Dilbert and Search’s Strategic Options

September 21, 2009

I was waiting for the exercise bike at the gym in Harrod’s Creek, and I flipped through the Sunday comics section. The Dilbert cartoon set forth three business trajectories:

  • A long slide to oblivion
  • A death spiral
  • A cartoon figure getting sucked into a toilet.


Copyright, Scott Adams, 2009.

My instinct was to map search and content processing vendors to each of these Dilbert strategic options. I have to tell you my 10 fingers want to whip up a table and populate it with search vendors who, based on my research, fit into these categories. I hate talking to my attorney. He charges more per hour than I do, so I decided to go another direction.

I want to talk about broad sectors of the search marketplace. I have a nifty chart that shows the 14 major sectors. I have plugged companies’ products into the graphic, so one can tell at a glance what company is moving where. Another graphic shows what specific products compete in a particular sector. Quite a few interesting observations emerge when you look at the comparison of Endeca and Vivisimo in terms of Autonomy, for instance. Another learning emerges when one looks at customer support and the search vendors who are chasing that sector.

What seems semi-useful and reasonably attorney free is a look at sectors mapped to one of these Dilbert options.

One of my sectors is basic search. In general, basic search allows an organization to provide key word indexing and maybe some very basic point-and-click features. An example would be looking at a result list by date, file type, or author. This sector is “getting sucked into a toilet”, to use the Dilbert phraseology. The good news is that an organization can download an open source search system (Flax or Tesuji.eu), use what the vendor includes with some other software (MOSS), or use a credit card and buy one of the lower-cost search systems; for instance, Gaviri, which also offers more robust versions of its interesting technology. The challenge the vendors face is that free puts significant pricing pressure on organizations. The economy described on CNBC and by some wild and crazy economists is still struggling. With search at best a utility, price pressure and a general sense of “been there, done that” means that basic search in the aforementioned Dilbert mode.

Another sector is the real time business intelligence entering a death spiral. Here’s why. The purpose of most enterprise software and systems is to deliver actionable information. The notion that “business intelligence” is a separate operation manned by expensive experts goes counter to what users want; that is, answers. The problem is that the “search” does not deliver answers. The systems that promise to mash up structured and unstructured data are exactly what organizations do not want—yet another expensive and separate system. Quite a few vendors are communicating that their systems’ ability to take information from a database and unstructured and semi-structured information from public Web sites  amuse me. Transformation and query processing are complex for this type of system. The winner is this sector will probably come from an unrelated technical sector; for example, advertising.  I have quite a bit of skepticism when I hear the vendors in this sector promise “actionable intelligence.” Some systems are close, but others, no cigar.

The third sector is the long slide to oblivion. The giant database companies that include search with their products are a good example. I find it painful to use the quasi open source search system provided by IBM. I am confused with the poor performance and the slightly off kilter results I get when I run a query on the Oracle Web site. Companies doing search as an after thought are likely to find themselves taking that “long slide to oblivion”. In fact, most organizations with search as a “sort of” business are on the way down and out. Why? Search is more than an access method. Search is the core of the data management and work processes that are increasingly in evidence.

In retrospect, I wish I could prepare for this Web log a table putting vendors in a table like the European search vendors listing we cobbled together last week. That type of detail is not Web loggy in our business environment.

Stephen Arnold, September 21, 2009


Comments are closed.

  • Archives

  • Recent Posts

  • Meta