Hunch, the Web site that asks users questions in order to help them answer their own, is going through a minor identity crises.Last week, cofounder Caterina Fake announced that Hunch would no longer work for anybody but registered users.
Caterina said that because of this change “traffic will plummet.”
We think Caterina and her cofounder, famed angel investor Chris Dixon, should take this moment to reevaluate what they have in Hunch and do something different with it.
What’s so impressive about Hunch is its machine learning. Asking users just 20 or so questions – like “Can you do 10 pull-ups?” or “Do you spoon? Or are you spooned?” – Hunch’s algorithm, built by a bunch of MIT grads, can tell where a user is from, their political and sexual orientation, their gender, their age, their race, what kind of car they drive, magazines they read, and on and on.
Right now, Hunch asks users these questions and then uses that data to serve better results in a shopping search engine.
We think this is a misuse of Hunch’s amazing technology. Are people really looking for new ways to search for what to buy? Don’t they just go to Google or Amazon for that?
Compete.com, which underestimates traffic but remains directional, confirms we might be on to something. It says traffic to Hunch.com peaked in March and then declined in April and May.
So, if Hunch shouldn’t be another shopping search engine, what should it be?
Hunch’s business should be helping Web publishers serve incredibly well-targeted display advertising.
To do this, Hunch needs to get as much information about as many Internet users as possible – and install cookies in all their browsers.
There are several ways Hunch could do that. With its ridiculous ability to ability to know seemingly everything about its users, we’d like to see Hunch create a free version of eHarmony.
But Hunch doesn’t have to do something so complicated as that: It could just offer players of Facebook games currency in exchange for answering a few questions (and accepting a cookie.)
If Hunch can cookie enough people and prove its ability to target ads, you can bet Facebook, Google, and maybe even Apple will come calling with offers they’d never think about making for yet-another-shopping-search-engine.