Way back in 2001, Seth Godin offered me some good advice. I'd added a pitch to Chowhound's front page to sell our newsletters, and Seth suggested we try a range of different wordings (and fonts, etc.), and then track how well each "converts" (i.e. persuades people to actually buy), eventually determining the optimal recipe.
It's a brute force method useful only when there's a large enough flow of people to create statistically significant results - and when there's an effective and inexpensive means for tracking their behavior. This makes it ideally suited to high-volume web sites. But while I appreciated the idea's ingenuity, I opted not to try it, for various reasons (foremost: I needed to focus ever-decreasing time on either improving the site itself or on improving our business, and, not being a businessman, I chose the former...hoping, perhaps naively, that a great, useful site with lots of traffic might eventually be of interest to someone).
Twelve years later, the Obama campaign used the same approach, and is being roundly congratulated for its cutting-edge geeky cleverness (note: they were, indeed, very, very clever; I'm not trying to diminish that).
Interestingly, Reed Harper, the guy running this aspect of the campaign, started out as Godin's intern!
Is there any difference between this and what direct mail advertisers have been testing and using for 50 years (and more)?
ReplyDeleteYes, much faster feedback and much lower cost.
ReplyDeleteThe combo of those two allows extremely subtle fine tuning, and trial of counter-intuitive iterations (e.g. misspellings, ugly design, etc).