Enterprise Search: You Know You Are in Trouble When
December 7, 2011
When I was jammed between two less than svelte individuals on a flight from Dallas to Louisville, I read a chubby article called “IT Inferno: The Nine Circles of IT Hell.” My view of the write up was that the author walked to the end of the pier and fell into the water off the shore at one of the beaches south of Rio’s harbor. Yikes.
I then noted “Six Lessons from a Lightning ERP Rollout.” I think the main idea is that if one goes fast, the outcome of an enterprise resource planning project is going to better. Speed does only good, which will bring joy to the wanna be F1 and Nascar drivers in information technology. I found this statement interesting:
ERP implementations have gained a bad reputation, in which merely late is considered very good, and spiraling out of control is considered common. There are always more ways of doing something wrong than doing something right. Beyond that, the act of defining an effort as an ERP implementation contributes to the likelihood of disappointing results.
The idea in these two write ups was a good one. Make some mistakes and end up caught in a Dante-esque world. In theory one could emerge in the land a paradise, but few make it. As far as I know, Dante, whose house in Florence looked over a street in which all manner of technical and non technical activities unfolded, did not think much about information technology, online systems, and Service Level Agreements with $100 billion consulting outfits like IBM. No Watson was in the neighborhood to assist Dante with a thorny question.
I want to ignore the reference to Dante and focus on a handful of ideas in the six page write up. Furthermore, I want to sidestep the generalizations about large scale information system projects in organizations and narrow my focus to search. I won’t even draft business intelligence or the other 13 euphemisms for search I wrote about a few weeks ago. (I picked up this theme and drilled into some specific companies in my December submission to the January or February Information Today for which I write columns for money. Amazing, I know.
So, no big picture, no Dante, and certainly no theological overtones. I remarked at a speech in November 2011 that I once studied and indexed medieval religious literature. Got a laugh. Too bad it is true. Now you know why I poke mercilessly at English majors, failed high school teachers, and self appointed experts. The goose is talking about his guru-ness or, I should say, his being a goose-ru.
So, you know you are in trouble when:
- You cannot control your budget for search. You will know when this occurs because the organizational equivalent of a six grade Catholic school teach armed with a wooden pointer and a habit (the good kind) tells you to stay after class. The CFO watches the money, and she will bring you up to speed on the magnitude of the problem you have created. Numbers make clear that a search project has gone south. I can hear the snap of the wooden pointer now. Ouch.
- Users work around the system. Now most information technology professionals deny that users are unhappy with any aspect of a system which is actually online and functioning to some degree. However, when we ask users about search systems, the message we receive from surveys and interviews is easy to understand: Search is not too useful. The proof, however, is not a survey. Just ask users what they do to find information and you will learn about intra-company networks, bootleg systems, and even hosted services which poke through the firewall to index content. The sticky notes are a dead giveaway as well.
- The infrastructure is not able to keep pace with indexing. We hear a lot of baloney about how fast a search system is, how small an index’s overhead will be, and how little latency exists within a search system. The reality is that most search systems cannot provide near real time index updates. As a result, there are numerous instances of old information used for proposals, financial summaries, and marketing reports. How does a cash strapped search system manager keep pace with burgeoning digital content? Easy. Index less and update the index on a relaxed schedule. Don’t believe me? Where is the marketing PowerPoint for the presentation given yesterday by the marketing VP? I get these materials by asking the person to give it to me on a storage device or dump it in an online file sharing service. That sometimes works. Search usually does not work.
- The vendor changes his marketing tune. Here’s how this goes. You license a search system from a company which asserts that it is 100 percent committed to enterprise search. then you read in Beyond Search that the company is in customer support, information optimization (whatever that means), business intelligence (an oxymoron at Bank of America or Goldman Sachs perhaps?), or taxonomy generation. What happened to search? In the quest for revenues. search vendors change more rapidly than we can track. When this happens, how easy is it to get the exact technical inputs your need without delay? If less than 10 minutes, you have a winner vendor. More than 10 minutes, well, well, well.
- Your open source wizards takes a job at Cisco Systems, eHarmony, IBM, or one of the other Open Sourcey firms. You remember the old space launches. I bet this triggers a memory moment, “Houston, we have a problem.” One certainly does.
Check out the IT inferno. Just make sure search is not throwing coke into the oven as someone increases the oxygen flow. You can have a coal fired meltdown without nuclear powered search. AtomicPR, however, will lay down a radioactive blast zone for its various solutions, including XML as a universal search solution. I have heard that phrase before, “universal search.” Those categorical affirmatives are a sixth indicator I believe.
Stephen E Arnold, December 7, 2011
Sponsored by Pandia.com
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
