What's in the Zone?

The Next Big Thing in Online Search

Posted by indroneel on March 1, 2007

Online information is growing at an unprecedented rate with more communities and users taking to the Web than ever before. In the face of this explosion, the top three search services (Google, Yahoo! and MSN) are encountering serious challenges to maintain the relevance of their results. New solutions are constantly being proposed to augment (and at times replace) the capabilities of these services. In this article we shall take a closer look at some of these approaches that have the potential to grow into the next big thing in the search technology space.

Visual Search

Visual search techniques utilize the user’s ability to specify and select items visually, combined with a keyword-based search. The visually selected items are used to further refine the search and arrive at the desired result. What makes this approach really appealing is the usage of perception over descriptive communication as the primary input for information retrieval.

Under the hood, visual search combines optical recognition techniques with descriptive metadata to build relationships among similar looking items. Popularized by Like Visual Shopping, this approach is expected to find wide acceptance among online stores dealing with durable and non-durable personal items.

Information Clusters

Information clusters are more of a “post-processing” feature — they address how search results are presented to the user rather than the techniques used in searching. Instead of returning just the first page of results for thousands upon thousands of hits (most of which are irrelevant and not query-specific), the search results are organized into categories and sub-categories effectively compressing the entire result set into one visual page. These “clusters” are created dynamically and organize the result set for easy drill-down and navigation.

You can think of information clustering as Web directories in Yahoo! or Google tailor-made for every keyword-based search performed. The differentiating factor here is the algorithm and logic that power the dynamic categorization with a high degree of relevance.

Information clusters are brought to you by Vivisimo Inc. Check out the Clusty Internet search portal for a live implementation.

Community Driven Search

Social or communal search uses human judgement to improve the relevance of results aggregated from top search services. The human judgement is captured beforehand using a variety of transparent (such as click-through profiling) and non-transparent (such as tagging and rating) techniques. Even though social search has been around for quite some time (the Open Directory Project being a good example), the novelty in this approach are the new Web 2.0 features that are being used to capture the human factor information.

Traditionally, it has always been the practice to present the search aggregator as a single entity (service or portal) to the end-user. A recent breakaway from this model is the Eurekster’s Swiki service that allows users to define and host their own search aggregators with little effort. Each aggregator maintains its own record base of human-factored meta information for relevance boosting. With discrete groups of users attached to specific aggregators, the swiki model is a definite move towards better integration of search with community powered portals and services.

Human Assisted Search

In the absence of a controlled vocabulary for the Web, automated searching does not return useful results for niche subjects and for broad or confusing phrases. While computer cognizance of the spoken language constantly improves, there continue to be gaps. Human intervention, in the form of real-time interactions with “expert” guides, have the potential to fill these gaps.

The success of human assisted search is derived from the expertise of the guides involved. This is pretty much subjective since expertise and its relationship to a context are something that cannot be measured on an absolute scale. Even as this expertise increases over a period of time, the time taken to arrive at the relevant results is still significantly higher in comparison to that of a more automated solution. Human assisted search is therefore expected to gain more acceptance as a premium service. It should be used only in cases a regular search service fails to return the desired results.

Check out ChaCha for a practical demonstration of human assisted search.


One Response to “The Next Big Thing in Online Search”

  1. Neil, nice coverage, very right about search. Search is not solved. According to Jupiter, 41.2% of users report that general search results are often not directly relevant to queries. 18% of users leave a search engine without having found the information they were seeking. Social search addresses relevance head on. After “on-the-page” and “off-the-page criteria”, Web connectivity and link authority, relevance is now increasingly augmented by implicit and explicit user behaviors, social networks and communities. Social search adds a new relevance layer to information retrieval. After all, we are trying to solve user intent more than content connectivity. This new human, social network and community layer adds context, freshness, significance and smarts to information retrieval.
    The wisdom of crowds – so well articulated by James Surowiecki – is at the root of emerging information retrieval tools such as collaborative content harvesting, directory building, voting and ranking.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: