The 9 kinds of context that will define contextual search


Yesterday I did the kick-off presentation and workshop at a strategy planning session of a major online media company. The intent was to provide some different perspectives on trends in digital media as input to their deliberations.

One of the many topics I discussed was the rise of contextual search.

Looking back over the last decade, I think it’s fair to say that the search experience has not evolved much. Sure we’ve had the shift to real-time indexing, experiments with multi-category results, predictive text in the search field, and a few other innovations, but if I was sitting in 2001 wondering how search would develop over the next 10 years I would be sorely disappointed to find out how little actually happened in that time.

Clearly it is a nonsense to always get the same search result, irrespective of who you are and all of the conditions surrounding the search. Yet for all major search engines there is currently minimal difference in the results from the same text string search performed by different people, in different conditions, very likely looking for different things.

At LeWeb ’10 Marissa Mayer of Google described “contextual discovery”:

“The idea is to push information to people,” Mayer said. She noted that on mobile devices this is particularly interesting because location can provide context. One example she gave was a menu when you’re in a particular restaurant. It would be great to show up and see that on your device — maybe with a bit of social flavor based on what your friends like, she added.

Here are 8 kinds of context that could be taken into account by information discovery engines (if the context is available).

* Individual. All of the history and context of a person.

* Demographic profile. Age, gender, occupation providing clues to likely interests.

* Interest profile. Interests in topics expressed explicitly and implicitly.

* Location. Country, town, and proximity to points of interest including shops, informing what is accessible.

* Device. Interface being used and previous use of that type of device

* Date. Day of the week (notably distinguishing weekday and weekend), time of year, and proximity to major events such as Christmas or Valentine’s Day.

* Time. Time of day, making coffee, beer, or movie guides more or less relevant.

* Weather. Searches for local destinations may be impacted by current weather, such as pubs with beer gardens being more prominent on sunny days.

* Mood. Positivity, negativity, excitement, hunger and far more, as reflected in status or other updates, could impact the content presented.

(Please add any thoughts on other relevant context in comments.)

The current greatest focus in using context in search is location. That’s useful to consumers, and is supported by the emerging business models around local such as coupons. However there are many other aspects of context that should be taken into account.

There are three other key issues to consider in how context is impacting search.

Push. The March 1997 cover story of Wired was on push technologies, how we would move from seeking information to the right information finding us. While the fad of the day collapsed ignominiously, we are now finally closer to being able to push to us information that we actually want to receive. Notifications on the iPhone are a simple example of this shift back to push. It will take a while to evolve, but we are ready for a limited form of push to replace a portion of the search we currently do.

Behaviors. The act of typing words into a search bar is so ingrained in us that it will take a long time to change that behavior. There are many opportunities to help people to discover content in other ways, or ways that combine rich context with keywords, but the activities that drive those will change slowly.

Agent technology. The concepts and early practice of ‘intelligent agents’ – that act on our behalf to find information and negotiate with sources and vendors – date from the 1990s. Here too many of us are now ready to share personal information that will allow us to be better informed, and to enable better search and content discovery through a variety of new mechanisms.

Contextual search is about to develop at a rapid pace, starting primarily with location and broadening from there. I look forward to seeing the pace of meaningful innovation in search accelerate in the next few years.