A few days ago I came acrossthis item in Google’s blog – looking at what they call ‘Social Search. This is a set of applications being developed by Google to allow image content that you and your social circle (as set up through your Google account) have posted on image sharing sites such as Flickr in searches returned by Google Images. So the idea is that you do a search on particular images using Google Images, and prominently featured in the results set would be images that your friends have posted up on these other sites. I assume that eventually this sort of thing will spread out to encompass other sites of user generated content – Facebook, MySpace, personal blogs, etc. Of course, this would require some cooperation between the companies running these sites and clearly there would be financial issues involved, but technically it’s not that difficult.
At first glance social search looks like a very cool concept. After all, we tend to ask our friends and colleagues for advice and guidance on where to buy things or find them online. We take their advice on what web sites are reliable, we are likely to at least look at films or books recommended by people who know our tastes, and so on. If it did become possible to pull together information about searches carried out by groups of friends, and include information posted or recommended by our friends in search results in a prioritised manner, then the results would probably be more immediately relevant to us, and would also be at least partially validated – rather than the results being the equivalent of a cold call, they’d be closer to a personal introduction.
However, it struck me that there’s a potential downside to this approach, especially the more integrated in to the overall search results the ‘personally linked’ social search results become. There is a phenomenon well known in management consultancy circles called ‘Groupthink’. It’s what happens when you get a group of people who’re closely linked in some way – members of the same close knit team or department, for example. What can happen during decision making and problem solving sessions is that the group may come to decisions based upon internal politics and ‘norms’, rather than objective facts that are presented to them. This effect has been seen to be responsible for poor decision making in a wide range of situations. It struck me that there is a good chance of this effect becoming evident in search results should the ‘Social Search’ really take off.
For example, if someone in a social grouping is particularly ‘active’ online then their comments and recommendations might turn out to have a larger impact than other folks who’re less active online but possibly more informed about issues. The overall effect would therefore to bias such social network search results towards the people with the largest online profile rather than those results that are possibly more accurate. Such individuals would thus become opinion leaders and formers in particular social groups, and advertisers could easily seek out these higher profile individuals to sell directly to them, working on the principle that they will sell to their circle of contacts either directly by recommendation or indirectly through the results of social search.
Slightly disturbing. Whilst influencing small groups of people it’s not the end of the world, but how long before we get a situation similar to that in the Phillip Dick short story ‘The Mold of Yancy’, where the behaviour of a whole civilisation was influenced by the tastes and preferences of one man? Far fetched? Perhaps not.