Case Studies: April 2009 Archives

One of the LinkedIn groups I belong to has a great discussion started by Tom Burgmans , Enterprise Search Specialist at Wolters Kluwer, a publisher. The group is Enterprise Search Engine Professionals and has over 1,600 members. Tom began a discussion with this question: FAST Technical Users Group? As I read his call to action by the FAST user community, and the subsequent cheers in response from group members, I was delighted to see the swell of support. Here's why.

This is a perfect example of where social tools meet a need. A suggestion I also made as a panelist at FastForward 2009 has emerged spontaneously as a direct result of market forces. My observation had been that the FAST user conference was largely attended by IT folks, and the overwhelming number of keynotes and session topics focused on social tools, not especially tied directly to search either. A recommended call to action directed to Microsoft was that they host a platform of social tools to facilitate genuine user community sharing around the FAST product. The people who most need this are search administrators and content managers who presumably have some governance responsibilities for searchable content.

In Tom's suggestion we see the effective use of a social tool to generate interest among members, a large and focused audience who serve as a great test of the viability of his idea. That is neat!

Almost 30 years ago when I ran a software company, we (the company) organized and ran annual user group meetings in tandem with a large professional conference that most of our customers attended. These meetings were very successful, well attended by 40 - 50% of our customers. Over almost 20 years the group spawned a lot of professional and collegial relationships that gave our small user community a sense of collective investment in furthering the improvement and support of the product around which they met. Efforts to turn over total control of the user group to the community were not successful because, in those days, the infrastructure needed for planning, organizing and running meetings across the North American geography did not exist. My company provided that support mechanism out of necessity.

However, three regional user groups began their own programs to share knowledge, and the entire user community collectively published a "cookbook" of source code for reports that many of the users had built for use with the database application and wanted to share with others.

Today the opportunities for building these communities of practice have a vast number of "free" social tools to employ, so that barrier has gone away. More important, the benefits to the user community are limitless. It gets to drive discussion about the product, share hints, workarounds, and tips for successful implementations. The user community gets to decide what is important, what is needed in the knowledge-base of operational information. It can call for product changes, improvements and use social platforms for galvanizing the community around specific issues.

One of the best outcomes we saw with our own user community was around a visitation day at our offices for customers to meet together to "test-drive" an alpha version of a major new release. We purposely stayed out of the meeting for an extended period. Later we learned that when each had developed a "wish list" of changes and tweaks to the release, some rather marginal choices had died a natural death as a result of the "wisdom of the crowd." This was an ideal scenario for us as a development company because we did not have to disappoint any individual users with a unilateral decision to reject their ideas.

Trust me when I recommend to the enterprise search user community, you will empower yourselves in ways you can't imagine when you join forces with other customers to drive the improvements and success of any product you use and value.

When search fails me, the reasons may be hard to discover as a user but once on the inside of an enterprise I can learn a lot about what is going on. After listening to scores of business case studies, personal experiences and reading about rampant dissatisfaction with search it is discouraging to recognize the simple reasons for most negative outcomes.

Consider this scenario. I was attempting to find the address of the office of a major global platform vendor (one of the largest) that sells an entire suite of enterprise search and content management software products. One can usually find business location information from links on the home page of any corporate Web site or at least from the site representing the division one is visiting. But there was no such link for this corporate site. Then using the "search" box and later the "advanced search" option, trying a dozen variations of the division name, town in which the office is located, and product names I struck out on every query. All paths lead to a page with a single corporate address, or a couple of other remote addresses, and links to web pages that contained no address. Even those pages with addresses had no link to directions. I followed up with queries using Google and these got me back to the same dead-ends. Finally, I found the address through various online non-specific business directories.

This experience lead to a couple of conclusions about why my search failed: 1. The content does not exist; there is no such listing of locations. 2. The search engine is not properly tuned or metadata is not supplied with labels such as "locations," "directions," "business offices," etc. The immediate solution for this case is to ensure that someone with practical business sense and usability competency has ownership of the overall web site experience to make sure that essential company data is available and easy to find. Or, if the company has made a conscious decision not to publish that information, at the least they should have a page stating the alternative for potential visitors as to how they can find their destination or to what office they can direct postal mail.

I had to two reasons for needing this information; one was a visit to an individual who was not available to give me the address in time to reach the office, and the second was a personal follow-up letter after someone from the company had been a speaker at an event I chaired. As things stand, I have been left with personal skepticism about the commitment of this company to build, produce and actually use content management or search products that will be truly responsive to needs of their potential buyers. When you don't or can't showcase your products, I question "why." This is not a technology problem; it is a human factors and human resource allocation problem.

This brings me to some search fundamentals:
• No content - If content that customers or employees expect to find is not included in explicit directives to the search engine for the repositories to be crawled and indexed, it will never be found.
• No metadata - Any content lacking explicit language likely to be used by a searcher will probably not be found if it also lacks sufficient metadata.
• Poor indexing or search rule base - If the content being searched is business documents without many unique contextual "hooks," such as product names, technical terminology or topics of narrow interest, the search engine being used must be "smart" enough to glean the intent of the searcher from the context of query. In my case, I supplied a half a dozen terms to layer the context, tried them in different combinations, with and without quotations around phrases, but nothing worked.

Conclusion, if you really don't want searchers to find what they want to find, it is not hard at all to compromise findability. I will not arrive at my destination and you won't get any first class letters from me.