By Chris Sherman
August 16, 2006
Many players, big and small, are working in the social search space. Here’s a look at who’s doing what to harness the power of human beings in improving search.
In yesterday’s SearchDay, What’s the Big Deal With Social Search?, I offered an overview of social search, touching on the pros and cons of injecting human judgement into what has traditionally been a heavily algorithmic process. Today I’ll map out the various approaches to social search and some of the key players in each area.
Most of the major web search services are dabbling with social search. Other, smaller companies are creating social search tools as the core foundation of their business. Most social search services have some distinguishing technology or approach. What follows isn’t a comprehensive overview of what’s out there. Rather, it’s a sampler of what’s going on in the space.
While all of these services can surface interesting content, they tend to reflect the biases and interests of their most avid users. And since even people with a lot in common can also have widely divergent interests, they often contain “noisy” results that may not be relevant.
Shared bookmarks and web pages
One of the first types of social search services to emerge sought to leverage the power of shared bookmarks. The idea here is that if people save a particular page as a bookmark or favorite, they’re effectively voting for the page—fundamentally the same idea behind Google’s PageRank, but counting the “votes” of web users rather than webmasters.
In the early days of the web, directories were the most popular information finding tools. Directories are compilations of pointers to web sites created by humans, who typically write descriptions of entire web sites distilled into pithy one or two sentence descriptions.
As the web exploded in size, labor-intensive directories fell out of favor in relation to algorithmic search engines which could easily scale to catalog many billions of individual pages. But the “open source” model of allowing numerous volunteers to collaborate has saved the directory model.
All of these collaborative directories have a limited scope compared to algorithmic search engines such as Google, Ask and MSN. But limited scope also means that most of the entries in these directories are high quality and you needn’t worry too much about spam creeping into search results.
The Open Directory Project is the original collaborative directory. The ODP is a great resource, but has suffered from years or neglect by its owner, AOL. Newer collaborative directories with fresher, and often potentially better results include Prefound and Zimbio.
Wikipedia, even though it’s ostensibly an encyclopedia, contains so many links to authoritative web pages that it should also be considered as a collaborative directory to the web.
Taggregators, or tag engines, for want of a better word, tend to focus on blog and feed-based content on the web. These services cluster content based on the labels that users have created to describe the content. While tags can be useful, especially for things like photos, music, video, and other content that doesn’t have text that can be indexed by search engines, tags can also be easily misused, with inaccurate, ambagious or even misleading words.
Personalized verticals are a relatively new approach to social search. These services make it easy for anyone to create a specialized search engine focusing on a relatively narrow topic. All you need to do is define the topic area that you’d like the search engine to focus on and then do a bit of tuning once the search engine has built the specialized index for you.
You can also include your own ads in search results for your personalized search engine, allowing you to monetize your efforts and compete, at least on a small scale, with the major search services.
Social Q&A sites
Question and answer sites are yet another category of social search sites that have been around forever. Most public library sites offer “aska” services, which allow you to post questions to librarians and get answers, by email, instant message or SMS. Online forums and bulletin boards have also served the purpose of allowing people to post questions and get answers.
Google started its Answers service several years ago. Want an answer to just about nay question you may have? Simply post the question and offer a bounty from $2 – $200, and you’ll get consideration from an army of Google-qualified volunteers who will research the question for you.
Tagging allows content creators to create descriptive labels for their material, but as I mentioned earlier, tags aren’t always accurate or even helpful for users. Sometimes you want recommendations from people who’ve already consumed the content and have found it interesting, useful or entertaining.
Collaborative harvesters are a relatively new breed of tools that tap into the collective wisdom of their users. When a user of one of these services finds something interesting on the web, they nominate it for consideration from other users. People then “vote” on the content, and any content that garners enough votes gets pushed forward as a recommended source for the rest of the community.
Among the more popular of these collaborative harvesting services are Digg , Netscape ,
Reddit and Tailrank . If you really like the idea here, a site called
popurls.com aggregates all of the top recommendations from these and other web sites including Flickr photos, YouTube videos and others, all on a single page.