July 06, 2010

False Positives? I Blame Reverse Scientific Method.

I guess one of the best times to make a public mistake on the blog is at 2 a.m. on a 3-day weekend. Or, in theory, given the absence of hard news, it could instead be one of the worst times, guaranteeing the blunder will fill the vacuum. I got to test the theory this Monday, when I pieced together a number of theoretically unrelated updates and manufactured speculation that two of my favorite and most-watched Web companies, Foursquare and Brizzly, were hooking up.

By morning, both companies denied it outright, but other news outlets wrote it up as a clear miss - which was their right, considering how high profile it could have been if I had been correct.

Fellow blogger Denton Gentry saw my new age sleuthing as registering a false positive, as I made something out of nothing. I've got a history of watching corners of the Web and referral logs for story ideas or new applications, so this route wasn't too much of a change - even if the outcome left much to be desired.

As I see it, the fault for this story had as much to do with reverse scientific method as anything. If you remember scientific method from your grade school science fair projects, the right way to prove a theory is to set up a hypothesis, collect data, analyze it and interpret the results, forming a conclusion. But many a time, once a hypothesis is made, one can see clues everywhere, in fact self-selecting the data. Once I had a buzz in my brain about Brizzly and Foursquare being more than "just friends", I kept digging until I found enough clues to tip it from theoretical to probable.

The combination of @VCMike's update about one startup receiving an offer and that of @cw signing legal papers tipped me over the edge - as at 2 a.m. I was looking for the "one more thing" that could morph a fever dream into a serious scoop. And we still don't know what Mike was referring to - or Chris for that matter, though it could be leftover work from the Plinky sale to Auttomatic.

I tend not to post speculation here. Most of the time, you see my analysis of what I am using and hearing, product news or observations. But to explain the motives behind this, for those who aren't part-time geek tech bloggers, you need only look back at how the news that FriendFeed was acquired by Facebook snuck up on us. I had had my own inklings something major was up when co-founder Paul Buchheit rolled in happy after a late night. Instead of the site's loyalists getting a fair shake at the news, the story debuted on a blog that had been very negative to the site in the past. Similarly, news of Brizzly's own Picnics feature (which I like a lot) had muted response, thanks to what I saw as a half-hearted outreach plan.

Those two contributing pieces, in addition to my unanswered hypotheses (at the time) thrown to @shellen and @cw, made me feel I could take the risk of getting ahead of a story even if I didn't have all the data signed and sealed. I've seen worse conjecture with fewer clues on other sites before, and had anecdotal data from third parties that helped me feel I had the story right - late at night when most people were turning in after an evening of fireworks.

So, yes, I got the story wrong - and many people couldn't see how I thought the two teams might leverage one another. I could see it ... with a new Web and iPhone client dedicated to Foursquare... a tie-in of Foursquare tips to the Brizzly guide, and somehow... Picnics to talk about venues. It could happen. I've seen worse fits. So for this round, I was wrong, but I'm not totally repentant. I don't mind being wrong every once in a while, having tried and missed, than having not tried at all.

See some of the ensuing fallout:
The Next Web: Could Foursquare be looking to acquire Brizzly? [Updated: No]
VentureBeat:  Thing Labs CEO: No Brizzly buy for Foursquare
Business Insider: Brizzly CEO Shoots Down Foursquare Acq-Hire Speculation
GigaOM: No! Foursquare Isn’t Buying Brizzly

No comments:

Post a Comment