                          Now we are in the creative, a lot of these ads, these ads are gonna show on desktop but they're also gonna show on mobile, right? You probably want to preview this in mobile and get a really good idea of what it looks like. You want to make sure that these ads work on mobile, I promise you they're gonna have great conversion rates on mobile. Your headline should be something like, subscribe, this is gonna be what is down here in the bottom. Maybe this is your company name. Something kind of informative right next to the button and then text is gonna be really about whatever it is you're offering people.

    Crystal King:                          Can you tell us if this is gonna also connect into Instagram?

    Marcus Andrews:                Yeah, exactly. Instagram and Facebook, these ads are gonna show on Facebook, they're gonna show on Instagram, they're gonna show on the desktop site, they're gonna show on the mobile app.

    Crystal King:                          Awesome.

    Marcus Andrews:                You can control that in the ad's manager, but the look really nice on Instagram. If you upload the right image, they look almost the best on Instagram, I think.

    Crystal King:                          Great.

    Marcus Andrews:                Yes. All right. You upload your image, for this, why don't we use a pic of our buddy, Gary V. Once you've done that you select the call to action. If this is for a newsletter, maybe we'd do subscribe. Then we get into some tricky stuff. All right. We're in the audience targeting. There's a few more advanced features you can do with Facebook. Things like look-a-like audience, things like retargeting, things like audience expansion, which are awesome. It's more of an advanced feature, it's very powerful and very cool. It's usually kind of step two after you start with more demographic and psychographic targeting. Let's just take a look at how that works.

                                                          Okay. Inside of the audience targeting, you can hit edit. This is where we just did ourselves a huge favor by building this persona. If we go back to our persona ...

    Crystal King:                          Good old Cathy.

    Marcus Andrews:                Good old Corporate Cathy.

    Crystal King:                          There she is.

    Marcus Andrews:                We can see that she is in her late to early 30's. She's female. She has an income over 70K. Let's start with that. We like those numbers, right? Let's say she is 33, 45, female, she's in the United States. Then we jump into interests and a lot of this is really open-ended so maybe we want to get rid of some of the starters that Facebook kicked us off with and we want to say income. All right, what did we say, 50 to 70, does that sound good for Cathy? Okay. We say income, we say that she is married, should come out here. Yep. Relationship status, married. What was our other one?

    Crystal King:                          She has a couple kids. She's a manager in the business world.

    Marcus Andrews:                All right. Well, let's stick with that we have here. This is gonna get us started, right? We can also narrow down by location so we can do things like ... A great way to go and get even a little bit more specific depending on your ...

    Crystal King:                          You could even get really deep and get into neighborhoods if you really want to get targeted.

    Marcus Andrews:                Well, I'd done it earlier, but yeah, you can target neighborhoods and specific addresses and things like that.

    Crystal King:                          Yeah, it's interesting.

    Marcus Andrews:                Yes, which is an interesting use case for like ABM or something like that. You could almost target somebody's building and very fun things.

    Crystal King:                          Yeah.

    Marcus Andrews:                Okay. Now we've got some basic demographic stuff. Let's dig into some of the psychographic points that Crystal was describing. I love the websites and what the read and what informs them, good signal to say this is our people. This is who we want to target. Let's check this out.

    Crystal King:                          There it is.

    Marcus Andrews:                There it is. You can really put in just about anything here. Facebook knows a lot. You're able to do a ton. Let's get even a little further down.

    Crystal King:                          She likes yoga, yep.

    Marcus Andrews:                Yeah. Yoga. As you can see, there's a ton that you can do here. You don't want to over target, you want to make sure you keep an eye on this. One million may be too big, depending on the size of your business, depending on-

    Crystal King:                          The budget, too, right?

    Marcus Andrews:                Especially depending on the budget. Then things like your lead time, your business model, there's a lot that goes into this. These are the basics. Once you've got it, you save it, and you can hit promote and you're gonna go live. Right? Then you've got a lead ad that's gonna be showing to this audience that's gonna be out there. Okay. Once you've done that and you've created this, what we're gonna do is show you some basic analytics on that ad. You can always see more inside of the Facebook ads manager, things like billing and advanced placement and advanced settings will happen inside of the Facebook ads manager. You can always find that data at This is kind of what it looks like here.

    Crystal King:                          Can you do A and B testing on the lead ads?

    Marcus Andrews:                Yeah.

    Crystal King:                          That's exciting.

    Marcus Andrews:                Yeah, so you can create a campaign and you can put different ad sets inside of the campaign. Facebook will actually optimize a lot of that for you. They'll take their machine learning algorithms and they'll look to see what's resonating and what's not  and help optimize your ads for you, which is awesome. Yes. You're gonna see a lot of information on these ads, but most importantly, what you're gonna see is, let's see if we gonna do here, most importantly what you're gonna see is contacts come in. Every time someone fills out this lead form, without the HubSpot integration, they just go into a CSV inside of Facebook, which is unfortunate, because what you really want to do is you want to reach out to that person as soon as they've filled out that lead.

                                                          They're interested in your piece of content. It's great if they can go to a thank you page to check it out, but what if they go to a thank you page, read it, and they also get a call from a sales person or they get a warm email that comes to them right after that.

    Crystal King:                          Exactly.

    Marcus Andrews:                Then they're still in that moment. If you wait a few days or a week, you might miss it.

    Crystal King:                          It's not the same.

    Marcus Andrews:                Exactly. Here we can see an example of a lead that came in through a Facebook lead ad into HubSpot. It's inside, if you have the free product,  you've got them in your CRM. You can easily call them. If you have them in the paid product you can use all of this data as a signal to inform your other marketing efforts. This is what it looks like. They'll come in, they'll look like a lead form. Then you can take, inside of HubSpot marketing, you can build a list based on those people.

                                                          All right, I want to build a new list and I want it to be around a Facebook ad's property. Let's call it a lead ad is true and I do that. It's gonna generate me a list of everybody who has filled out a lead ad. Boom. There you go. I can take this list, I can plug it into a nurturing flow and so that anytime somebody new comes in, they get an email that's introducing them to a sales person. There's another work flow that tells the salesperson that, hey, if this person opens the email, you should call them. You can start to do some really advanced stuff now that we have this integration between Facebook and HubSpot.

