September 07, 2010

The Future of Search: Personal, Persistent and Implicit

The world of search is changing before our eyes - from one where there is one right answer for all people at a specific instant to another where an answer to one's query can differ based on who they are, their social graph, their location, or even the time of day. No matter how fantastic the algorithms, combined with measured human editing, the assumption that there are "correct" responses to single word queries or short phrases for all is practically gone. That's why users are increasingly using multiple search terms in their queries, and modifying the searches until they find a satisfactory result. Pile on the conundrum of how to approach the world of real-time, social feedback, and geolocation, and a simple If A then B problem starts to look a lot more like calculus.

At the IFA home electronics event in Berlin, Google CEO Eric Schmidt was widely quoted following his remarks where he said the future of search is well beyond the Web, but through all your information, and done automatically, as you "walk down the street", the example saying that the smartphone would be "doing searches constantly."

While some may think this is Google over-reaching its boundaries and extending its view into every step of our waking lives, it's not evil that is seen here, but instead evolution, the same evolution that is the future of products like my6sense.

Stated directly:
  • The traditional search market is powered by one time explicit search.
  • The future search market is powered by persistent implicit search.
What do I mean? An explicit search is one where you are clearly searching for something specific and you are taking that action. It is often one done for a specific occasion or time. Today, I may search for "movie times Inception Sunnyvale" to find when the film is playing in my area. I may never search for that again.

A persistent explicit search is one where the query is known and unchanging, but continues over time. For example, take Google Alerts. You may have a persistent explicit search on your name or your company, and every time Google News has a positive hit, you get an update.

An implicit search is more vague, keying off related activity or your own interests and alerting you to things you might want to know. Blinkx, the video search engine, was talking about leveraging implicit search back in 2006 with a tool it calls Pico, which found relevant results to display alongside your usual activity. Tools like this, including Google Reader Magic, some aspects of TiVo recommendations and Amazon, are implicit, showing related items you didn't search for outright.

Enter the world of persistent implicit search - and think about how this could be applied. What about dating? What if over time, a digital algorithm figured out who your soulmate was, and was constantly searching to find if they were around the corner from you? What if you were looking for a new car or a new dress or a new home, and your smartphone alerted you that the right place at the right price had just become available?

As humans, we are in a state of persistent implicit search. We are always looking for things that may make us more comfortable, more happy, more rested, more rich or more successful. The trick becomes training the machines to think like humans, and to mimic that human ranking function that separates the good noise from the bad noise and brings you the right information at the right time, even if you didn't look for it yourself.

Any time you have a big company (like Google) talking about getting to know you even better, based on information discoverable, either public or private, it is not surprising to hear cries of concern over privacy. Schmidt was careful in his comments to say this new search would be "with your permission - this is personal search, for you and only for you". You opt in and you get the new world of personal, persistent, implicit search, which can find you the right answers anywhere you are from any device.

The world is no longer one where there is one right answer for everything. Google knows that, and so do the rest of the search engines, big and small. But Schmidt is talking about it clearly and correctly. It's very close to the language I've seen Barak Hachamov of my6sense use when he and I discuss the future of the apps and the API. This direction is more than just cool apps. It's about the future of discovery and artificial intelligence. Watch and see which companies get the transition right, and which are solving for yesterday's needs and problems.

Disclosure: I am the vice president of marketing at my6sense, which is also a Paladin client.