The impact of Facebook’s Graph Search on traditional search

Facebook yesterday announced their well anticipated entry into the search world with the unveiling of their new ‘Graph Search‘, an upcoming product allowing users to search based on the recommendations and interests of their connections.

How does Facebook Graph Search work?

How does Graph Search work?

In what Zuckerberg aptly referred to as ‘some really neat stuff’, Graph Search will aim to index all the posts, shares, photos, and ‘likes’ of users and deliver them through the use of simple and concise language. Typical uses of Graph Search will differ from the every day queries searchers bombard Google with; searches are entered in natural language such as “What are some good restaurants nearby?”, with results generated from each individual’s social graph. One senior Facebook engineer gave the example of “Who are single men?” as a query which would return a list of single man connections, raising the additional possibility Facebook may also be entering the online dating market. Non-descript and custom filters can also be applied, again in natural language, to any search. So continuing with the dating theme, the query “Single women who appreciate E.L. James” could be used to better filter down results.

Facebook Personalised results

This important demarcation of fully personalised results is only possible due to Facebook’s unique position, and vast data stores. The sheer volume of information Facebook has at its disposal will allow search queries to potentially be answered with a greater layer of relevance than traditional search engines, who while able to filter results based on location and search history, know nothing of a user’s social life and connections. Given the success of social music search engines and sharing tools – such as Spotify and Last.fm – it is likely that users will quickly see the benefits of recommendations from their friends in other markets, when searching for content. An additional blow to Google has been dealt by Facebook choosing to partner with Bing, in delivering additional search results where the Graph Search finds little of relevance. The enhanced relationship is likely to be mutually beneficial, given Bing already uses social data from Facebook to enhance its search. The partnership should also prove a welcome boost to Bing’s share in the search market (which is currently hovering around 5%), given that Facebook boasts over 900 million active users every month.

Social search potential

For now at least, Graph Search sounds like a useful and practical addition to Facebook’s social armory The potential to quickly cut through a mass of personalised recommendations with natural search language appears much more powerful than the similar efforts of Google+, which currently only apply a thin layer of relevance adjustment based on your individual connections. What remains to be seen is how useful results are in nicher categories, and how much bias is introduced in certain instances. For example, if twenty of my younger cousins are die-hard One Direction fans, will Graph Search filter out these results based on age, individual preferences, and how strong a connection I am with them? If Facebook engineers are clever about the relevance drawn from each connection (and it is likely they will be), then social searching could be poised to take a fair bite out of the traditional search market as users discover new experiences tailored to them.