    Crystal King:                          I know that lead ads are available for people that have HubSpot Marketing Free. What about some of the other things that you were talking about, the lead flows, and things like that? Where does it end?

    Marcus Andrews:                There's some basic stuff inside of HubSpot Marketing Free. You can really do a great job of getting started with HubSpot Marketing Free. Building your database, getting started building a list, starting to run ads,  starting to capture interest from people who are visiting your website. When you get to the point that you want to start doing things like email and marketing automation and have more in-depth analytics and really build a bigger database, you'll want to upgrade to Marketing Paid. There's lots of ways to do it, there's lots of options. If you're using HubSpot Marketing Free, use it to it's full advantage. Take it and max it out to it's limit and then kind of look for interesting things that you might want to do that you need the marketing tools for.

    Crystal King:                          But it's a great way to try it out before you have to make a decision, right?

    Marcus Andrews:                Yeah. Absolutely. Absolutely. Great way to get started, too, if anybody's entrepreneurs or start ups out there, it's a great tool. When you're a small team, especially. We want to get into questions so to keep going here, all right. Where was I? Awesome. This is what the leads look like and then I think that was everything there. All right, so we talked about reporting, and all right. If you want to see some more advanced stuff you can jump into the Facebook ads manager. This is a really great tool. The cool thing about it is that there's lots of ways you can go really in-depth. If you want to use HubSpot for the basic stuff and then graduate into this, it's totally fine. You don't need to necessarily use one over the other. You can use  both in how it suits your needs.

                                                          They also have a very cool mobile app, which I'm gonna show you guys. Okay. You see this? This is the Facebook ads mobile app, and it will show you basically everything that you see inside of the ad's manager. It's an awesome way to kind of stay up on your ad's campaigns. I think when I launch a Facebook lead ads campaign I want to know everything that's happening. As soon as a lead comes through or the spend starts to go up, it's really kind of fun to watch and the app will send you push notifications. Then if you need help, this is a great place to go. Also, if you need help, you can jump into Messenger. Facebook Business will actually send you some updates on Messenger about your promotion. You can preview it, you can look to see it, you can look at other promotions that you have running and get updates here if you're interested in getting out of email a little bit and using Messenger for some of this stuff, it's a great way to do it.

                                                          You can also ask HubSpot. So, if you got a notification from HubSpot about the event today via Messenger, or if you just go to Facebook Messenger and search for HubSpot, you can sent us a message. You can say hi and we'll reply when we see it. That's a more personal thing. You can also do things like search resources.

    Crystal King:                          Very cool.

    Marcus Andrews:                Yeah. If I search resources I can put in just about anything and what if I say, Facebook lead ads ... HubSpot is gonna go to work and it's gonna look for resources or blog posts and, whoops. Let's try social, but it's gonna go through ... I wanted to search the blog. That's fine . If you search the blog, it's gonna go to work and it's gonna find information about lead ads. You can search for SEO advice or marketing tips. How to create Facebook lead ads, a beginners' guide. That is exactly what I'm looking for, Facebook, thank you for ... Or, sorry, HubSpot. Thank you for bringing that up. It's gonna show me the content right there. I can look at other related content and then I can jump to the page.

                                                          You can honestly think about this as a little HubSpot CSM, kind of in your pocket. You can easily find what you're looking for, it's conversational. We're gonna help you out with things here. We have lots of cool stuff happening with Facebook Messenger. Use it for fun, check it out. We encourage you to use it to get through our content and that sort of thing.

                                                          Whew. We're almost done here-

    Crystal King:                          Yeah, that was a lot of information.

    Marcus Andrews:                Getting into questions. Facebook Messenger is very cool. It does lots of things. It's also a camera. So, here we are. You guys can even say hi to our crew. What's up guys? My favorite part about it is that you can even do stuff like-

    Crystal King:                          Tacos?

    Marcus Andrews:                Eat tacos.

    Crystal King:                          Yeah.

    Marcus Andrews:                This is what social media is really all about, it's about tacos. You can have a lot of fun with it. It's pretty cool and there's a lot that's inside of Facebook Messenger, more and more, every day. There's messaging apps like Line and kind of other things that are happening in different parts of the world where people are using messaging apps for all sorts of stuff. Facebook Messenger is very fun if you check it out. Check out the taco filter, highly recommend it. All right. Let's get out of here. Okay.

                                                         Great. Okay.

    Crystal King:                          I think we have a bunch of questions to answer.

    Marcus Andrews:                Let's do it.

    Crystal King:                          All right, Micheal has asked ... It looks like I can answer the first couple of these. How many people do you have to talk to before you can confidently make a persona and how much data do you need? That's a great question. I think that it depends on you and your business and the amount of information that you can gather. I would say you'd probably want to talk to several dozen customers at minimum. Definitely try to do some polling if you can. Get some industry reports. Find out as much information as you can. The more information you have, I think the more informed you're going to be. I don't think there's a hard or fast rule on that. It's really gonna be depending on the patterns that you start to see as you start talking to customers, I would imagine.

    Marcus Andrews:                Absolutely. I couldn't agree more.

    Crystal King:                          Then, Eric asked what about negative buyer personas. It would be easy to probably create some negative buyer personas but we're really looking for your ideal buyer. Negative is probably helpful in a different kind of way, I would imagine.

    Marcus Andrews:                Sprinkle it in maybe with she wouldn't do this or she definitely would do this.

    Crystal King:                          I don't think you need a specific buyer persona for somebody that you don't want. Yeah. Spend your time on who you really want to target. I would say. There's a few more here, too.

    Marcus Andrews:                Yes. There's question, can Facebook lead ads be used for B2C? Yeah. I mean, I definitely think that they're a B2B ad unit. I think there's better ... Facebook, in a lot of ways, started with their ads building for B2C.  Their video ads are very cool. Their carousel ads and canvas ads are awesome. If you're looking for really good B2C stuff, I would look in those directions. There's probably a use case for B2C with lead ads-

    Crystal King:                          I would say maybe getting people to sign up for your mailing list and things like that.

