A couple of weeks ago, I had a conversation with Michael Arrington (of TechCrunch fame) on Twitter. The first time I’ve ever conversed with him, in any form. The conversation took place at 2:33 a.m. (ET) at the tail end of a long weekend (Jan 20th, 2009).
The conversation was about HubSpot’s latest free tool, Facebook Grader. That fateful day, we had made Facebook Grader publicly available for some “limited” testing. We told a few friends and colleagues, tweeted about it a bit. No announcement, no press release. We just needed some users and I was too lazy to write the “alpha invitation code” feature so we could control how many users actually tried it. The tool was very, very early — even by my standards. But, we needed some testing so we could make it better, so we put it out there. As fate would have it, Michael tried out Facebook Grader night and tried it out that night and wrote about it in an article. The feedback was, to say the least, not positive.
So, I tweeted about Michael’s article:
Me (@grader): @techcrunch bashes Facebook Grader tonight calling it beyond useless. [sigh]. Lots of work to do.
I wasn’t expecting Michael to read my humble tweet, let alone respond to it. But, respond he did. At 11:33 p.m. (his time).
@techcrunch: I’ll give you $100 if you just delete it from the Internet. Or bump my score to 100. you pick.
So, on a lark, I decided to stick my neck out there a bit and came back with:
me (@grader): I’ll make you a deal. I’ll donate $1,000 to your favorite charity if you give me 2 weeks, retry it and still think its useless.
Michael took me up on the offer. We struck a deal.
That was 2 weeks ago today. Time’s up on the Arrington challenge. We’ve been working away furiously on Facebook Grader since then. Here are some of the things that have been done in that time:
1. Grading Facebook Business Pages: In addition to grading Facebook users we now also grade Facebook pages. Facebook Pages (or “Fan Pages”) as they’re called, are mostly setup by businesses or other organizations that want to build a community/following on Facebook. Grading these business pages addressed one of the biggest complaints we got about Facebook: “Grading people/users doesn’t make a lot of sense, because it ends up being a popularity contest and it shouldn’t be about how many friends you have.” Yada-yada-yada. I still think that grading users/people is just fine (done correctly), but it’s clearly useful to be able to grade Facebook Pages for a business, and try to figure out how successful they are (compared to other pages) based on how many fans there are — and what the power/influence/reach of those fans is. I think few would refute that a fan page with more fans (and more connected fans) should get a higher grade.
2. Business Page Search Feature: It’s now possible to search pages by keyword and see a list of matching pages in rank order. So, for example, you could search for “mortgage” and find the highest ranking business pages with that keyword. Still crude, but at least semi-useful.
3. List of “Top Friends”: Now, when a user grades themselves, the application shows a list of their top friends, based on their Facebook Grade. As it turns out, Michael Arrington is my #1 friend.
4. Improvements To The User Grading Algorithm: The first pass at the algorithm wasn’t working as intended. We weren’t properly crediting the 2nd order network of friends as intended. We have improved this considerably and the database itself has grown, so the grades are now much more accurate. Coincidentally (and I do mean coincidentally), Michael now actually gets a grade of 100/100 instead of the 90/100 he got when he first tried the app. He is also at the top of the Facebook Elite page. Further anecdotal evidence that things are working better is that folks like Loic Le Meur, Pete Cashmore and Mari Smith are also on the list.
Now, here’s my dilemma. People are going to think that I coerced the
algorithm and adjusted it to give Michael a high grade just to make him happy
and so I wouldn’t have to pay the $1,000. This isn’t it at all. We simply
fixed a bug, and the database of users is bigger now, and the combination is
delivering better results. I was surprised as anyone with the outcome. Basically, Michael has a lot of connections on Facebook, and on a relative basis, those friends are better connected.
Message to Michael: Here’s what I’d like to do: I go ahead and write that check for a $1,000 on your behalf to a not-for-profit of your choice — no strings attached, regardless. You stop by and check out Facebook Grader and assess whether the application is still “beyond useless” or if it’s gotten a wee bit more useful. Candidly, we’re not quite where we want to be yet (some of it limited by what’s currently possible with the Facebook Connect API). Regardless, thanks for taking the time to review Facebook Grader. There’s nothing like a bit of a challenge to focus energy and get things done. Let me know what organization you want the money sent to. Hope all is well with you.