A Prescient Search Egnine?
Monday, November 9th, 2009xConomy Seattle presents a forum about the Future of Search on November, 30th. It made me think about Search trends, and what can we expect in the Future…
First, mobile will become more predominant. Mobile devices as the only truly ubiquitous computing platform, and recent improvements made them, at last, usable as Internet devices (hello, iPhone). But search on mobile will remain very different than on larger computers. Mobile devices are mostly used in interstitial periods. The ‘always connected’, always busy’ mantra of contemporary life makes use draw our smartphone as soon as we’ve got 30 seconds of idle time. Your partner gone to the restroom? Check mail… Waiting at a traffic light? Download news… The main consequence is that mobile devices are only used for short, interruptible periods of time.
Whereas some companies saw Local search as the killer search application for mobile, the main hurdle remains monetization. Imagine an End User looking for the nearest Home Depot. Giving its address and directions for free means no revenue at all; displaying a Lowe’s advertisement instead is unacceptable; asking to pay for the information is just wishful thinking. There’s much less serendipity in search when done from a mobile device, and therefore much less monetization opportunities. Not because of device limitations, but because Users are themselves time and location constrained. While major search engines will continue to invest massively in improving their mobile experience, mobile search per se (i.e. on the go) will never match the money-printing ability of Web search. Unless it brings serendipity in a few seconds. Unless it brings relevant search results even before an End User entered her query…
Even notebook and desktops users want instantaneous and effortless gratification. Internet Users, who spend less time in front of the TV, still appreciate the mindless, motionless comfort of drifting away beholding their screen. Because they have introduced a fundamental new way to consume information, Search Engine are poised to bring the next revolution as well. Continuous, time-unified consumption (e.g.books to TV shows) is progressively exploding into fragmented, self-guided time segments. This is why the Prescient Web will be brought by Search Engines, and not by PointCast revisited.
Automated clustering (grouping alongside most common keywords) never delivered its promise for better results organization. A better approach is simply to organize results based on their source or media. By separating Image, local and Web search results for instance, Search Engines are able to provide a simpler interface to each media. Engines continuously adding new sources (music, social media,…) will further discriminate their origin in order to keep the results intelligible enough for End Users.
But the real change will come from intense personalization. A personalization that goes well beyond the filtering or sorting of results based on previous history, but fully leverages the social web to deliver meaningful, pertinent search results… even before entering any keyword!
People are now connected enough, both technically and socially, to make it possible to ‘map’ all their inbound and outbound influence, their topics of interest, and their information consumption habits. Take all the browsing history, search history, tweets, blog posts, followers, RSS subscriptions, eMail exchanges, and you’ll get a pretty accurate understanding of somebody’s work and hobbies. If of one side, you know what End Users search or consume, and on the others, you monitor the ‘background noise’ in their information sphere, you’ll be able to pinpoint news or data they did not even know they should search for, at the first place.
For instance, I’m into photography. My Search Engine should know that, given the queries I sporadically enter to check out new Canon cameras. It could automatically tell me when a new model’s out, even if I don’t know it could be.
A Search engine could even analyze the transmission path of informations, and significantly reduce the propagation time by short-cutting intermediary nodes and feeding users automatically.
One strength of such Prescient Search is that it doesn’t need to be perfect, for it only supplements the current interface. It’s simply a matter of presenting users, by default, a set of data thought to be interesting for them. Not unlike Amazon.com, a Search Engine home page will display plenty of personalized items, while giving unhindered access to the whole data. It can also be applied one field at a time, and leverage the efforts already made for recommending music or movies for instance. Prescient Search will also be continuously improved, alongside the emergence of the semantic Web, new social networks and the simple fact that computing resources are becoming more affordable.
Needless to say, unlike mobile search, the monetization potential is real. Both on a consumer side, through advertising, and on a business side, where predictive information discovery could revolutionize organizations. Imagine an Enterprise search engine that would surface the problems even before people realize it, or a system that would automatically spotlights emerging trends. How much would it be worth, in the modern information economy?










