Tuesday, May 22, 2007

Universal search and personalization

Gord Hotchkiss posts some good thoughts on how personalization will be part of Google's universal search:
Personalized search is the engine [that] is going to drive universal search.

When you look at the wording the Google throws around about the on-the-fly ranking of content from all the sources for Universal Search, that's exactly the same the wording they use for the personalization algorithm.

[Google's personalized search] operates on-the-fly, looks at the content in the Google index and re-ranks it according to be perceived intent of the user, based on search history, Web history and other signals. It's not a huge stretch to extend that same real-time categorization of content across all of Google's information silos.

As Google gains more confidence in disambiguating user intent, more specific types of search results, extending beyond Web results, will get included on the results page and presented to the user.
Gord makes a good point. When talking about universal search, Marissa Mayer motivated it by saying, "The best answer is still the best answer."

But, the best answer for me is not necessarily the best answer for you. Especially in cases where there is ambiguity in intent and multiple possible verticals that might be relevant, the deciding factor could be what I have done in the past, my search and web history.

If Google is not personalizing universal search already, it likely is only a matter of time before they start.


Anonymous said...

I think Gord is confusing personalization and Universal search, or at least it seems so based on the way he is phrasing it.

The way I understand it, Universal search is like merging n sorted lists -- for a given query, take top n results from web search, top m results from news search, top p results from image search and so on, and merge them to come up with a single set of results stack ranked in order of relevance.

The trick clearly is how you merge all these disparate result sets. Regardless of how you do it, if it is done the same way for all users, it is still Universal search, but without personalization. But if the way you merge these result sets is different for each user, and is a function of their preferences (as exhibited by their historical behavior and that of similar users), then it is personalized Universal search.

So again, personalization is a bias/filter you apply on top to re-order the rankings you computed. And contrary to what Gord says in his post, I don't think the two are inextricably linked, nor do I think personalization is the engine driving universal search (the need for universal search might just be a result of the realization that (a) very few users use Google's vertical search interfaces, and (b) it is a much more seemless user experience to organize all info. relevant to a query regardless of its format into one result set).

Greg Linden said...

Right, the two are separate, universal search and personalized search, but I don't think Gord is confused on that.

For example, as I quoted, Gord talks about personalized search, but then talks about how personalized search could be extended to help pull information from "Google's information silos" and "disambiguate user intent."

I think we are both arguing that personalization will be used in universal search to help with the ambiguity on user intent and which vertical results should be shown.