Hunch.com: Training the Web to be Your Friend

Would you like the Web to understand your inner GPS?

Then go to Hunch.com and start training the Web. After answering a few questions about your tastes, preferences, and opinions, you will get a set of recommendations from other Hunch members for movies, books, restaurants, recipes, music, vacation spots, shops, gadgets and other goodies.  Hunch will also predict how well you’ll like each recommendation. The more you interact with the site (e.g., answering more questions, rating Hunch’s predictions, adding a descriptive tag, making a recommendation), the more accurate and relevant its recommendations become.

In the wide range of topics it covers, Hunch.com goes beyond topic-specific recommendations offered by sites like Netflix and Amazon. More significantly, its recommendations are based on understanding of who you are, not just what you bought or liked in the past. Hunch predicts what you would like with the help of its vast index of “affinities”—the connections between a person and any item on the Web (another person, a product, a page) as measured by likes, dislikes, ratings, check-ins, and other online interactions. This index of the Web, or its ”taste graph” as Hunch calls it, is continuously refined  with input from members of the Hunch community, their social networking activities, and applications using the Hunch Open API.

Hunch also advances the state of “social search” (see Jason Kincaid’s take on its current limitations), improving the function of social networks as information filters.  Some of your closest friends may have very different tastes than yours, and a friend who has a similar taste in books, may have very different taste in food. Over time, Hunch understands the nuances of the affinity between you and the people in your social circle and adjusts its recommendations and predictive ratings accordingly.

There’s no advertising on Hunch. Some of its recommendation pages link to external sites where you can purchase the product or service that Hunch suggests, resulting in the merchant paying a referral fee to Hunch. But the presence of a link to a retailer has no effect on Hunch’s recommendations which makes sense given its mission to accurately map the web of Web connections.  Another way Hunch makes money is by helping partner sites improve the quality of their own recommendations. Hunch says it never exposes users’ input without their permission, and by default, users’ answers to Hunch’s questions are set to “private” in their profile.

So far, Hunch has documented more than 30 billion connections between people and items on the Web. Its correlations-obsessed team can predict whether you are for or against the death penalty based on your skill in handling chopsticks or guess how you drive based on whether you save your receipts or leave them with the cashier.  A fun way to get a feel for Hunch is to play the Hunch correlations game, correlations do the darndest thing.  Play, follow, lead, recommend, interact —use Hunch to train the Web to be your friend.

In the Age of Big Cloud Data (ABCD), you want to be in control of your Web profile.

In the Age of Big Cloud Data (ABCD), successful companies will go beyond search and social to discovery and from business models based solely on advertising to selling deep and predictive analysis of what makes us tick.

Advertisements

About GilPress

I launched the Big Data conversation; writing, research, marketing services; https://whatsthebigdata.com/ & http://infostory.com/
This entry was posted in Analysis, Business Impact, hunch.com, Recommendations, Search, Social search. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s