Yesterday's announcement from Google about including Web history in search personalization marks a fairly significant development in disambiguating intent on Google. Consider the implications. One of the issues I had with the initial implementation of search personalization was that it really only worked when there was existing search history. That really only covered one in five searches for most of us. That also meant that personalization showed up most often in areas where you tended to do a lot of searching. For example, if you search within your industry a lot and tend to go to the same sites over and over again, you would find the site lifted on to your top page of search results. Of course, if you were doing the typical "vanity" search to see where you rank and you end up clicking on your own site, this would have the effect of lifting your site into the top 10 results. If anything, this implementation of personalization works to make navigation search a little more efficient. But I'm not sure it went too far in disambiguating intent, which is the holy grail for any search engine.
With the introduction of Web history, it's a whole new ballgame in disambiguating intent. This allows Google to move far beyond the well tred search path and actually taps into your current browsing behavior to try to determine what's on your mind right now. If Sep Kamvar's personalization algorithm is as powerful as I suspect it is, this could dramatically alter the results that you're seeing. The promise of personalization is greatest when it can be applied in areas that are new territory for you. It helps Google interpret just the kind of site you want to see, given your behavior at the present time.
Let me give you an example. Let's say you're looking at buying a new vehicle. Let's further say that you're fairly early in the consideration phase and your visit a lot of sites like Edmunds.com and Autobytel. This tells Google that you're looking for information and you're probably looking at sites that could be comparing your alternatives. If you've already visited sites like Edmunds.com, Google would probably lift those sites into the first search results page. If Google's algorithm truly makes a move towards a recommendation engine, what it can then do is find similar sites you may have never considered, based on the characteristics of the sites you have been visiting and make you aware of these sites. That's where the real win for the user comes in personalization. It's not just providing you a shortcut to sites you are already aware of, it's in making you aware of new sites you never knew existed, ranked and prioritized according to the PageRank algorithm. With Web history, Google can track your progress through the buying cycle to be able to match the information site you're looking for to where it believes you are, based on your current click stream data.
There are other implications that are very interesting to advertisers. Click paths tend to indicate the life events that you're currently in the middle of it. The life event could be a major purchase, planning a holiday, buying a new house, planning for a wedding, or graduating from university. In each of those instances, there are a number of linked consumer needs that tend to go together. There's been a significant amount of research done on how life events generate predictable consumer patterns. Web history gives Google a window into exactly what is happening in your life right now. I had written a column about how surprised I was with the glimpse that search history provided into my mindset at any given time. If you combine that with Web history, you would have a very finely detailed snapshot of both big and small events in my life for any time period. It gives Google the ability to precisely target search results based on exactly what's happening to me right now.
But let's face it, it's not the search results that Google is focusing on. Google is altruistic enough to make organic search results the testbed to play with the personalization algorithm, but the monetization opportunities in this are mind-boggling to say the least. When you combine the ability to precisely target and interpret the mindset of any given consumer with the multiple touch points that Google now owns to provide advertising messaging to that prospect, you have a marketer's dream scenario. When I asked Marisa Mayer about this she made it clear that organic results are what they are working on now, but they don't want their advertising network to be too far behind the curve. I'm still working my way through the interview making notes but I did want to get this post up because I think from a user perspective there's some important information here. For me, the promise of personalization is moving Google to be a true recommendation engine when it gets confident in disambiguating my intent based on my current behavior. Folding Web history into search history moves Google a quantum leap forward in being able to do this reliably and consistently.
The interesting question will be to see what kind of user pushback comes from the privacy concerns. Danny Sullivan touched on this a little bit in his post
. Will the trade-off of increased search accuracy be enough to have lots of users opt in? Obviously this is what Google is counting on and that's why they're introducing the enhancement in the organic results first. If they can provide a clear win to the user, than the trade-off seems a lot less formidable. And when they're introducing that usability lift in something as benign as organic search results, it seems a little less ominous and invasive. If they can get us using Web history by giving us a win-win on our search functionality, is a greater likelihood that we'll leave Web History turned for when they do decide to start rolling it in to their advertising presentation algorithms. Enough users will have it turned on it will give them the critical mass they need to appeal to the early adopter advertisers who want a take it for a spin.