The dangers of search personalisation


Search is getting more personal. Google in particular is striving to make the search experience as customised to each individual as possible. If Google knows I live in the south of England then it assumes I am more interested in results from websites with a connection to this part of the world than to the north of Scotland. If I am looking for somewhere to eat tonight, then yes, it’s very helpful. If I am looking for a major capital project for my business, then no I want best fit for my needs.

As someone working in SEO, I have of course been familiar with the increasig personalisation of search for sometime. Good or bad for the searcher, it certainly causes issues when talking to clients and trying to explain that different people looking on different machines may see different results.

But it was not until I read a book recently by Eli Pariser called ‘The Filter Bubble’ that I really appreciated the full impact of personalisation. The internet has been heralded as opening up information in way no other medium has managed to achieve and for that we must all be grateful. Personalisation though is having the effect of funnelling what we see by delivering what the search engines knows, or thinks it knows, we are interested in. I thoroughly recommend reading Eli’s book for an in depth analysis, in my view he in danger of becoming a little too concerned and even borders on paranoia as the book progresses but he makes some very interesting points that we need to think about.

We all know that everything can be, and therefore is being, measured on the internet. It is also reasonable to accept that all this information is, or certainly could be, available to anyone willing to pay for it. So it not unreasonable to assume that anyone who wants information badly enough has access to it – and this is perfectly possible.

So here are a few scenarios that Eli highlights in his book and which made a particular impact on me.

Personalisation and social media

Scenario 1 surrounds social media. Let’s say I come from a fairly poor neighbourhood and have lots of friends with dubious credit histories. I on the other hand, have ‘made good’ without leaving my roots behind completely and have an excellent credit history. Is it unlikely that financial institutions are going to ignore the company keep and not take it into account when assessing me for a loan?

Personalisation and social mobility

Scenario 2 is about how difficult it might be for someone to make good from humble beginnings. Let’s say from my less than privileged background I go to a modest University though I am very bright. Recruiters looking to fill the most prestigious jobs are most likely to target Russell group candidates or Ivy League ones in the US. This has always been the case but it is so much easier now to target them on social media. If I do not see an advert for that great job, there is no way I am going to get it. My application never gets submitted.

Personalisation and news censorship

Scenario 3 has perhaps the most serious implications for society as a whole. Increasingly people get their news from the internet. If I go an buy a paper, then I may dive straight for the back pages to read the sports news but I am probably at least aware of the front page headlines about phone hacking or the banking crisis. The internet on the other hand feeds me news based on what I interact with. So if I never show any interest in the banking crisis, gradually less and less news about it is put in front on me and gradually I become totally unaware of its existence.

Perhaps the biggest danger with personalisation is that it is insidious. As an SEO I am aware of it and can, if I choose take some steps to mitigate its effects. But most people are totally unaware that in effect their information is being censored. And that is dangerous. In fact, if I could argue that personalisation is the start of the end of the information age and threatens our very democracy. But that would be getting paranoid – wouldn’t it?  Perhaps it does just enhance the search results.

Update-24 Aug

Eli Pariser uploaded a video about his filter bubble idea a while back – explains the concept very well.


Comments are closed.