According to research by Martin Feuz and Matthew Fuller (University of London) along with Felix Stalder (University of Zurich for the Arts), contrary to the popular belief and expectation that Google’s personalized search results would be more accurate, it was found that personal search results are not particularly high quality and actually reflect Google’s central perception and even worse – they serve the interests of advertisers more than those of the searchers.
A Bit About Personalized Search
Google started what is called “personal search history” about 6 years ago, and a year later (in 2006), personalized search based on that search history was introduced. The official launch of personalized search into Google’s search world occurred about half a year later, based on the user’s behavior on Google – the results they tend to click on, the sites they save under their Google favorites, as well as on their Google home page (iGoogle), the time the user spends on sites they enter from Google (or rather – how long it takes them to return to the search engine and perform a search again), etc.
It should be noted that personalized search was not widely implemented until about two years ago (2009), when users began to receive personalized search results even when they were not logged into their Google account (Gmail, AdWords, Analytics, and other free services offered by Google, through which it continues to collect more and more information about user behavior).
The study raised 3 hypotheses:
1. Personalization is supposed to be subtle and relatively unnoticeable. It was found that personalized search results actually rose quickly, accounting for more than 60% of the results on the first page. However, it seems that the subjects indeed struggled to identify these results.
2. Long-tail terms will be more common as the information about the user’s search history grows. This hypothesis was not confirmed at all.
3. Personalization reflects only the user’s past searches and their online interests. It was found that all philosophical subjects received personal results for certain searches even when there was no connection between the search history and the search terms. The researchers concluded from this that Google created a group of profiles based not only on search terms but also on demographic information and various social preferences. This, as mentioned, actually leads to more similar results (=less personalization).
The study is, of course, not perfect, both based on the dynamics of Google’s search results (see frequent algorithmic changes) and because the nature of personalization makes it difficult to set a uniform starting line for all subjects.
On Personalization and SEO
In the context of SEO in search engines, there have been quite a few discussions on the subject. Some were quick to declare the death of the SEO field with the claim that it would no longer be possible to influence search results, as every user would receive a different results page according to their search history. However, from what I see, personalization has only benefited the field of SEO and search engine marketing in general – nowadays, it is no longer about promoting keyword phrases and their long tails. A good SEO must examine the behavior of visitors to the site they are promoting (see investing more time in statistical analysis) and ensure that the target site indeed gives the user what they expected to find in it. In one word: relevance.
Just as search engines seek to provide, through personalization, more focused results, so must website owners provide more focused and accurate information. For example, if your site is a virtual store for computer games, your content should focus on this topic and not drift to related content (something done in the past, as mentioned from trying to bring more traffic to the site) such as online games or mobile gaming apps. Irrelevant content can bring a lot of traffic to the site, but a significant part of it will be irrelevant traffic, which will increase the bounce rate from the site and actually make it clear to the search engine that your site does not really meet the search queries of the users.
Thus, the next time a user types in “mobile gaming apps” in the search engine, the search engine will no longer display your site at the top of the results.
On the other hand, if someone searched for the computer game Battlefield 2 and your site indeed offers this game, it is likely that the user will stay on your site (and not return to perform another search on Google) and will return to it to purchase other computer games. Such user behavior will make it clear to the search engine that your site is relevant to search queries related to the world of computer games.
In conclusion,
Even if the aforementioned study is not perfect and in fact only points to the tip of the iceberg, there is no doubt that the subject should continue to be researched. From a marketing perspective, it is essential that SEOs and online marketers pay attention to the effect of personalization and ensure that their sites are more focused and in other words – more loyal to the user. To some extent, this is just aligning with Google’s initial statement – it all revolves around the relevance index.
