The Impact of Social Sharing on Google Search Results

The next session at Digital Methods is a plenary panel which begins with Christina Schumann, whose focus is on Google and other search engines as technological actors on the Internet. Search engines are especially important as they now serve as a kind of gatekeeper on the Net - but the criteria they use for ranking and structuring information are often far from transparent.

The basic approach of search engines is to crawl or otherwise gather Internet data which are then indexed and processed into a database; this database is queried as a search query is entered into the search engine. Factors in returning search results include on-page information (content, programming, and design of Web pages) as well as off-page metadata (especially the link networks surrounding each page, relative to the theme of the query).

Such links were traditionally made through Web hyperlinks, but sharing of URLs through social networks - social signals for the relevance of a site - are increasingly important in this context, and are being integrated into search engine evaluations in various ways. This may also take into account the authority of the sharing user (their activity and network profiles) and other social network factors.

How is this operationalised in the case of a search engine like Google, then? What is the impact of social sharing on the overall ranking of Web pages? Christina's study examined both the internal and external impact of social sharing: she created four different test Web pages about a unique, fictive topic, which were shared across three different social media platforms (Facebook, Twitter, Google+) using both newly created test accounts as well as pre-existing social media accounts with a reasonable level of existing network authority.

Sharing occurred in two phases, to enable an examination of the initial and later impact of such sharing activities, and results were monitored using an SEO monitoring tool. The Google+-promoted site appeared most quickly, but overall the sites only began to rank once the pre-existing social media accounts began to promote them; Facebook and Twitter promotion eventually outperformed Google+.

A second experiment took a similar approach but built on a pre-existing topic (student housing) and used only pre-existing social media accounts of varying authority in the three networks. But in this case, the sites didn't appear in the generic student housing search results; only searches for the specific site names generated results, and here especially in the weeks where social media promotion was strongest.

This means that social media shares are only of limited influence on search results at present, but that they do have an impact, especially for unique topics - but the impact is short-term and doesn't necessarily generate lasting prominence. Of course, it's impossible to know whether and how Google is continuing to tinker with its ranking algorithm, so things may well have changed again by now - longer-term observations are necessary to trace the role of social sharing on search results.