The evolution of Facebook's search engine has once again brought the intersection of social and search into the headlines. Powered by Microsoft and a semantic web vision consisting of Facebook's enormous user base and the open graph protocol, the potential can't be ignored. But while there is plenty of controversy, there is little clarity for online businesses on how to comprehensively optimize their own sites for all of the traffic-driving capabilities that bridge social and search.
How search taps into social
As the chart below illustrates, there is little doubt that social networks have become an important source of traffic for online businesses, requiring specific optimization strategies.
Source: Compete, April 2010.
And while social networks are leveraging their assets to make semantic search a reality, the familiar search engines are also looking at social as the future, evolving their algorithms and structures to incorporate social information
The benefit for online business? Increased traffic. The benefit for people? Greater relevance and a more personalized web. At the end of the day, it's about discovery -- how people find where to go on the web -- and the winning companies will be those that are able to deliver the most relevant, personalized discovery experiences.
So where do social and search intersect today? And how can online businesses use the tools available to socially optimize their own sites for increased traffic? What are the other benefits? Let's take a look.
Channels of traffic for on-site social optimization
Let's begin by looking at the three sources of traffic, or channels of discovery, for which you can socially optimize your site:
- Traditional search
- Social network search
These differ significantly in terms of social influence, search intent, and volume driving potential. The following graphic plots the three types based on those differences (credit goes to Danny Sullivan, editor-in-chief of Search Engine Land, for the inspiration):
Feeds are live streams of activity shared on social networks and across the web by someone or something with which the user has a relationship.
Feeds enable passive discovery and facilitate the transfer of information from one to a highly interested and relevant many. The majority of traffic coming to websites from social networks is the result of people clicking on an item they see in their feed -- an item shared by someone with whom they have a relationship.
Feeds as a source of content discovery are characterized by 1) low search intent (the person viewing the items is not necessarily looking for that content or any content in particular; the item was pushed to them) and 2) a high degree of social influence -- meaning that the item is viewed as worthwhile by a person or entity to whom they are connected. Feeds can also drive a large volume of traffic, as noted in the first chart.
Traditional search is characterized by a high degree of intent and is capable of driving an enormous volume of traffic, but it has relatively little social influence. Intent is high as a search is actively made on a specific term for which the searcher has both interest and a goal in mind. However, the results rank low on the social influence spectrum -- they are not determined specifically by your personal preferences or the wisdom of your personal network.
The major search engines aren't satisfied with this anymore -- particularly Google, which took steps to incorporate real-time searches on Twitter into its search results after failing to deliver current, relevant results about Michael Jackson on the day of his death.
Today, Google incorporates real-time results into its main search results pages (SERPs) when the volume of activity warrants it, but in most cases places them into a separate section of search called "updates," as pictured below:
With the introduction of Google Buzz, Google is also experimenting with search results that are based on what they know about your social graph, as in this example that appeared at the bottom of the search results for a search on "recipes":
Social network search
Social network search is exciting because it sits in the upper right quadrant of the discovery chart -- ranking high on both social influence and search intent.
In order to properly describe social network search, let's briefly review the concepts of the semantic web and the open graph. The idea behind the semantic web is that the internet can be more usefully or relevantly organized and described by the relationships between people and social objects, rather than just the relationships between pages (i.e., the hyperlink-based system on which traditional search has been based).
So what is a social object? Think about it by starting with the Facebook social graph -- a combination of connections among people, groups, and entities like business or celebrities.
(From Dare Obasanjo, "Facebook's Open Graph Protocol From A Web Developers Perspective")
But of what use are these connections? The social object, in a nutshell, is the reason a specific conversation takes place. Human beings are social animals. We like to socialize. But if you think about it, there needs to be a reason for it to happen in the first place. According to Hugh McLeod, that reason, that "node" in the social network, is what we call the "social object."
So how does semantic data affect social network search results? Let's look at an example. On Facebook, a search on "Inception" pulls up search suggestions as the user types that look like this:
The first result displayed is the "Inception" page on imdb.com. As it happens, this searcher has already clicked the "like" button on this page, which connects her to the social object "'Inception' movie on imdb" -- a connection that Facebook will be aware of.
Note the search results displayed in positions No. 2 and No. 3 -- one a website, one a Facebook page. As it happens, one of the searcher's Facebook friends has also clicked the "like" button on this "Inception" page on imdb.com, and the search results list both his name as well as the number of other people (who are not necessarily the user's friends) who have liked it. The hierarchy of results displayed appears to be 1) display items "I" have liked, 2) display items "my friends" have liked, and then 3) display items "other" people (Facebook users) have liked.
If the user declines to select any of the suggested results, Facebook displays a full page of search results consisting of 1) popular Facebook pages with the term "Inception" (not shown in the screen capture), 2) posts by the searcher's friends that contain the term "Inception" and 3) web results from Bing.com that match exactly those displayed when a search is conducted directly on Bing.com, as shown below:
Social network search today is still quite small, comprising only 2.7 percent of all searches in March 2010 (according to GigaOm). Furthermore, the majority of searches on Facebook consist of two words, suggesting that most searches performed there are done to find people. But the size of the Facebook user base and the high activity level of the Twitter user base suggest that once social network search as a product is ready for prime time, Facebook and Twitter can place it front and center in the user experience and quickly gain share.
