Social Bookmarking: A New Search Solution?

Integration with social bookmarking websites may represent a revolution in Internet search.

Social bookmarking websites like Digg, StumbleUpon, del.icio.us, and Reddit are among the most high-traffic websites on the Internet. On these sites, users post links to websites, and the community of users rate the site according to its quality. Some web analysts speculate that social powered search engines may be the future of search. It is believed that the human categorization and quality assurance that is already done on these sites may be combined with existing search algorithms to create a stronger search engine with higher quality results, but there still remain questions and problems that plague the current models of social search.

Already, the social search experiment is being tested with several ventures. 50 Matches, for example, is a webcrawler that only searches sites rated on social bookmarking sites. This selective filtering yields a small number of matches that are expected to be of higher quality than a computer-generated search.

A "human-optimized search" - especially one implemented on existing structures of human activity - sound ideal. However, some questions still remain when considering the possibility of social search. For example, how is a search engine going to know what the real "best" or most relevant results are based on these rankings? Social bookmarking ratings are highly based on sensationalism and the types of sites that offer interesting, but likely not-too-useful information. A social search site that has a mission to keep users entertained may be a great product, but as far as relevant search goes, the results may be less than perfect.

Next comes the concern about spam and unwanted results. It is likely that, if social search catches on, spammers would begin to assault popular social bookmarking sites with ratings to boost their own sites' "reputation". There would need to be some mechanism in the engine to confirm the validity of a user ranking on the social bookmarking side.

Next, web search tends to cater to the concept of The Long Tail, wherein the vast majority of search queries are performed very infrequently. A search engine that crawls social bookmarking sites alone loses the quantity of indexing that has made Google famous (though it should be noted that Google's PageRank system bears similarity to many social bookmarking models). The search results would almost assuredly suffer from a lack of breadth in websites indexed and would fail to generate relevant results for the majority of queries in the Long Tail.

For now, social search remains to be a big question mark in the future of search. Developments in social search may yet influence the future of search, but there are still a few big questions that need to be answered before a strong solution can be engineered.

User Comments