At an analyst dinner with some industry bigwigs last week, the conversation mainly focused on the arrival of Bing (which does not stand for “but it’s not Google”) and the CPC (define) across a number of verticals.
The conversation really got interesting when it shifted to social media and the notion of social search.
This is an area of research I’ve been steeped in for quite some time now. Having lived with Google’s (and others’) general purpose search with its non-validated 10 blue links for so long, I’m intrigued by the idea of validated search results.
Let me clarify what I mean by non-validated results. Google and other search engines rely heavily on a connectivity server and link analysis of the Web (or at least the fraction of the Web they capture). When you search for something, a mathematical formula based on hyperlink data ranks the top 10 results for you.
This means the results are based primarily on Web pages that have the best linkage data according to each search engine’s own criteria. But this is not based on the entire Web, just part of it. Results aren’t verified or validated by human beings, other than by click-stream data. So how do you really know if these results are genuine or even truthful? There’s plenty of evidence online regarding black-hat SEO (define) and the ability to manipulate search engine results.
This is why social search is rapidly emerging as a key paradigm on the Web. As the term suggests, social search deals with search within a social environment, perhaps more easily described as a community of users actively participating in the search process.
The online world has become a very social place rich with people interacting with each other in various ways. And this provides new and unique data resources for search systems to exploit — along with a ton of privacy issues to be considered. Tagging and searching within communities are two main areas of research. Other examples include P2P (define) and metasearch.
Searching within communities examines how users search within such environments. For instance, a musician who enjoys golf and photography may well be a member of music, sports, and digital camera communities. Interactions in these communities range from passive activities, such as reading Web pages, to interactive, such as writing a blog or participating in online forums.
Being able to automatically determine which communities exist in an online environment and which people are members of each will be very valuable information for search marketing’s future. A rapidly growing area of social search is that of question answering. These are systems (frequently expert systems) whereby a person posts a question and community members provide answers to the question. This type of search task is much more social, interactive, and focused than a general-purpose Web search.
Filtering and recommendation systems have been around online for some considerable time. As social search gains popularity, these systems become very important because they are attached to a person’s profiles. For example, in standard search tasks Google returns documents in response to many different queries. And these queries usually correspond to short-term information needs. With the use of filtering, a fixed query represents a long-term information need.
For example, a number of online services provide alerts. CNN, for instance, provides alerts allowing users to specify various topics of interest, such as international politics or sports. When a new story matches the user’s profile, the system sends an alert. In this way, the user doesn’t need to continually search for relevant articles. Instead, the system is tasked with finding relevant documents that match a person’s long-term information needs.
Recommendation systems work within peer groups to attempt to predict how much a person would prefer certain items. Recommendation systems are social search algorithms because predictions are estimated based on ratings given by similar users, thereby linking people to a community of users with related interests.
For those interested in the future development of information retrieval on the Web, including the emergence of social search, I strongly recommend, “Search Engines: Information Retrieval in Practice.” It’s not a bedtime read, but it is a great reference for the design and implementation of search engines and where search is headed.
Join us for a one-day Online Marketing Summit in a city near you from May 5, 2009, to July 1, 2009. Choose from one of 16 events designed to help interactive marketers do their jobs more effectively. All sessions are new this year and cover such topics as social media, e-mail marketing, search, and integrated marketing.
When you’re just starting out as a business owner it’s easy to become wrapped up in the seemingly endless number of metrics ... read more
Visual search on the web has been around for some time. In 2008, TinEye became the first image search engine to use ... read more
We’ve written an awful lot about Google’s open source accelerated mobile pages project (better know as Google AMP) over that last 12 ... read more