Yesterday I posted "Tweaking My Algo", a brief experience report on using a site search utility to inform content development. Today I would like to give a few ideas on what the optimal site search utility might include and solicit your ideas for additional features.

1) A deep database of both keywords and querries should be maintained. You probably already have a good idea of what the most common querries are going to be and if you plan to use this information to inform content development it would be great to be able to prospect the top 1000 querries and not the top 50 or top 100. For a very large and very busy site this database could grow very large.

2) You should be able to archive the database at any time. This is very important if you install a new site design or even change your persistent navigation. This will allow you to see how your new format is working.

3) Allow you to clear the archive at any time (this pairs with item #2 above).

4) Give you control over the search algo. After experimenting with one of these systems I am very confident that my algo settings would need some tweaking if I moved it to a different site. Plus I like the ability to give heavy weight to specific files or folders. This could be useful if you have different development styles on different parts of your site or if you want to boost the visibility of commercial pages.

5) Allow you to set-up channels so that search querries from different parts of your site can be studied separately. If you have an info site with a store it will probably be profitable to know what the folks in your store are looking for.

What else would be ideal for these goals? Does anyone know of an inexpensive product that does all of these and more that I have not thought about? Any developers thinking about building one?