So how should marketers and publishers be driving increased traffic by leveraging the connection between social and search? Here are four "musts" of on-site social optimization to get you started:
1. Facilitate social referrals
Social referrals are the No. 1 way to drive traffic to your site from social networks, and are also a growing element of traditional search. Here are six key areas on which to focus:
- Remove friction from the sharing process. The social networks and portals all make second-generation sharing tools (those tied to the use of an existing identity), enabling users to share without leaving your site, and with a single click.
- Prompt people to share. Prompt users to share after they complete activities, such as watching a video or taking a poll. And provide options to share right within the site flow. Prompting people at the right times, and incorporating auto-sharing features can increase sharing by as much as 500 percent.
- Enable sharing to multiple feeds. The feeds on Facebook, Twitter, Yahoo, Google, Windows Live, and more have all become viable and sizeable sources of traffic for websites that have implemented these second-generation sharing tools. Be sure to give your visitors multiple options for sharing, as well as the ability to share to multiple feeds at once.
- Optimize your content for performance. Second-generation sharing tools give your site the ability to specify image, copy, links, and more for every item shared. Be sure that each of these elements is optimized to drive the maximum number of return clicks whenever someone views it in a feed. Is the image compelling? Does the copy make sense? Is there a call-to-action? If you're not driving many return clicks, your content could be the culprit.
- Shorten all links. Not only is link-shortening critical on limited character platforms like Twitter, where you might lose all return clicks if your link doesn't fit, but it is also critical for tracking return traffic from all providers. This effort is critical for tracking ROI. If brand is a concern, it is possible to create branded or vanity URLs.
- Remember your social content. The moment at which people contribute content to your site -- from experiences to photos -- is the point in time most likely to result in a share, so make it easy for them.
2. Ensure social content is both search-engine friendly and feed friendly
While most sites make their core content accessible to search engines, many don't realize the importance of making their user-generated content both sharable and search engine friendly. Comments, reviews, and forums are the most popular type of user-contributed content, and most of the platforms that offer these features are built with SEO in mind. But check that you are taking advantage of all of the tools available to structure this content. Also be sure you are incorporating all of your content opportunities; for example, if users can contribute recipes, photos, or other rich content, ensure that this information is accessible and well-structured.
3. Add semantic search classification
While social network search volume is small today, there are several reasons to add semantic classification to your site's pages, as well as to tweets made from your site. First, they offer another opportunity to generate feed items, increasing the amount of sharing happening on your site and driving traffic back to your site from the feed. Second, classifications like Facebook's Open Graph and Twitter's Annotations are clearly a foundation on which to build a search product.
Best practices include:
- Capture as many levels as possible when classifying each social object. For example, don't stop at tagging an Elizabeth George novel as "book"; be sure to also classify it as "mystery" as well.
- Use FBML (Facebook Markup Language) to get maximum exposure in the feed.
- Include higher-level objects like your brand and website.
4. Enable social discovery from within your own site
Optimizing your site for social discovery also includes providing social and semantic web context to the content and activity discovery mechanisms on your own site. There are two primary ways of doing this:
- Social sign-on: Websites can best leverage social graph and semantic data by enabling their users to sign-in with an existing social identity, which gives the site access to rich user and friend data, depending on each user's privacy settings. This is a capability the social providers make available to third-party websites; however using a vendor can simplify the process of both adding and maintaining this functionality for one or more providers.
With social sign-on in place, a website can apply both social network and semantic data to create custom features. For example, a site can add an activity feed that incorporates users from multiple social networks, as well as the specific activities that a site wants to display. At the other end of the spectrum, consider Amazon's new personalized pages. In the example below, after signing in with a Facebook identity, the Amazon user is presented with two social discovery features: One is a list of friends who have birthdays in the coming week. Clicking on the gift suggestions link below each friend's photo brings up search results based on the books and movies that friend has "liked." Amazon is taking the semantic data of the likes and applying their own search algorithm to it. No. 2 is purchase suggestions for the Amazon user based on books and movies that their friends have "liked." The suggestions are accompanied by a display of how many of the user's friends have "liked" the item, as well as photos of those friends. The friend's name appears when the user mouses over any photo.
- Simple social network plug-ins: Facebook and Twitter also provide simple activity feed plug-ins for third-party websites that present what a user's friends have liked or shared on that site, incorporating both social graph and semantic data. The plugins are simple to add. However, "like" data cannot be accessed and applied by the website, the plug-ins are currently only available for users of Facebook or Twitter, multiple plug-ins need to be implemented to serve users of both networks, and the plug-ins cannot be customized.
Applying social information to improve the way people discover information is the future; both the search engines and social networks see that future, and their products now reflect it. For online businesses, the landscape has become more complex, but it also means greater opportunity as power shifts back toward websites that are able to successfully optimize for these new channels of social discovery.
On Twitter? Follow iMedia Connection at @iMediaTweet.