    Marcus Andrews:                Yeah. Events, if you are B2C and have an event or some kind of thing like that. Lead ads are great for that.

    Crystal King:                          Yeah. If you're crunched for time and unable to set up interviews, this is from Noah, by the way, to build a persona, what are the recommended sources to build a detailed one? Again, industry reports will be helpful. Do a survey. That would probably be the fastest think you could do. You could even whip one up in Survey Monkey, actually, and just send it out to your customers. I would say that's probably the quickest way to gather the most amount of information. You should probably be doing this as early as you can in your process. Again, if you build it into your sales process, too, it would be super helpful as their closing leads. They're asking certain questions. Also, any of your marketing campaigns can ask some of these questions, perhaps as well.

    Marcus Andrews:                Do they change over time? Is there a good time to refresh your buyer persona or update it or is it always locked in? How do you think about that?

    Crystal King:                          I would say maybe every year or so, you should probably at least take a look and see. Particularly if your product has changed. The markets change a lot. There's a lot of companies that decide, oh, you know what? Our product's evolving. The kind of customer that we need to sell to is different. Yeah, I would say maybe once a year if you're in a very fast moving industry, maybe a little bit more frequently.

    Marcus Andrews:                Cool. There's a question here from Mickey. Is Facebook forms saved in HubSpot forms so that you can create a workflow? Yes. Absolutely. It sounds like you have the paid HubSpot Marketing tools. What it will look like is it will look like someone filled out a form even though they came through the Facebook lead ad unit. It'll look like a form fill, it will have some information about the Facebook ad campaign so you can kick off a workflow and that's absolutely what you should do because then you can automate this process. If you have a lead ad that says, download our white paper or subscribe to our newsletter, you can have them automatically get an email from a workflow as soon as they submit the button, which is kind of a cool experience for the user.

    Crystal King:                          Great. Josh asked, how should the audience size of your target segments affect how much money you should spend on that particular campaign or audience?

    Marcus Andrews:                Yes. You have to start to think about your budget and how much inventory you have. There's a relationship between the two. If you have a million dollars to spend but you're only targeting five people, you're never gonna spend all of that money because Facebook can only show ads to those people so much.

    Crystal King:                          Right.

    Marcus Andrews:                Your ads are probably not gonna get served very much because they're not gonna just bombard these people with ads. If you have the inverse, if you have a million person audience but you're only spending five dollars, you're really not gonna reach that audience very well. You kind of want to say, all right, if I've got a $100  to spend and I know that I'm gonna get a certain amount of clicks, you can use that ads calculator to start to understand what your budget size should be and do some of that targeting. It's imperfect, you just want to ... you don't want to be over-targeted and you don't want to be too broad. If you've got a few thousand dollars to spend, looking at 100 thousand people audience, 200, 300, 400 thousand, you're gonna spend your budget and you're gonna be able to do it in a pretty tight time frame, but not bombard these people with your ads.

    Crystal King:                          I would say also, too, I would imagine, the great thing about Facebook is you can actually do a lot of testing for a very little amount of money.

    Marcus Andrews:                Absolutely.

    Crystal King:                          See if your ads are actually reaching the right people for $50 before you spend five thousand.

    Marcus Andrews:                Yep.

    Crystal King:                          You can always try.

    Marcus Andrews:                And you can test audiences, too. You can say, I want to try a few different audiences. I'm gonna take my five hundred dollars and I'm gonna put 100 dollars on five different audiences and I want to see which audience comes back with the best quality leads. Then you figure that out, and you say, oh, that audience is more valuable, this audience is less valuable. I'm gonna do more of what's working, less of what's not. You can continue to do that process again and again with ads and really start to increase your ROI.

    Crystal King:                          Kayley asked how are lead ads different from regular Facebook sponsored ads?

    Marcus Andrews:                Yes. Lead ads are those ad units that have the form built into it. A lot of what I see people  do is I've been getting hit with more B2B ads on Instagram's. 

    Crystal King:                          I'm seeing more of that, too, actually.

    Marcus Andrews:                I know, it's interesting. Right? All of a sudden. My friends at the sales lion, George, saw a great ad on Instagram from those guys. I clicked on it and it said, download our se-book or something to that effect. I clicked on it and it took me to their landing page. That's a typical ad flow form. You get to their landing page, you click on the button, and you get to a form that's a little janky on your phone. It's hard to fill out. Maybe you accidentally click on something and it goes off. Then you run out of patience and you click off of it. With those lead ads, you stay inside of the ad unit, you never leave, go into a mobile browser. The form is automatically populated for you so you just have to click two taps.

    Crystal King:                          Yeah, that's exciting.

    Marcus Andrews:                It's nice, that, if you want to do an offer, campaign, with ads, look at these lead ads.

    Crystal King:                          All right, I think we have time for one more question. Nora asked how many fields do you recommend using for a lead ad?

    Marcus Andrews:                That's a great question. You want to balance the information you need to get from leader lead ad campaigns with what a user is going to take. There's, I think, there's basic stuff that people are okay with sharing. Name, email address, title, name of the company, I think those are all great. You get into phone number, it's a little bit more personal information, people may be less willing to share. Get into address, your zip code. You're kind of starting to go down ... you're asking for a little too much.

    Crystal King:                          Well, it's also a lot to type in, too.

    Marcus Andrews:                Yeah. That's true. That's true. Yes, so you want to be careful with all of that, but you can get a lot of information and you can test it, too. Like you were saying earlier, you should just test to see where you're getting the best conversion rates.

    Crystal King:                          Great.

    Marcus Andrews:                Awesome. Okay. Well, cool, thank you all very much for attending. Let me make sure I promote the next session. This is day three, it's the end of day three. We've got one more day of the Four Days of Facebook.

    Crystal King:                          It's gonna be a good day tomorrow.

    Marcus Andrews:                It's gonna be a good day tomorrow-

    Crystal King:                          Yeah.

