Enterprise Search Summit 2007 — Endnote

Stephen Arnold delivered the endnote at the Enterprise Search Summit, held in New York City on May 14 and 15.

The format of the session included remarks by Mr. Arnold, followed by comments and rebuttal from Robert Peck, Managing Director, Bear Stearns, and Sue Feldman, VP, IDC. The audience offered comments and asked questions during the 50 minute session.

The highlight of the session from Mr. Arnold’s point of view was a question from a 20-something, perspiring Google employee. He asked Mr. Arnold, “May I have a video of your remarks? I was caught in traffic and missed your speech.” Alas, Mr. Arnold’s analysis of enterprise search and his opinions were captured by Dan Farber in his ZDNet blog. Click here for his summary: http://blogs.zdnet.com/BTL/?p=5064

Below is the text from which Mr. Arnold delivered his introductory remarks.

We’ve spent two days and heard from more than 24 speakers about enterprise search. What I’ve learned is that there is no such thing as enterprise search and what search implementations exist are in crisis.

Let’s look at my interpretation of what I will call “the crisis in enterprise search”. For me enterprise search is the service that indexes some organizational content and maybe other types of information. In most cases, enterprise search focuses on unstructured information. Some systems handle the structured data tucked into database tables. Regardless of the mix of content, the systems in organizations don’t work very well. In fact, our research shows that most organizations have five or more enterprise search systems and none of the organizations we studied have indexed all of the content that could be indexed. Who wants employee compensation, health data, and confidential agreements in a system that may be compromised, crash, or not work reliably.

Why is there a crisis? There are five reasons:

1. IT in enterprises is in crisis. Existing staff are overwhelmed, underpaid, and overworked. Search--arguably one of the most complex computational tasks--just adds complexity to the IT department’s work load. Once a system is up and running, the users learn that search is not what is needed. Who can search for something if you don’t know what you want, need to answer a question, or use the keywords in a document containing an answer. If the answer is in a video or podcast, most enterprise search systems can’t index these documents, so the user has to rely on brute force methods. For most enterprise users a list of results is * not * what’s needed. Answers are. IT departments are lucky to get the search system up and stable. Niceties like answering questions are a lower priority than indexing new and changed content. Many IT departments rely on vendors to handle enterprise search. While this may reduce short term stress, it causes additional management work to fall on already heavily laden shoulders. There’s no resolution to the IT crisis in view. Enterprises don’t trust hosted solutions, don’t have the resources to invent a search system, and don’t have staff able to deal with the scaling and performance issues an enterprise search solution brings to the IT department’s cubicles.

   The result? IT staff can’t cope. Companies are in danger of losing market share or being forced to sell or close their doors. Is anyone prepared to argue that in most organizations, IT is a healthy, vibrant and respected unit of the organization?

2. Costs. Enterprise search is a famished monster that feeds on money. Enterprise search costs are not well understood and when they are figured out, the costs are difficult, if not impossible, to keep in check. CFOs will be forced to intercede and look for ways to control, then stabilize IT costs. Forget improving existing systems. Accept replacement of existing systems with one that fit within the company’s resources. Services to fund making a system work. Anyone care to name a vendor with this business model. The cost crisis is a time bomb for some vendors. If you dig through the financials of publicly-traded companies dependent on search, you may find what I discovered. Customers, angry with their new search systems, aren’t paying their bills. Cost overruns combine with unhappy customers to fuel this “crisis” within the enterprise and within the vendors themselves. It’s an exciting time to be involved in enterprise search from the accounting angle.
3. You. A general failure to develop requirements, do thorough head-to-head testing, and manage the vendors. A lack of ownership creates an environment of sloppy work. Does anyone want to offer a case to refute that most organizations have five or more “enterprise search” systems? I don’t want to dwell on this beyond reminding you that you need to learn about search technology. Conferences that offer presentations that skim the surface or feature speakers who say, “I haven’t spent much time on this diagram” will prep you for blipping burgers at McDonald’s, not soaring through the executive ranks.
4. Vendors. It’s a painful fact. Vendors take advantage of customers who do not follow procedures to ensure that what’s bought does what is required. Overcharges are part of the business model. Litigation is difficult because the documentation and knowledge required to Pravda vendor error is not available. Vendors have documentation, so your company pays through the nose. Is there a vendor in the audience who wants to refute my assertion that scope changes and unnecessary consulting fees are not increasing in popularity among some vendors? You won’t hear me mentioning any names. I do maintain profiles of more than 52 search system vendors, but I don’t have the energy or money to deal with the swarms of lawyers these folks wield launch at me if I nail the top ten specialists in what’s called a “scope change.”
5. Sea change. Management gurus from Aardvark and elsewhere have jumped on the technology bandwagon. If you are a technology company, you are doomed to failure. So there’s a sea change or earthquake in thinking about how to “do IT.” Most companies want to “play safe” when it comes to technology, which is setting the stage for a massive dislocation in the enterprise marketplace. Will your company chug along until a major breakdown occurs, or will your company embrace a more radical solution. Either way, organizations are on an AirTrans flight to Crisisville without a return ticket.

And where’s Google? I’m not going to talk about Google. I talk about an imaginary company called Googzilla. It has two leaders, an imaginary Sergey Zilla and his twin brother Larry Zilla.

This imaginary company doesn’t know where it’s going because it follows the clicks. Let me explain. You and I click on a service offered by the Zillas. The Zillas are good at math. When a service gets a lot of clicks, the Zillas offer more services to feed the need of the users.

Therefore, if enterprise search is in crisis chichi it is) and it implodes (which it will), a vacuum is created, and the Zillas--whether they like it or not-- will be sucked into the void. Those clicks are addictive and the Zillas are hooked on clicks.

The bottomline is that existing vendors are creating a vacuum, and nature abhors a vacuum. The Zilla are a force of nature and will obediently have an opportunity to dominate enterprise search. The Bills don’t agree with me. That’s okay. I’m 64, recovering from a heart attack, and ignore “received wisdom” from PR people.

To the gentle, kind audience I say, “Learn, demand accountability, focus on costs, go beyond marketing hype.”

Thanks for your attention. Bob, Sue, rebuttals?

Spoken remarks vary from my manuscript.