Can Social Bookmarking Improve Search?

Josh Catone has a post on ReadWriteWeb titled Will Social Bookmarking Pay Dividends with Search Result Augmentation? that points to research out of Standford University that seeks to provide some answers on whether social bookmarking can augment traditional search results on the Web.

The report titled "Can Social Bookmarking Improve Web Search?" includes eleven experiments using del.icio.us designed to evaluate "different aspects of social bookmarking and their impact on web search". The main end results lead to the issue of needing critical mass which is still not here- the paper's authors estimate that only about on tenth of the web has been bookmarked and tagged in del.icio.us and therefore (at least when using only this domain) tagging is not yet ready to make a significant impact on search results.

However the report makes some interesting points that i highlight below that they believe can still be used to improve how search engines work.

• about 25% of URLs entered into del.icio.us are not seen in search engines for 4 weeks to 6 months after publishing. So one of the conclusions is that social bookmarking could be used "as a (small) data source for new web pages and to help crawl ordering.". I think that especially for niche vertical search engines this would make a huge impact- and could help them win traffic over the general search engines
• Users of social bookmarking sites are probably acting largely out of self interest rather than in concert. This means that their actions are largely uncoordinated. del.icio.us has relatively little redundancy in page information for perhaps 50% of URLs and high redundancy for perhaps 20%. Once again i think that bigger benefits could probably be seen in the vertical search space and SPECIFICALLY in the Enterprise space- where users tend to be more community and task focused.
• Tags might be able to help a search engine if tags can be matched to a user's query. This is applying traditional taxonomy search engine optimization models which have proven to provide great value. The report states that the tags that users are applying are on the whole accurate- of course the definition of 'accurate' is wholly dependent on the users and what and why the tag was applied to. For example, a tag of 'produce' could mean a banana or a command for the user- so when a user searches for bananas a work order to produce an art commission might come up so further disambiguation is the key.
  
• del.cio.us is already experimenting with including bookmarked results results as art of the Yahoo! search interface so it will be interesting to hear more about this from them
  

One thing the report did not evaluate is the display of the results when the tags are applied- which i believe makes a big impact on usability.

The last item that i thought was very interesting (and of course if you have been reading my blog for a while you will know why!) is this one conclusion around Domain tagging in which their recommendation is to 'pay' librarians to tag entire domains.

We also asked our users in our user study whether tags applied only to the URL involved or whether they applied to the entire domain. Our users said that about one in five tags applied not just to the URL in question, but to the domain as a whole. Both of these statements, both in terms of prediction and in terms of our user study suggest that a well financed search engine could pay "librarians" to tag entire domains rather than single URLs. Our broad conclusion was that paying such librarians would be more efficient, and might obviate the need for about 20% of the tags on URLs bookmarked.