« Ruling by committee | Main | Top 5 killer intranet mistakes »

Search sucks, but could it be any other way?

Let's face it, search sucks in most organisations. But could it really be any other way, as things current stand?

Taking a step back for the moment, the question is worth asking: even if an organisation is committed to delivering a good search experience, exactly how are they going to achieve this?

There are few resources and sources of assistance that an organisation can call upon when designing and implementing search:

• The right search vendor is chosen, to match the organisation's needs. So far, so good.
• The search engine comes in a shrink-wrapped box, with an install CD, and some technical documentation.
• The default install of the search engine is terrible, unusable, over-laden with features.
• Nothing in the search engine documentation talks about how to design the solution. There are no examples, samples, case studies or methodologies. Just a lot of under-the-hood technical configuration details.
• Most search engine vendors provided very limited (or no) professional services to support clients in designing good solutions (except for those vendors who offer products that can't work without professional services).
• There is almost nothing written about how to design good search, whether articles, case studies, reports or books.
• There are few consultancies that provide professional services around the design of search (as distinct from product selection).
• In many cases, organisations are entirely unaware that there is even work that needs to be done to make search work, beyond just installing the software. Primarily, because no-one told them.

So how are organisations supposed to find out how to deliver good search?

From where I stand, this lack of support and resources is one of the key reasons for the poor usability (and usefulness) of search. Clearly something needs to be done. While it is beyond the scope of this blog entry (and my thinking) to propose a complete solution, here are a few practical steps that could be taken:

• Much more needs to be written about designing search, by as many people as possible.
• Vendors should publish case studies, talking not about why a client selected them, but what the client implemented (and why).
• The default (out-of-the-box) configuration of search engines should be improved.
• Vendors should include in their manuals more about designing search, the business benefits, and the methodologies that can be used.
• Vendors should become more client-centric, in both their sales and implementation processes.
• Organisations need to start setting aside a little bit of time and money to design the search solution, beyond purchasing the software.
• The industry as a whole should start talking about how search is designed and implemented, with the goal of identifying some "best practice" approaches.

Posted by jamesr on November 29, 2005 11:50 AM
Categories: Search tools