Social Search Is Useless Without Context

I know that Google’s Social Search was released days ago and is now old news, but I wanted to use it for a bit before really commenting. Generally speaking it is an interesting idea to use your contacts from various social sites to drive search results. The idea they are exploring is that your social network contacts have more influence over you than a generalized application. For example, restaurant reviews from social network connections are probably more influential than a review on Yelp from someone that you do not know. This is definitely true from personal experience and I am sure there are various research studies that would back up this sentiment.

Mahendra at Skeptic Geek feels that this is only partially true. His post is in response to the various Facebook vs Google vs Twitter posts currently being discussed. Mahendra focuses on the fact that Google is probably going to lead the way because it pulls content from various networks as opposed to just Twitter or Facebook:

Real-Time Search will not be useful unless it is filtered by Social Relevancy.

Today, I follow several people on various social networks, who make up my “social circle”. The problem is not everyone uses the same social network for the same purpose. I follow some people on Google Reader as they share great content there, but I don’t follow them on Twitter, because they tweet mundane stuff I don’t care about. And vice versa.

This is an excellent point because many people do the same thing. Many people still work on the assumption that LinkedIn is your business network and Facebook is your friends network. The other social sites also have their strengths, like Digg with technical articles and anything Apple. So, having a search based purely on one site will likely not give you the results you are hoping for in all cases.

However, the same reason that Mahendra uses to say that Facebook will lose to Google, is the same reason why Google or any other social search player will struggle. Basic search treats all content types the same, meaning blog posts are the same thing as an article from a newspaper. Relevance is determined by a number of factors including authority (PageRank in the Google case), which is typically measured at the most basic level by the number of incoming links.

Social search is a little different due to the purposes of each social site. First, if you are looking for images, then you really do not care for results from Digg, but results from Flickr should automatically be ranked higer. This basic case is likely a feature of a social search engine, but what about a traditional topic search. Let’s say I am looking for “javascript charts and graphs” because I want to put some pretty charts on some website. I want content that comes from technically savvy sites, with heavier weight from programming related sites. How do I get that information with social searches? I would need to know what sites are best for programming related information, which still could use PageRank for basic authority. However, what if I am following someone on Twitter for their programming related items, but not on Digg because they share a lot of silly pictures there? Also, how deep does the relationship go?

In some basic tests with Google, it is gathering information from my direct contacts as well as some information from the extended network. In my “javascript charts and graphs” example, there is content from a solid technical source in the extended network, but I am only connected to them through a connection with someone else on Granted, I think I got lucky with this contact as my contact on is also a technical person, but I would not want programming related searches using my connections.

Without the context of why I am connected to a person, an aggregated social search may not be a great solution. Maybe site specific searches give us the context that we need to find information that is relevant. Another major point that is still missing is the personal authority and influence, which has been a major topic of discussion for many people. How does that fit with social search? Should I get technical articles from a connection that is an authority on marketing? Again, context rules, and without it, social search will be nothing more than an experiment.

3 thoughts on “Social Search Is Useless Without Context

  1. Rob:

    1. First, thanks for referencing my post. 🙂

    2. I absolutely agree that social search is useless without context.

    3. My post was about who is likely to win in the long run. Needless to say, there are many obstacles in the path towards social search nirvana. My point is if anyone is likely to reach there, it will be Google, rather than anyone else.

    4. Considering the example of searching for “javascript charts and graphs”, if the person you are following on Twitter (but not on Digg) has ever tweeted anything about javascript charts, you should get that tweet (or even the linked site) in your social search.

    5. I think the mechanisms to evolve an understanding of relevance in social search will combine algorithms and user actions. Today, Magic Sort in Google Reader works off your likes and sharing patterns. Clicks on search results are monitored to refine the relevancy algorithms. UI features like “Did you mean” or “More like these” or “Similar” can be further enhanced so that social search augments regular web search.

    The end is not yet in sight, but Facebook doesn’t seem to even qualify for the race.


  2. Mahendra,

    I have no problem referencing one of your posts 🙂 I tried to not look at who is winning or even likely to win because we do not know how these tools will evolve. If you look at Google’s social search, it really is not that good yet. Google also tries to make everything fit the search model, and they do not have a good history of really changing things. So, I am thinking that the playing field is fairly open. It will be interesting to see how all of the tools evolve.


Comments are closed.