I’ve noticed that many new startups try to include an element of serendipity in their products. If you look at Wikipedia, the current definition of serendipity is “a happy accident or pleasant surprise; specifically, the accident of finding something good or useful without looking for it”. My good friend Lenny Ratchitsky did a great talk at TEDxCondordia entitled “Losing Serendipity” which explains how serendipity is important in our daily lives, and how he integrated elements of discovery and chance in Localmind (you can view his presentation on YouTube here).
With this in mind, there are two different approaches you can take with a product:
- The Serendipity approach, in which you can surface useful information that your users are not expecting. A good example of this is Foursquare’s Explore feature. It shows you interesting locations around you ranked by your friends checkins and tips. Using explore, you can discover an interesting restaurant or cafe based on recommendations.
- The Personalization approach. Using this approach, you can profile the user and determine his/her likes and dislikes. Using these signals, you then provide information that is the most contextually appropriate for the user. Hunch (which was recently acquired by Ebay) is an example of a recommendation engine which statistically looks at your likes/dislikes and provides tailored recommendations. Another example, Facebook made changes to their News Feed a while back to display only “Friends and pages you interact with most” (which was partially rectified with the recent Timeline update).
What approach should you take? Take your pick. The missing element is profiling the user’s affinity for risk or change. Some people enjoy having the same breakfast every morning, others actively seek out new places to eat. The winning formula in my mind is a system that looks at behavior, then dial up serendipity or personalization on a user by user basis.
