There is a megatrend on the Web today, and it’s called personalization. I didn’t say a “new” trend: Amazon, after all, has provided you with remarkably accurate personalized product recommendations for years. But it’s a bigger trend today now that more crunching power is cheaply available, more and more of our preferences and behaviors become susceptible to reasonable approximations by algorithm, and more smart entrepreneurs move to take advantage of it.
My quick scan of the TechCrunch headlines today provided, at a glance, three examples of hyper-personalization (links point to the TC articles, by three different writers – what a story if an editor had thought about weaving them together!):
- My6Sense, an RSS content filtering tool that resulted in an “a-ha moment” for the reviewer when absolutely relevant posts floated to the top of his iPhone screen, without the user having had to do anything else (no ratings, no preferences) than using the app as a reader for a couple of days.
- BeeTV, a personal TV recommendation system from the brains of Gavin Potter, of Netflix Prize competition fame;
- Covet.com, a clothing and accessories recommendations engine tailored to your style.
This trend will brilliantly simplify our lives if it helps us save time that we waste today. If you don’t like browsing through clothes, surfing through channels, scanning your RSS feed reader, flipping through bookshelves, reading movie reviews, turning the pages in a recipe book, choosing a toy for your nephew, researching holiday destinations and so on, then recommendation algorithms will solve the problem for you: voilà, you don’t know it yet but this will become your favorite TV show. If you don’t like dating and want to be in a serious relationship from day one, there are matching algorithms that will find you a compatible partner for life. And personalized medicine holds – of course – huge promise.
But we also need mechanisms for serendipitous discovery; for stretching one’s boundaries; for challenging one’s opinions; and for getting out of our comfort zone. (If schooling were organized by personalized preferences, how many people would ever get any basic algebra?) The personalized universe freezes us in time. It narrows our horizons. If not executed with a fondness for adjacencies and the odd curveball, it will let us dig ourselves into a deep tunnel of 1970s progressive rock, if that’s where we start from, and never even discover 1990s grunge. It will keep suggesting backpacker hostels when we can afford four-star hotels, or four-star hotels when we can only afford backpacker hostels. It will make us into Burgundy experts, while we don’t know we might enjoy Bordeauxs better. It will reinforce us in our particular religious and political bias. It will perpetuate our teenage Ayn Rand infatuation.
Social media may come to a partial rescue of the algorithm: you can follow a friend’s recommendation for Industrial music if all you know is Alternative. But you must still have become friends with that person – or “social media friends”, if you’ve never met in person – on the basis of some shared worldview. A social media recommendation mechanism to open up our horizons would refer back to Granovetter’s “strength of weak ties” and perhaps upweigh the tastes and preferences of our weaker connections, so that we may discover and learn something new.
Personalized recommendations are undoubtedly efficient. If you want your life to become more efficient, you will use hyper-personalization to minimize the drudgery and get to a good enough solution quickly. But if at times you enjoy discovery, you like being challenged, you want to try something different – you will step out of your personalized universe and explore someone else’s, or create one that does not exist.