My post Why Can't You Just Make it Work Like Google? last week surprised me by going viral. Well, as viral as a blog post about information management can go. It certainly seems to have struck a nerve with people from all across the legal industry. It turns out that making search work effectively inside the organization is something a lot of people are attempting to tackle. After posting it, however, I realized there is also a reason why you would not even want to use Google as it functions out on the Internet for use inside the organization.
Allow me to recap: documents or information or intranet pages inside the organization cannot benefit from the "recommendations" of links from other websites, unlike the way Google search results are created.
However, one big reason why you would not want to use Google as we do on the Internet for searching inside your organization: the results it pulls up are just "good enough". That is, unless you have a specific webpage or document name to work from, you may be getting inaccurate results.
You are pulling up results that do well in the search results rankings according to Google's search algorithm, but this does not necessarily mean they are the exact results you need. You are seeing results that are pulling up as a result of all that linking from outside, and possibly from other Search Engine Optimization (SEO) tactics being used by organizations to drive their results up in the list.
Even if you are an expert researcher well versed in all the tricks of searching Google, it can be pretty darn difficult to run an accurate search. That is, you may be looking for the best law firm in Winnipeg, but quite possibly that law firm could be falling on pages 2 or 3 (or 10 or 12) of the Google search results. And who among us spend the time to click through to page 2 or 3, much less 10 or 12?
Compare with the searching we do at work inside our organizations: we look for the exact answer, the exact document, or ALL the documents that follow a specific criteria. Unless you are looking for a "quick and dirty" result such as a sample letter, there is little room for fuzziness.
If you are looking for all the documents created on a client file, or all documents on a specific subject, or all documents to be pulled out of the system for the purposes of destruction under a records management program, quite frankly "good enough" is not good enough.
One solution that we see floating around frequently is the idea of full text searching — "We just need a search engine that is powerful enough to search the full text of all the documents." This would be nice if we worked in organizations that only produced a few hundred documents. But in the case of legal organizations, lawyers can produce millions of documents. Searching full text in this case quite likely means sorting through hundreds or thousands of documents that pull up in the search results.
What do do? Again, it seems that indexing the documents as they are added to the system is the best bet. Which means creating some sort of classification system or list of index terms or taxonomy is required.
A lot of people offered up their thoughts for alternatives to me on Facebook. What about you: what do you think is the solution? Do you have search working well in your organization? How is this being achieved? Please, I invite you to share your thoughts with us.