I've had a depressing month, despite spending time in lovely Seattle and San Francisco. I've been finishing the text of a book on how to implement and manage search applications for websites and intranets, as well as running workshops for the Nielsen Norman Group on the same topic. I wanted to show some examples of good search implementations, but finding any was almost impossible. Even the search software vendors seem to have paid no attention to search usability. One particular problem when you are working in the intranet sector is being able to demonstrate intranets in action, so I have to use websites as a surrogate.

Not that intranet search is at the pinnacle of usability; quite the opposite is true. All too often search software (or more usually a Google search appliance) has been implemented to provide a solution to failing information architecture. The result is a total disaster. Rather than making search effective, organizations opt to re-engineer the brains of employees instead.

The supposition is usually that an employee undertakes a search to find out something that they had absolutely no prior information about, which is rarely the case. We use search even when we know the answer to a question. That is because we like to have the assurance that we have not been overtaken by events, and we can go into a meeting confident that we have found all the information that arose out of a particular project. Certainly the core function of search has always been to help us add to the knowledge that we have.

Especially at my age I can't remember what I know. Memory is far from perfect, but often we need a pointer of some sort to jog our grey cells into action. Finally, the biggest challenge of all for a search engine is to help in situations where we don't know what we know. Amazon trades on this in a very clever way by suggesting books that other readers of the book of our choice have read, as a way of informing us about things we did not know. Clustering and visualization technologies can help alert us to areas of information and knowledge that we did not know existed.

Each of these four requirements has to be delivered through the search interface. All too often the emphasis seems to be to deliver the maximum number of relevant documents on the first page of search results. Leaving aside (for this column) any discussion about relevance, the issue in intranet and enterprise search is not the provision of a Google-like instant search gratification. It is about providing a range of parameters that enable the searcher to move on beyond the initial results list and develop and refine their search query. The process is like going through a shopping mall looking for a birthday present, with each shop suggesting new ideas, rather than looking at the store directory for the Birthday Present Shop and being disconcerted to find that there is no such store.

On most intranets, the search box is hidden in the top right hand corner of the screen, as though the designers are almost embarrassed by what will happen if it is used. What search does display very easily is the quality of the content. Incorrect, misleading, or missing titles are quite common, as are the duplicates of internal memos that every department has put up on their intranet, which now occupy two pages of search results.

This of course assumes you can read the search result. Companies have gone to some lengths to meet the Web Accessibility Initiative guidelines and then thrown them out of the window when it comes to search results, where (for example) it is often difficult to enlarge the font size. And as for those really useful dynamic URLs that are tens of characters long—do they really assist the search experience?

I have to stop now and go and get my blood pressure tablets. Put down this copy of EContent, run any search on your intranet, and ask yourself two questions. First, is that the best that the software is capable of delivering? Second (because I know the answer to the first one) don't your employees deserve something better? I know the answer to that one as well. Searching is about finding.