    Marcus Andrews:                ... I feel like, I'm super excited for it. Our founder of HubSpot, Darmesh, who is really passionate about bots. He built growth bots, which has been super successful. Him and Anand from Facebook, who is the lead of the messenger ecosystem over there at Facebook. They're gonna talk about bots, they're gonna talk about messenger. MK will be back to interview them. That's gonna happen at 11:30 Eastern tomorrow morning. Make sure to tune in for it and also this session, I know there's a lot in it on a customer blog tomorrow, you've got a post coming out, right?

    Crystal King:                          Yes. So if this is a lot of information to take in and you want it actually just outlined a little bit more specifically for you, there will also be templates you can use for your buyer persona. There's a video of Marcus walking you through the lead ad one more time. On the user blog tomorrow morning, it will be posted.

    Marcus Andrews:                Great. We'll make sure that there's links in there, so check that out. We'll probably follow up via email and Messenger with that blog post.

    Crystal King:                          Yeah.

    Marcus Andrews:                Thank you all for tuning in-

    Crystal King:                          Thank you.

    Marcus Andrews:                Yeah. It was great to chat. Thank you Crystal.

    Crystal King:                          Thank you.

Meet the Speakers

  • Crystal King

    Crystal King

    Social Media Professor, HubSpot Academy
  • Marcus Andrews

    Marcus Andrews

    Senior Product Marketing Manager, HubSpot

Day 4: Making Sense of the Messenger Ecosystem

September 15, 11:30am - 12pm ET

Why are messaging apps exploding? What’s the opportunity for marketers? How should you use messenger for service or events? How can you build an experience that solves for the customer? Here to answer these questions are HubSpot’s co-founder and CTO, Dharmesh Shah, joined by Facebook Messenger Product Growth Manager, Anand Arivukkarasu.


  • MK Getler:                              ... and welcome to our fourth and final day of our four days of Facebook. I don't know about you, but it has been an exceptional journey up to this point. Today, it's gonna get that much better.

                                                          During today's session, we're gonna be talking about this messaging app phenomenon. I don't know about you, but I have used not one, not two, but three different types of messaging apps today alone. [00:01:30] One of those three messaging apps that I used was actually Facebook Messenger. I'm not alone in that. In fact, there are 1.2 billion active users on Facebook today.

                                                          What's even crazier to me is that there are two billion, I didn't mess that up, two billion messages being exchanged between businesses and their customers on a monthly basis. Now, [00:02:00] that probably has something to do with the fact that 56% of people actually say that they'd rather communicate with a messaging app than with a live representative of a business.

                                                          So there's something to this. With me today are two experts, two pros in this space, to walk us through what's happening here with these messaging apps. To my far left, we have Anand Arivukkarasu. Did I say that perfectly?

    Anand A.:                                 You got it right.

    MK Getler:                              Yes. Nailed it. I rehearsed that many times before this. Anand is actually from Facebook, and he is the product growth manager for Facebook Messenger. He's had over 10 years of experience in product, dev, and growth in this technology space.

                                                          At Facebook, he now works with developers and aggregators to help them build experiences on the Messenger platform as well as actually creating the bot and AI ecosystem.

                                                          So he's also an expert on all of this. He's a speaker on AI [00:03:00] and machine learning, and tech, and data, and mobile apps, and anything else I missed there?

    Anand A.:                                 Ad tech.

    MK Getler:                              Excellent. So he's an expert on many of the things.

                                                          Also to my right here, Dharmesh Shah, the founder and CTO here at HubSpot. Dharmesh has founded multiple companies, created the startup community on He's wrote about the HubSpot culture code, a tiny little thing we carry around here, pridefully, in our hearts.

                                                          Most recently, Dharmesh has created the GrowthBot, which is a chatbot for marketing and sales or anyone that's growing a business. In addition to GrowthBot, Dharmesh is helping to lead the charge on the messaging bots for HubSpot.

                                                          So thank you both for being there today. How are you guys both doing today?

    Anand A.:                                 Great.

    Dharmesh Shah:                  Yeah, thanks for having us.

    Anand A.:                                 Thank you for having us.

    MK Getler:                              Excellent. Thank you. Honestly, the pleasure is all mine. We are all honored to be here with you.

                                                          Before I dive into some of the questions that we already have for you today for this panel, I also want to send a shout-out to everybody who's watching virtually. If you have any questions for these two powerhouses in the messaging and in the bot space, please leave your comments on the video stream as well.

                                                          I have more phone here. I'm not just looking at pictures again of my dog, Kody. I'm actually going through and crawling through messages from each one of you so I can field and send those messages and questions over to Dharmesh and Anand.

                                                          So with that, let's begin. Anand, the first question is for you. Are you ready?

    Anand A.:                                 Sure.

    MK Getler:                              Excellent. So in terms of basic communication in technology, why are these messaging apps so different that it's causing people to be so heavily attracted to them?

    Anand A.:                                 Absolutely. If you see the evolution in the last six years, if you see, mobile phone has become, from a secondary medium, has become a primary medium. With that being changed, people have moved into this new form of communication, or it's been there, but more and more using this new form of communication more extensively.

                                                          The traditional forms that are desktop-oriented, or other ways, like phone or email, are not just being the more primary ones, messaging apps have gone in there.

                                                          I see three main areas why messaging apps have become strong and why they are there right now. One, they are instant. People are just like, "Oh, I just want this message right now," and they'll immediately be able to open it and interact with it. It gives that immediacy to them.

                                                          At the same time, they are able to work on simultaneous things. With a phone call, your whole attention is being driven into the phone call, unlike messaging apps, where you can do simultaneous things. With email, there's so much more spam and scam coming in there, so the open rates have become so low there.

                                                          So compared to all that, the short form medium of messaging apps is very instant. It's expressive. There's] a lot of rich media content that you can put it in. It makes it so much more personal, so much more expressive for the audience, and they love it. We see that trend across the board, not just with millennials.

                                                          Then it's a convergent area, where we're seeing more and more integrations happening in Messenger. It's proving as a payment system. You can access transport with messaging apps these days. So it's kind of become that one-stop shop where a multitude of functionalities can be built on top of it.

