I was asked the other day about companies who specialise in providing advanced search or matching technologies for web sites i.e. products or services to help improve the 'findability' of your own products or services for users of your site.The 4 main players that we most often comes across within e-commerce are:

• Endeca - http://endeca.com/
• EasyAsk - http://www.easyask.com
• Fredhopper - http://www.fredhopper.com/
• Verity - http://www.verity.com

Others that we're also aware of include:

• Working Learning - http://www.workinglearning.com
• SLI Systems - http://www.sli-systems.com
• Matching Systems - http://www.matchingsystems.com
• Match4J - http://www.match4j.com
• WCC Services - http://www.wcc.nl

Any others out there? Anyone got any experience of implementing such technologies and can comment on how well they've worked?Ashley

Here at Coherence Design (www.coherencedesign.com) we are currently putting the finishing touches on a matching application based on the Match4j match engine technology. The value proposition can be broadly described as follows:

• Current search facilities within databases or websites tend to produce a large number of results, often irrelevant, despite applying sophisticated searching techniques.
• The matching solution from Coherence Design eliminates the problem of either too few or too many irrelevant results. It’s also non-exact, which means that it will also pull in results that match closely to the request.
• The more criteria specified in the search, the more specific and targeted the results are. The results are scored in relevance order and uniquely you can drill down to see how each individual criterion matches.
• The application is powered by an underlying matching engine technology which is complementary to existing applications, databases, and technology platforms. This is all configurable by non-technical domain experts, lowering the costs and increasing the speed of implementation.

Whilst matching is not entirely new, it is early days regarding its positioning to the right of advanced searching and artificial intelligence based solutions. In response to some obvious questions about how is this different to Google and other search engines the following goes some way to help highlight the failings of search and benefits that matching can bring.

• Firstly Google is keyword based and exact - you have to know what term to use and it will find exact matches, but not near matches. It also ranks results using an abstract relevance measure based on keyword frequency, number of links and popularity (the equivalent to only listening to music from the major record companies and missing the Indy scene).
• Other search engines use artificial intelligence to indicate a relevance percentage - the scores often don’t reflect the relevance they indicate with no means to explain how the score is calculated.
• The net effect can be frustration and critically loss of confidence, not to mention missed opportunity in failing to find the very information that may have satisfied the request or its intent.

I’d be happy to hear feedback about exposing the problems with current technology and how we might move forward where more real intelligence is applied to complete the job of making things more findable.