Vertical Search Engines - Specialist Custom Searching

While much of the search world keeps its eyes focused on every move made by Google, Yahoo and Microsoft, many specialised small specialist vertical search engines have sprung up, along with several technology providers that enable them. The world of search is such a dynamic place that you can find major developments taking place right under your nose that just aren't getting that much attention. Vertical search is one area where this is happening.

Vertical search engines that specialise with listing your business will most likely offer you a higher conversion rate than a horizontal search engine, (Google, Yahoo etc) since users will have implicitly self-selected themselves as interested in your custom specialist vertical. PPC (pay per click) campaigns with these vertical search engines should fare better as well. It should prove worth the effort to seek these vertical engines out, and develop a relationship with each one individually.

Outsell reports that the vertical search market will reach $1 billion in revenue by 2009. Enterprise search technology provider Convera's recent survey of more than 1,000 professionals found that only 43 percent of business professionals always find what they want from a horizontal search engine after several attempts, and half of those who don't find what they want will then turn to a vertical search engine to improve their results.

Why is tradional Horizontal Search (ie Google, Yahoo etc) not working as well as it used to?
The Convera study above shows that horizontal search does not always do the job required of it, which is driving the rapidly growing vertical search revolution. The problem is that search engines often do not know what you want based on your search query. For example if your query is "surf" and you live in Sydney, what is it that you really want? Or if you type in "aortic dissection", the search engine does not know if you are a doctor looking for research data for aortic dissection or a patient looking for information on the symptoms of aortic dissection and what to do if they display these symptoms.

Search engines are trying a number of initiatives to address this problem, such as search query refinements that allow users to focus their searches more quickly, such as Yahoo Shortcuts or the Google Onebox. However, these refinements are not entirely scalable, as it requires human editorial input for each query to make them accurate.

Vertical search engines address this problem by allowing the custom design of search engines for a specific purpose. The human input is built in from the beginning. If you decide to use a search engine which says it's a "health information search engine for doctors", instead of one that is labeled a "health information search engine for patients", you have already helped to reduce the ambiguity of your search queries even before you type in your query.
In addition, traditional horizontal search engines cannot always determine the target audience of a page or site. Vertical search engines naturally address this issue, because the sites included in the results have been selected according to more specific criteria, and perhaps even by human input.

So what does this mean for the world of search engine marketing?
From a marketing perspective, there is really no feasible way to approach every person or business designing a vertical search engine. It's simply not a scalable activity. There may be individual vertical search engines that grow to be large to make it worth approaching them directly, but the majority will have to be found in other ways.

Traditional SEO link building will get you nowhere for any custom search engine built on the vertical search platforms, because they rely on humans to decide which sites to include. Among the many vertical search engine start-ups that do not use one of these vertical search platforms, there are some that do their own crawls of the Web. Traditional SEO should help here, although the nature of the algorithms they use will vary from engine to engine, and may not be at all link-based.

(Edited from original article by Eric Enge
January 3, 2007)