                                                          All this is giving a unique advantage for these messaging apps. That's why the trend is growing and a lot of people are using it more and more.

    MK Getler:                              Interesting. So I understand how the ease of use is alluring, the native application in people's daily lives is so ingrained, honestly.

                                                          So Dharmesh, but why is it exploding now, today?

    Dharmesh Shah:                  Partly because we're just messaging each other more often. That's the primary thing.

                                                          If you look at the closest analog to messaging apps and why they're becoming popular in my mind is, if we think back to when the browser first came out, the browser applications, as they existed at the time, were rudimentary compared to your desktop apps. You didn't have drag and drop. You had this very limited set of UI widgets.

                                                          But the browser apps won for a couple of reasons. There was nothing to download. You could experiment. You could try something out. You could kind of interact with it immediately. You could navigate to any particular website. There was search. All these things were happening.

                                                          We're having the same kind of thing happening with messaging apps now, where the messaging application is an operating system. It's a container for apps to exist, and you have the same benefits. There's nothing to download. You can just say, "I want to send a message to x, y, z person or x, y, z company," and it just happens.

                                                          I'll go on a mini rant here, and I promise this will be the last one. I won't promise that it'll be the last one.

    MK Getler:                              Please don't ever stop these mini rants.

    Dharmesh Shah:                  I've been in software for decades now. For the longest time, the user interface has been primarily about pushing pixels around on a screen: [00:08:00] big screens, then smaller screens, yet smaller screens, then bigger screens again.

                                                          For the first time, with messaging apps, we have conversational UI, where instead of talking to software with touches, and swipes, and clicks, and drags, and things, you can actually take the thing that's in your head and express it, and have some semblance of hope that the software, if it was designed to do that, will do something meaningful, essentially.

                                                          We've never really had that. We'd like to believe that we build intuitive software, but there's nothing intuitive about, like, I know how [00:08:30] this application works. I'm gonna intersect my knowledge of the software with the software and then hopefully, the report of the thing that I'm trying to do gets done.

                                                          So this is, I think, the biggest shift we've seen in terms of app development as a new operating system that's bigger than mobile, in my mind.

    MK Getler:                              Fantastic. So I shouted out that stat at the very beginning of two billion messages being sent between people and between their businesses. So what are they even talking about? Two billion messages?

    Anand A.:                                 Right. It's awesome. Actually, that number is growing, from what I'm seeing. It's great because I see a spectrum of things happening in that case, like starting from, I would call, the marketing part of the spectrum, which would include something like brand engagement, the deeper engagement that you could build for your brand with Messenger, or something like getting new users to come and come to your events or come to sign up for things.

                                                          Or the next step would be something like purchase or e-commerce, or an app-like experience, into the next spectrum of things, which is traditionally under the customer service and customer care bucket, which would include sending notifications, updates, to product care, to customer loyalty building. All the spectrum is happening across the board.

                                                          Different companies are taking advantage of different features that are available to use this channel very effectively, and that's why I'm seeing a multitude of areas in which messaging is used today.

    MK Getler:                              So given that you've seen such a wide and vast array of applications of this business to consumer communication style, do you have one that's your favorite, the best communication style you've seen thus far?

    Anand A.:                                 Yeah. So there's definitely a lot of things that I've seen that are awesome. I'll give you some examples. I saw Chevrolet, was one of our good success cases, where they used this bot experience to build us an assistant DU, which will drive you to keep up with your New Year resolution. This is a very unique way.

                                                          It says, "Chevrolet will drive you," is their brand thing. Then they use that driving motivation thing into a bot assistant that'll remind you of your New Year resolution and keep you moving on day-to-day, this phenomenal engagement in that.

                                                          Or being Call of Duty bot, where there was gamers were given this inquisitive way, not just interacting and playing, but also, a popup will come up within the game, and then they have to answer that particular riddle that they're having within the game in a messaging channel, making it much more interactive, and all the way up to seeing a case from Rogers.

                                                          Or I would say the Globe Telecom case is a good example, where they're using this hybrid, human-bot experience where they're able to reduce the number of call volume that goes to their call center and then increase the customer satisfaction. I think it was 22% increase in which people are feeling that it's more personal and more expressive when customer care is done over a messaging channel.

                                                          So those are awesome, awesome favorite examples of mine.

    MK Getler:                              Excellent. Really cool. So translating that, Dharmesh, for you, for these small to medium-sized businesses, what's the business opportunity for them?

    Dharmesh Shah:                  Twofold. I think the most exciting one ... there's two buckets of use cases. The first is using messaging apps and using bots to service your customers better. So the analog here is, back in the 90's, small businesses put websites up, and the idea was to answer questions that people had about the business when the business wasn't necessarily open, like, "What hours are you open?" Common things.

                                                          What bots all you to do is essentially take that up a level, because the kinds of questions you can answer through a bot interface are just wider in scope than a website can conceivably do well. So instead of expecting the user to navigate this beautiful website they designed, which is awesome, what users really want is, "Just tell me the answer to this question."

                                                          Like, "Do you require an annual contract or not?" Or, "Do you have an office in Singapore?" Or, "Do you support Portuguese?" That's what they want to be able to do. That's why they're there. So that's the first bucket of use cases, just serving your customers better, making that initial connection.

                                                          The second bucket of use cases for businesses around internal tools, can you make yourselves more efficient in terms of CRM applications, HR things, like, "Oh, a new employee just joined. Do you want to enroll in our 401(k) plan?" That should just be a message that pops up, and you can click on a button, and it just happens, versus going through a set of forms or some web application.

                                                          So those are the two primary buckets. I'm more excited about the first one in terms of how they can reach their customers and how they can service customers better. We have problems in customer service generally, overall, as a society, and the nice thing about software ... It may not necessarily understand you all the time. It's early. But software doesn't get angry. Software actually has more compassion, even when it's really busy, as it turns out. I'm pro-bot, as it turns out.

    MK Getler:                              I was like, what a coincidence.

    Dharmesh Shah:                  Not that humans are bad. But I'm just saying that for certain kinds of things ...

    MK Getler:                              Again, if anybody has any questions for Anand and Dharmesh, please fire them through the comments below.

                                                          But thinking about that customer experience and thinking actually about curating a specific experience for those that are interacting with your Messenger, how can a business start to think about that experience [00:14:00] and curating that experience?

                                                          I'll open up this to both of you. Feel free.

    Anand A.:                                 Yeah, absolutely. I think it's a really good question. I would see, again, it's a range of things. They can start with basically opening up the channel and starting doing that manual reply, to all the way to use our smart replies and then set up instant replies also, like have a good welcome message. All this can be arranged.

                                                          Or if they really are thinking about, "Hey, there's a lot more people are coming in [00:14:30] and interacting. I want to scale this platform up. What can I do right now?" So that's when they can either connect them to the live support agent, the entire customer service channel, or they can also automate it and then go all the way into the layer of AI, NLP is getting better and better as we see it.

                                                          I'm surprised when I'm interacting with some of these bots where I see a line, like, wow, this is amazing that the bot is able to take all their FAQs, which traditionally, you have to go to scan into that website and look at different things, it's able to give me as a human reply, and that's amazing. Companies have gotten into that stage where they're using NLP, AI, to build that great experience.

                                                          So it depends on the stage and the scale that you want on this platform, and I believe the world will be going into this case where you will be able to provide that one-to-one communication at scale with, as Dharmesh mentioned, the bots and the Messenger experience.

    MK Getler:                              What about for you, Dharmesh?

    Dharmesh Shah:                  Yeah, so it's interesting. I think most of the bots that start now should start with a relatively simple UI. I think people get caught up in the NLPs, like, "Oh, we can have all this AI and all this stuff," which you should totally do. I'm a big believer in it.

                                                          But essentially, from the beginning, even if you do a guided conversation that says, "Okay, well, here's the use case that we're trying to solve for. Do you have a question about this, this, or that?" So it's almost choose your own adventure, and it kind of reveals information in a much more structured way versus the blank canvas syndrome of, oh, you're interacting with this bot. Type in the thing that you want.

                                                          So I think along that spectrum, I think organizations will move further and further towards the, "Just let them ask what they want to ask, or let them make the requests that they want to make the request." I think we're still in the early stages of bot development for that, especially for small and medium-sized businesses. I don't think the world is yet ready, and the technology is still coming along.

    MK Getler:                              Gotcha. So keep it simple to start off with. Keep it simple and also keep it human.

    Dharmesh Shah:                  Yep.

    MK Getler:                              Excellent. Cool. So we do have some questions from our audience members. This question is, "Will messaging become a part of other apps and websites beyond [00:16:30] Facebook Messenger itself?" So I'll give that to you, Anand, first, and then Dharmesh, you can go ahead and respond as well.

    Anand A.:                                 Yeah, absolutely. I think that's what we are seeing overall. Right? One, there is a whole concept of app fatigue that's happening. I don't want to download an app for a conference that I'm going. I don't want to download an app for a retailer. It's just too much to ask for the end consumer to do that.

                                                          That's why messaging platforms are going into this stage where it is becoming a one-stop solution for many of these areas. A lot of micro-app experience, micro-web experience, can be built on top of them. Then, as I said, mentioned earlier, payments is gonna be integrated in there. Much smoother notification methods can be integrated in that, and top of it, we are creating this layer of NLP, which will help you to go a little bit more advanced and scale it up.

                                                          Definitely, the messaging apps will be going beyond where [00:17:30] they are right now.

    MK Getler:                              Nice.

    Dharmesh Shah:                  This notion of UI and platforms overall, it's interesting. In the mobile world, there was debate raging when mobile first started. Is it gonna be native apps, or is it gonna be mobile web, apps that sit within a mobile browser? As it turns out, native apps won, and the reason is because they were closer to the OS and it felt more natural versus things that were inside of a mobile web browser, at the time. So native apps won.

                                                          In the messaging world, I think what we're gonna see is a similar development. But I think the overall ... there's a difference now, 'cause when you're typing a message into Facebook Messenger or another messaging app, it's basically native. There's no difference between that experience versus what you would have anywhere else.

                                                          But the advantage of not having to download something is massive. I cannot overstate the ... Just the fact that the user does not have to say, "I'm gonna download this." They don't do it anymore. All the statistics show ... It's not that we can't ask them. It's like, even if we did ask them, they wouldn't do it.

                                                          So messaging now gives us this opportunity to have a low-friction, let people try something else, similar to what we had with the web, where the mobile web I had hoped would've wound up. But this gives another bite at the apple. Can we create something low-friction that provides functionality and it's very easy to get to, very easy to try because there's no download? Then I have to remember to delete or whatever. It's like, try it out, just like a website.

                                                          But it doesn't feel awkward compared to the native apps. Your dog won't bark at it because it's like, oh, well, native versus mobile. It's like, I know it's not right. Something's wrong. But, yeah.

    MK Getler:                              Gotcha. So thinking very similar along those lines about the ease of use, about the responsiveness of this, I can't imagine that customer service is going to not be leaning into these platforms as well. So how would apps be used for customer service? Maybe you can give us a couple of examples, Dharmesh.

    Dharmesh Shah:                  The simple ones are around, "I have a question about X." It's like this is the FAQ on steroids model, which is the early pass at some of these things. But over time, I think what ends up happening is that there's only so much you can put in an FAQ. There's a limit to how much you can do that.

                                                          A customer wants to not just get questions answered, but they want to conduct mini transactions. It's like, "I want to change my billing date to be the end of the month instead of ... whatever. Charge me per hour." Whatever it is.

                                                          So I think what's going to happen from customer support's perspective is that people will initially have this blank canvas, like I don't know what this thing supports or what it allows you to do. I'm seeing this with my own bot development now. Over time, people will just try things. Just like, "I'm not sure if this customer support bot supports this particular thing. I'll just type it in, and I'll try it and see what happens."

