By Sarah Douglas, Senior Search Optimizer
There is a lot of buzz right now about how social information is the future of search. Some predict a simple incorporation of Facebook "likes" as a replacement for traditional reviews from strangers on the net; others go so far as to imagine a future where social is fully integrated into the search engines, allowing your social circle to influence your personalized results by default. At G5, we’re excited about social media and its impact on the future of the search industry, but we’re exploring the limitations to this new concept for personalized search.
Does the Integration Benefit Companies More Than Consumers?
It isn't hard to figure out who stands to gain monetarily from the integration of search and social. Facebook and other entities that possess coveted consumer information are obvious winners in this arena; so are those that sell advertising space, like Google and Bing. The ability to tailor ads and search results to a user's probable preferences is clearly a valuable goal that will make these companies even more profitable. Consumers can benefit, also. Custom ads based on social data and user preferences make those ads more relevant to the end user. And for searchers who trust reviews and friend recommendations to make decisions, seeing what their friends “like” in results (when searching for a restaurant, for instance) could be very useful indeed. Still, not all searchers are created equal, and ultimately a significant assumption is being made: the preferences of an individual's social circle are highly influential on the individual. While social/search integration may help introduce new potential customers to a given product or service, it may not be as useful to the consumer who has individualized tastes and values.
Social Data Isn't Always Relevant
Conspicuously missing from the discussion about integrating search and social is the fact that social data is not useful for all queries. For transactional searches and those with local intent, there may be value in incorporating the social data. As a consumer, I may be curious about what brand of shoes my friends recommend, or which restaurants my family members enjoyed on a recent trip. However, this is not the case when it comes to queries with an informational intent. When researching something specific for work or school, it is unlikely that social data will be available to begin with. If it is present, the utility of social data is diminished since my objective is finding a good informational resource – traditional determinants in the search algorithms are going to be far more effective at selecting information that is reputable and most often cited.
Social Will Always Depend on Elective Participation (We Hope!)
Seamless, automatic sharing of information between social and searching is something that we hope to never see become a standard practice. Currently the social information used most often in search results are "likes" and "interests" pulled from social network profiles. However, one imagined integration is a future where the individual user's search and click through habits would be automatically incorporated into the search results of their social circles. The potential privacy issues are numerous; it is nearly impossible to imagine how social could be implemented without elective participation. There are several topics searched for that are not normally shared, even among friends – medical conditions come to mind as a tame example.
What Are the Likely Applications of Social and Search?
Based on the previously mentioned concerns, it seems likely that social data may be incorporated on a limited basis, much like reviews are already incorporated in both ranking algorithms and search results. Targeting ads to a friend of someone with a given interest is better than random targeting, so that will probably become a standard practice in advertising. The fact remains that no one knows you better than yourself, so it makes sense for Google rely more heavily on user history rather than social data when serving up results.