This week has seen much press coverage of Google’s initiative to collect and analyse individuals' search histories in order to provide more relevant, personalised search results.
If you type in the search time 'golf', it points out, its current search algorithm doesn’t know if you mean Golf as in VW Golf the motor car, or golf as in the game golf. But if it has a history of your previous searches, it will have a pretty good idea of which one you mean.
So collecting personalised search histories represents a win-win-win, says Google: the searcher benefits from more relevant results, advertisers benefit from search-related ads that are also more relevant, so Google benefits.
But Google’s initiative has caused a minor uproar. How much more intrusive can you get than a search engine collecting a personalised history of your own personal searches? The privacy implications are huge.
Peter Fleischer, Google’s global privacy counsel has responded to these concerns with this argument:
"Our policy puts the user in charge,” he says. “It is not something Google seeks to control. At any time they can turn off personal search, pause it, remove specific web history items or remove the whole lot. If they want, they can take the whole lot to another search engine. In other words personalised search is only available with the consent of the user."
With this, Google has made a big step forward. It has understood the difference between ‘permission’ in abstract (which in many marketing circles is taken to mean permission to spam and do whatever we want with your personal information) and permissions, plural.
Permissions management is one of the key ingredients of tomorrow’s information management infrastructure. Permissions are always contingent and context based: what I want to do right now, with whom, how much I trust them, and so on.
It’s also intriguing that Google are creating a facility that allows you to take this history to another search engine. This recognises the fact that strictly speaking this data is not Google’s, it’s yours to use and share (or not share) as you wish.
There’s another way in which Google’s initiative represents a step forward. By accepting that blanket algorithms don’t deliver personalised value, Google is accepting that the real power in search is not its algorithm per se, but the input of information from the user. It’s moving further towards a bottom up approach, rather than a top down one.
Nevertheless, two big issues still remain unresolved.
First, which side of the fence should the information reside on: in my database or Google’s database?
Fleischer claims that for personalised search to work, "search engines must have access to your web search history." But what does 'have access to' mean? Does it mean that Google collects and keeps the data unless told otherwise; or that the individual is given the means to keep the data and then allows access to it?
Second, is this really the best way forward for personalisation?
The traditional corporate mindset assumes that personalisation is delivered by the company to the user via the expensive and cumbersome process of collecting as much information about the individual as possible and then data mining this information to create guesses about what might be relevant to that individual.
This is Google's approach too. It is still making a guess about which golf you are interested in when, in reality, it could simply ask you.
This alternative approach goes in completely the opposite direction. It is based on enabling individuals to provide ever richer specifications, using ever-easier processes to do so.
In the search for 'golf', for example, why doesn't Google develop a pop-up or drown down menu which simply asks 'do you mean VW Golf or the game golf?'. This would allow the user to specify, without
Google needing to collect any personal search histories at all.
So, even though Google is saying the right things, it’s still travelling in the wrong direction: two steps forward coupled with two steps back.
As a result, privacy concerns about Google can only grow and grow.
Until corporations understand and accept that the future lies in individuals owning and managing their own personal data, these stalemates will continue.
But the breakthrough in understanding seems to be getting closer.
25 May 2007