                                                          The nice thing there is, from a product management perspective, bot developers can say, "Oh, just look at the logs every night, and that's your product management." See what people want to be able to do, what kinds of questions do they want to be able to answer through the bot, and then go support those things, essentially. That's like the focus group to end all focus groups, is just directly look at customer behavior, essentially.

    MK Getler:                              Interesting. So you did mention that bot that you're working on, GrowthBot. Can you give us a little bit more insight about what's going on there?

    Dharmesh Shah:                  Wow. I was diabolically clever there. I didn't even notice that.

    MK Getler:                              Go figure.

    Dharmesh Shah:                  Amazing. Yes. I've been working on a bot called GrowthBot, and it's a digital assistant or a bot for marketing and salespeople that's been out for about a year now, coming out of our HubSpot labs project. The quick idea here is that if you could hire an intern or a very clueful associate ... let's say you could hire me. I'm cheap, by the way. You can do that. Reach out.

                                                          You can ask questions like, "Oh, I want to know what my traffic was last month versus the prior year," or, "How's our PPC campaign doing?" Or, "I'm looking up this competitor. I want to know who their other competitors are. Have they increased their overall spend on this versus that?" Those kinds of questions that you might have for a very clueful marketing intern, GrowthBot tries to answer a roughly wide range, some would say a crazy wide range, of things that it tries to do.

                                                          It's been fascinating to see the overall response. Responses fall into the three buckets. One is what I call the random internet, profanity, marriage proposals, and GrowthBot gets those all the time. I think, oddly though, more marriage proposals than profanity outright. Maybe there's hope for humanity yet. I don't know.

                                                          But then there's the things that it's like, oh, well, the GrowthBot supports it. It's already there, and GrowthBot will come up with an answer. But the third bucket is the one that's the most interesting. It's like someone asked GrowthBot something, and it's like, well, GrowthBot doesn't support it today, but that was a reasonable thing to ask. So then I go stub it out, and say, "Oh, this feature's coming soon."

                                                          Based on how repeatedly that thing happens, that basically is a roadmap for the next week, or month, or quarter, and then we go do those things. That's how it's grown into what it is today. It's been popular. It's been lots of fun to work on.

    MK Getler:                              It's been fun to actually interact with as well. So thank you.

    Dharmesh Shah:                  Thank you.

    MK Getler:                              So Anand, there has to be a time and a place where the bot shouldn't be the one responding and you, yourself, as the human, should be the one to respond. What is that pivot moment?

    Anand A.:                                 Right. I think that's a very important question. I think that's one thing that the businesses should understand. At the end of the day, the experience that the customer gets in the bot is very important. They are waiting, they're standing, they want to get a particular answer.

                                                          If there is a clear signal, you would get a lot of clear signals upfront saying, "Hey, the customer is asking for this, and this, and this," and able to build an automated experience, which understands that, saying, "This is not covered in the list of things that I'm gonna say," and then seamlessly move them into that live chat agent is very important, because at the end of the day, customers are there, coming to find a value or an information out of there.

                                                          Then not giving them what they are asking for, you don't want this to be an IVR kind of a channel, where I'm going there, I'm pressing one, two, and waiting, listening to this crappy music, and getting annoyed about it. You don't want that to be this experience. That's why it's an open platform, where you can design that experience in which you know the customer cannot be satisfied with the automated list of content and subjects that you have, seamlessly transfer them into a customer care agent.

                                                          We built a couple of things, including something called Handover Protocol, which will enable you to transfer seamlessly between the automated experience to the live chat agent.

    Dharmesh Shah:                  I have something controversial to add.

    MK Getler:                              Please drop it. I love it.

    Dharmesh Shah:                  I have this on my list of rants. The first one is around I don't believe ... This is controversial. I don't believe software should pretend to be human. So if you're talking to a bot, or you're developing a bot, it's okay to give it a name. You should. It's okay to give it a personality. You should. But you shouldn't fake trying to be human for a couple of reasons.

                                                          One, it's disingenuous. We'll put that aside for a little bit. But number two is you're basically setting yourself up for failure, because it's just a matter of time. The technology is not to the point where people are not going to figure out within a roughly short period of time that this thing's not human, and then they're gonna be disappointed.

                                                          But if you just open it up with, "This is a bot," they've already lowered their bar. [00:24:30] It's like, fine. Then they'll ask things in a way that they think a bot will respond to, which is good. So that's rant number one.

                                                          Rant number two is less controversial. I want bot developers, when you do assign names to bots, because these are digital assistants, essentially. That's the other name for them. Stop naming them female things. We have Alexa. We have Siri. We have Cortana. So if I had named GrowthBot a name, I would've given it Gary, not Gabrielle or something like that.

                                                          Anyway, that's a personal ... It's 2017. We can fix some stereotypes before they start, essentially.

    MK Getler:                              I love that. I love that. So what you're saying is, first off, let's stick with maybe some gender neutral names moving forward, for our bots. But also, bots and AI are not going to replace us when it comes to working with our customers.

    Dharmesh Shah:                  We can't get autocorrect right yet. So when that happens, then fine. Maybe the robot overlords will take over. But, yeah.

    MK Getler:                              Fantastic. Well, so I'm bought in. Yes. I want a bot. I'm a small to medium-sized business owner. Should I go build a bot by myself? Should I buy one? Should I hire someone? I'm assuming you can't buy these at Target. How do I get one? How do I get my hands on a bot?

    Dharmesh Shah:                  Well, first thing is I think you should use a bot. Use lots of bots just to see what's out there. Every business there is, you're a customer of other people. Try some bots out so you get a sense for what bots make available.

                                                          Then don't jump in. It's like just don't build a bot to build a bot. This is what happened in the mobile world. It's like, "Well, everybody has a mobile app, so I'm gonna build a mobile app." Well, it turns out your yogurt company might not need a mobile app in the time.

                                                          Once you discover, it's like, okay, I've seen these kinds of bots. But here's how those kinds of bots may apply to my industry, to my business. Then find a way, based on what your resources are. There are platforms out there now to make bot-building easier. If you're really, really committed, you can kind of build a team and start doing it.

