Enterprise Search: Cruising on the Concordia
January 19, 2012
I keep my eyes peeled for useful management examples. Whilst recovering from a minor hitch in the goose liver, I watched the drama of the Concordia cruise ship unfold. The horrific event reminded me of several enterprise search deployments I had analyzed. I was not the “captain” of these enterprise search voyages. I was able to do some post-crash analysis.
To get the basics of the event, you will want to familiarize yourself with the write up in the UK’s Daily Telegraph, “Concordia Disaster: Should a Captain Go Down with His Ship?” In my opinion, the key passage in the Daily Telegraph’s story was:
…leadership entails an obligation to be courageous – morally, physically or both. It is the price of leadership; it is why leaders are more highly regarded and rewarded than the rest of us. But even subordinates in certain professions have the duty to be brave, as the rest of us do not. A soldier is expected unquestioningly to put himself in the way of bullets as a civilian is not.
(But my favorite news item was Cruise Captain Says He ‘Tripped’ Into Lifeboat, Couldn’t Get Out.”
Not Taking Responsibility
The alleged behavior of the captain shares one similarity with enterprise search implementations that sink. The person running the operation shirks responsibility for the disaster. My view is that ego plays a part. The more important factor may be the person’s character. I have reviewed a failed search implementation and had a difficult time determining who was responsible. The procurement team has the thick linen of committee think under which to hide. The information technology manager often keeps well away from search, a behavior conditioned by knowledge that making information findable is often impossible. The chief financial officer just counts the dissipated dollars. Accountants are not implementers. The person charged with the failure is often a young engineer whom those ultimately responsible deem expendable.
The first similarity is that in big disasters those who are responsible do whatever is needed to avoid responsibility. In enterprise search, there is a ship captain. Pretending that a captain does not exist is one interesting characteristic of today’s organizational life. Think Jerry Yang at Yahoo. Recall Leo Apotheker. You get the idea. What about the search system at your company? The National Archives? Amazon’s online store? There are captains responsible. Unfortunately these captains do not get global news coverage for their behavior.
Show Boating
The crash and sinking was a consequence of show boating. The idea is that doing something fancy is appropriate and within the perimeter of the job description is allowed. In enterprise search, the show boating becomes visible when one or more people make suggestions along these lines:
- We need to deliver answer to users, not laundry lists
- Natural language processing is essential to the success of our search system
- We need a taxonomy and semantic technology to make information accessible
- Our system has to work just like Google.
Each of these is similar to the Concordia’s buzz close to shore. Few of those involved in an enterprise search implementation realize how downright expensive, complicated, and resource intensive these “suggestions” become. Vendors go along to keep the contract. The deployment team is thinking about making search headlines and maybe getting a raise and a promotion. Great idea but when the effort sinks the search project, the result is a disaster.
The second similarity between the Concordia and the ill fated enterprise search system deployment is that getting cute can wreck havoc. Now you may say, “Hey, semantic methods will only help our search system.” Maybe, maybe not. My view is that show boating is one characteristic of doomed enterprise search system. The fix? Just do the basics well, then add some special sauce.
Exploding Costs
The cost profile of the Concordia problem is far from complete. The broad outlines are visible. There is the cost of clean up. There is the cost to the Israeli-American owners of Carnival Cruises. There is the cost of the ship. And there is the cost to the cruise industry itself at a pivotal time in the booking season. I can summarize other costs, but you have a base upon which to plan the amount of dough the captain of the Concordia placed at risk with his alleged course adjustment.
The third similarity between the Concordia’s sinking and enterprise search is that the cost of a failed project are far greater than those involved perceive. The license fee is irrelevant. The stakeholders of the company with the albatross have to pay for remediation, lost opportunities, and possible increased risk associated with regulatory or legal requirements. Sweeping a failed enterprise search system under the rug is as difficult as making the half sunken hull of the Concordia disappear from TV news programs. The wreck is too big to hide.
Bottom Line
The Concordia, like most enterprise search systems, should have unfolded in a predictable, acceptable way. Human error, not vendor methods or a specific search algorithm, is the cause of most enterprise search failure.
The question I am considering is, “How many of those involve in an enterprise search project will abandon ship?” My hunch is that there are more Captain Francesco Schettinos than many believe.
Stephen E Arnold, January 19, 2012
Sponsored by Pandia.com
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
