The Filter Bubble – Is Personalisation Damaging the Web?
In the recent SEO Quiz I posted on the SEO Consult blog, I mentioned a book written by Eli Pariser called 'The Filter Bubble: What the Internet is hiding from you'. It's a very interesting read, particularly for anyone interested in internet marketing or any of its offshoots, particularly as its subject matter is so divisive and controversial.
At the crux of the book is the argument that personalisation is fragmenting the web. Websites are increasingly tailoring their services to each individual user – this means that when we search Google, or one of the other search engines, for something such as 'climate change' or 'immigration' we are likely to be served results that support the preconceptions we already have on these subjects. The search engines are able to do this through monitoring what websites we have already visited, and determining which results in relation to a query we are likely to find more useful. This personalisation also extends beyond the search engines, Facebook use EdgeRank to determine which friends updates we see in our news feed, and it is argued that the recommendations Amazon generates are causing its users to become myopic because they aren't being offered an unbiased selection of results. In this search marketing blog, I'd like to explore the argument in a bit more detail – and hopefully at the end you can leave you view on the subject.
Are we in a Filter Bubble?
I'd like to focus specifically on Google throughout this blog because, as a search engine, it plays an incredibly large part in how we receive information on any subject. It is true that personalisation affects most of the searches conducted on Google, and the majority of casual users don't even realise that the results they are shown are tailored to their idiosyncrasies. The main argument that Pariser puts forth is that because the information we receive is suited to us, it is serving as a form of propaganda, solidifying our biases and leaving users with tunnel-vision as they are unwilling to listen, or just unaware of counter arguments against their beliefs.
Take, for example, a person that is searching Google for the query 'climate change' for the first time. The first link they click might be a webpage arguing that 'climate change' is a fallacy, from that webpage they follow links to other websites supporting this argument, and from those websites they delve deeper into this particular theory. The next time they search Google for 'climate change', instead of the SERPs being populated with unbiased information representing arguments both for its existence and against its existence, the results supporting the idea that it is a fallacy will be ranked higher, obviously making them more likely to be clicked. This starts a cycle where the searcher may just become unaware of arguments stating that 'climate change' is a reality, leaving them with a biased view and perhaps an unwillingness to accept other arguments. This is a bit of an exaggerated example, as it would generally take Google longer than one search, to really start influencing the results it shows, but it indicates how Pariser's argument is structured.
The second part of the argument is that it creates a privacy nightmare, because websites are capable of harvesting vast data on individual users, and this is creating a 'big brother' level of paranoia amongst some digital and social commentators including Pariser. There is truth in this argument, particularly as Facebook have recently had to restructure their privacy settings by setting on limits on how long they can retain data on ad-clicks, amongst other things. Another argument of Pariser's is that in authoritarian states, the algorithms that tell help filter results to suit our needs, can also filter out news that we shouldn't see – it increases censorship. So, while there are some significant arguments supporting Pariser's view, in the book he tends to ignore the counter claims to this argument – dare I suggest a hint of irony?
While personalisation does affect almost all internet searches, it is possible to remove all personalisation with a single click – a fact that goes unmentioned in the book. This may be because, as I stated earlier, many casual users are perhaps unaware that their results on Google are filtered. However, simply by clicking the 'Verbatim' search tool on Google, all personalisation filters are removed, and each Web Browser has the option of 'private browsing'. It can also be argued that personalisation makes it more likely that serendipitous results are more likely to be favoured as interesting results are generated along with relevant results. This is supported by a study conducted by Microsoft in 2009 that states, "rather than harming serendipity, personalization appears to identify interesting results in addition to relevant ones." This means that, perhaps as a consequence, a result we may not have come across in the deep 50+ pages of results the search engines generates, may be elevated to the first page based on our previous search history. This would suggest that personalisation is in fact opening more doors for us, rather than restricting them.
I would have to conclude that I disagree with Pariser to an extent, because I don't believe personalisation is harming users as much as he states, however, I do feel it would be beneficial for more users to be aware of how personalisation affects them and what they can do to limit those effects. I would suspect the majority of internet marketing agencies and search engine optimisation specialists would agree, as there is also the argument that personalisation is driving digital marketing innovation and requiring agencies such as Click Consult to work harder than ever to produce the best results for clients. I'd be interested to hear your thoughts on the matter, so please feel free to comment.