                                                          But step one is learn about bots. Use them instead of just jumping in. Step two, don't do it just because it's the new technology wave right now. Do it because you're trying to solve a business problem.

    MK Getler:                              Nice. Do it with a purpose.

    Dharmesh Shah:                  Yeah.

    Anand A.:                                 I do agree totally on that point. Businesses jump onto something just because it is there without understanding what business problem they want to solve, what end value that they're gonna create to the customer. That's super important for them to understand.

                                                          In many cases, I believe that bot may not be the solution at all. It can be just human replies. As long as your volume is in that certain way, keep it human. Keep it like it's a channel to be used, both on the live chat side, and also, if you want to automate, if you have the scale, and if you want to showcase all those things, that's fine.

                                                          But create a really good value proposition and a compelling experience. Otherwise, it's like, oh, I'm creating an app. I'm creating a thing. It doesn't have a direction. Also, have a clear objective as well because this not like another fad that you want to do, but it's more like an impact that you want to create for your customer.

    MK Getler:                              Interesting. So if you are going to be an early adopter, do it with a purpose. Do it with a vision, and add value to the consumer's lives.

                                                          So what kind of future do people need to get ready for when it comes in terms? So you are an early adopter, but you're trying to get in front of the curve here. What future do we need to be prepared for?

    Anand A.:                                 Yeah, absolutely. So there are a couple of things I say, right? As I mentioned, part of the Messenger experience, I would call it, is a place where you should start thinking about building personal, one-to-one communication at scale. Think about it in which it's a new channel where you want to create most value to the customer and build this channel as a long-term re-engagement channel.

                                                          That's one thing that you should be thinking about as you're building your audience in there. It's not like I'm getting this customer. Okay, I'm gonna maximize the output of them in the next two or three days and leave it there. It's about thinking about long term. You're bringing in your customer into this channel.

                                                          How do we build that long-term commitment with them? How do you re-engage with them? That's really happening phenomenally well with this kind of an ecosystem. Think about that.

                                                          The second area I would think about also is how do you adopt to scale? What are the set of tools that you can use? There's a set of things that you can already use, something like Instant Reply and all that, that's already built into the tools.

                                                          Also, think about the features that are needed, for example, payments. Payments are being opened up. How can you make the customer life much easier? People love shipping notifications because you want to track your order when it's coming up. How can you use the things that people love and then give that value to them?

                                                          So those are the areas that they should start looking at and probably be on top of the curve.

    MK Getler:                              What do you see in the future there, Dharmesh?

    Dharmesh Shah:                  A couple of quick things. One is I think we're gonna see a lot more multiparty chat. So imagine you're working with a customer in the B2B space or whatever. You have the customer, you have your internal salesperson, you have the sales engineer people that know something, and you have a bot or [00:29:30] maybe two that are all collaborating in a new channel that's spun up particularly for this project or for this opportunity. So I think that's going to happen.

                                                          The other weird thing I'll say that I think is going to happen in the future is we're gonna have much more bot-to-bot communication, bots knowing about other bots. It's like, "Oh, I need to get an answer to this thing." So it's like the HubSpot bot will go off and talk to the Shopify bot and say, "Oh, well, they have this question. In order to fulfill the request the user had, I need to do this."

                                                          So in my mind, it's the future of APIs. APIs have been around to kind of expose systems. What it's gonna turn out to be though is that more and more organizations will expose data and services through this messaging interface, and that sometimes will be the most convenient way to get stuff because the APIs haven't been built yet, but the messaging interfaces have. So that's bots talking to each other to get things done.

    MK Getler:                              That is actually super fantastic, evolving the technology that we have currently day in and day out.

                                                          Well, thank you both so much for this stellar insight into today and where we're going in the future of messaging and bots, as well. Any final thoughts? Any final takeaway we should have for our virtual audience members?

    Dharmesh Shah:                  I have one closing thought.

    MK Getler:                              Hit me.

    Dharmesh Shah:                  We gave you the 1.2 billion number on Facebook Messenger, which was announced at F8. So yesterday, they updated that number to 1.3 billion, which is a bigger number than 1.2 billion. But the thing that's fascinating to me, that means 100 new people in the last few months started using Facebook messenger. That's huge.

                                                          If you're ignoring Facebook Messenger overall, you're missing out on one of the biggest movements we've seen in technology, in my mind.

    MK Getler:                              Fantastic. Get on the train, people.

    Dharmesh Shah:                  Get on the train.

    MK Getler:                              What about for you, Anand?

    Anand A.:                                 Absolutely. I'm totally on board with that. We're seeing a lot of initial successes on the platform. It's definitely an area that you should start thinking about. Not just thinking, go and experiment and talk to your customers through the channel. Build that personal experience, and start using that edge that you're gonna have with this channel. That's my call.

    MK Getler:                              Nice. Excellent. Thank you.

    Dharmesh Shah:                  Thank you. Thanks for having us.

    Anand A.:                                 Thank you. Thank you for having us.

    MK Getler:                              Thanks to everybody who has dialed in virtually for our Four Days of Facebook, and I'm sad to say, that's a wrap. That's the final of the four days.

                                                          I just want to quickly end by saying thank you to all of our speakers, not just those that are here physically with me, but those that joined us on Tuesday, Wednesday, and Thursday. Thank you to all of you who took some time out of your day to join us virtually.

                                                          Thanks to Facebook, obviously, for joining us for this event. Shout-out to everyone who made this event possible, from the technology, from coordinating people like Anand to come visit us here at HubSpot.

                                                          But just because these four days are over, doesn't mean that the fun has to end. I'm gonna be sending you some recap emails of some of the tools, some of the tactics that we've talked about this week. We want to hear from you. We want to know what you've been able to do with all of the knowledge that we've imparted to you in these past four days.

                                                          All of the content that we've shared in these past four days, too, will be hosted at Thank you, and again, keep an out for partnerships between HubSpot and Facebook moving forward.

Meet the Speakers

  • Dharmesh Shah

    Dharmesh Shah

    Founder/CTO, HubSpot
  • Anand Arivukkarasu


    Product Growth Manager, Facebook