• Home
• About
• Writers
• Landscape
• Wizards Index
• Fraud
• Wanna Be Like Larry?

Selecting an Enterprise Search System: The Mid-Sized Company Dilemma

Earlier today (May 5, 2008) I received a telephone call from a journalist seeking my thoughts about this question: How do mid-sized companies select an enterprise search system. As you know, I call this type of search “behind-the-firewall search”. There’s considerable confusion about Web search, search on a particular Web site, ecommerce search, and the other denizens of the search phylum in the kingdom of information within the domain of knowledge. (I feel biological at the moment after a day of considering how Google vapor sucked oxygen from the Microsoft-Yahoo deal.)

This morning (May 6, 2008), my RSS reader proudly displayed an Information Week article penned by George Dearing “Why Is It So Hard to Be Found“. The article was interesting because Mr. Dearing used the phrase “within the firewall” to describe enterprise search. As you know, I prefer the phrase “behind the firewall”, but he’s close enough for horseshoes. You can read this essay here. Click quickly, articles on the pop-up beseiged Infomation Week Web site can be, as Mr. Dearing notes, “so hard to be found”.

The point he makes that stuck me was:

For something so critical to content as search, you’d think that companies would have more to show for it than misguided enterprise search implementations… I’ve always had a hard time getting my arms around the space, much less the application of specific search-oriented approaches…. Sam Mefford, a search consultant with Avalon Consulting, made me feel a little better recently when he told me, “I’m moving away from the terminology Enterprise Search wherever possible, and moving to just Search, because most organizations simply aren’t ready for Enterprise Search.” After talking to him, it seems the challenge for enterprise search is the same as for other enterprise software sectors: A lot of work was put into technology and software development but the needs of users have largely been ignored.

My bandwagon is no longer holding me. Avalon consulting is on board. I was encouraged by the journalist’s call as well. Some are looking at the search vendors’ assertions and seeing the handiwork of PR mavens, not programmers.

In Beyond Search, I provide quite a bit of tactical information for fixing a search system gone bad. But in the 3rd edition of the Enterprise Search Report, I slogged through the formal procurement process applicable to organizations of almost any substance. And what about the journalist’s questions? The young lass had done her homework. She wanted to know about user dissatisfaction (hovering around 60 percent), methods of selecting systems in mid-sized companies (an underserved sector), and the pay off from a good system (it is easier to explain the cost of not having information access).

The business end of a piranha. Imagine a procurement team swimming in a calm Brazilian river. Above the giant search vendors circle like hungry vultures. In the river, a swarm of ravenous up-and-coming search engine vendors want to nibble on the procurement team. Mid market companies find themselves in the middle when it comes to licensing a search system. Big, aggressive folks above and fiesty smaller ones below make it tough for mid-sized firms to make a well-reasoned search system acquisition.

After the pleasant telephone talk with the reporter, I continued thinking about the characteristics of a mid-sized company. I define “mid sized” as a firm with revenues between $50 million and $300 million in revenues. This is a company size that is caught in the middle. Vulnerable to incursions by far larger companies in search of new revenue, the mid-sized company is a tempting target. IBM, Microsoft, and Oracle have signaled an interest in the companies in this sector. Nibbling away like tiny piranha on the toes of swimmers, start ups and small companies with revenues below the magic $50 million threshold want to gobble the swimmers’ calves, maybe the entire swimmer.

In general, occupying the mid-market requires considerable attention to bigger companies and to an innumerable swarm of smaller outfits. With 15 or 18 million businesses in the US, most are smaller. The 2,000 or 3,000 giant-sized enterprises have the resources to prey on the mid-market.

To survive, mid-market companies have to work hard, deliver acceptable customer service, and market effectively. One slip up, and the weaker mid market company is a snack for a larger organization or a feast for a smaller predator.

Not surprisingly, selecting a behind-the-firewall search system boils down to one of three broad strategies. I’ve substantiated these via my survey and interview work. I am on the look out for more anecdotes, including survey data, that can illuminate the interesting world of mid-market companies.

How Mid-Sized Companies Select Search Systems

There are three broad types of procurement strategies in play in this sector. Remember, your mileage may vary. Let me describe each, offer a few differentiations, and conclude with several observations.

The first procurement tactic is to use whatever search system is available from vendors providing enterprise software. Microsoft is a popular platform in small and mid-sized companies because it’s reasonably priced when compared to the industrial-strength solutions from IBM, Oracle, and SAP, for example. SharePoint has almost 100 million users, and I encounter a large number of companies with SharePoint search. What happens is that if the “free” or built in SharePoint search doesn’t deliver, the Microsoft-centric company looks for a Microsoft certified partner with a search solution. There’s a good selection of vendors in this category, including Coveo, Exalead, and ISYS Search Software.

The second procurement tactic is to go with what the senior information technology person or consultant knows best. If an organization has experience with Lucene and a powerful text processing system like that available from Siderean Software, that’s what the IT manager will license. Who wants to learn a new system when one has a base of knowledge and a positive experience as guides? Answer: not too many IT managers.

The third procurement tactic is to go through a formal RFP process. This means writing down requirements, identifying a short list of candidates, getting dog and pony shows, and picking the “best” vendor. The criteria can be just like those in a software engineering text book, or they can be based on subjective criteria such as “the vendor was nice” or “we heard good things about the system from another firm we know well”.

Based on my research, the popularity of these three approaches is above equal; that is, 30 percent of the mid-sized companies use what’s easy and available, 30 percent select a vendor based on what the system manager knows, and 30 percent slog through an RFP or formal process.

What about the other 10 percent? What I found surprising is that the other 10 percent of the firms in my sample bought a search toaster. Google’s Search Appliance, followed by the Thunderstone appliance, are the big winners. But some companies opted for hosted search. Blossom Software, located near Boston, has a rapidly growing business because commercial enterprises and government agencies want to have an outside firm handle the search technology, upgrades, and customization.

What’s This Tell Me?

You may draw your own conclusions from my summary of my research findings. Here’s what I think these data suggest:

1. Mid sized firms take more short cuts than I thought. The idea that a prior experience with a vendor or just using whatever is included in another enterprise system disconnects user needs from the selection process. As long as employees are thrilled with the search system, no problem. If users don’t use the system, then a recompete is required. Time and money are wasted. The IT department’s credibility takes another painful punch.
2. The small but significant 10 percent looking for a search “toaster” underscores the need for simpler, less hassle ways to provide access to information. I’m watching appliances and hosted solutions to see if the 10 percent grows or shrinks.
3. A large number of mid-sized companies look at Autonomy, Endeca, and Fast Search and up and coming firms like Coveo, ISYS Search Software, and Siderean Software. In many cases, the Big Three are too expensive, too complex, and too needy of on going technical babysitting to be a good match for the mid sized firm’s overworked systems staff. Up and coming vendors of search are making significant headway in the mid market. If this trend continues, the Big Three will have to come up with some effective ways to slow down up and comers. A big firm might just buy one or two of these usurpers to “solve” the problem.

Disrupting the mid market is Google. Vendors criticize Google’s enterprise search solution as little better than lousy. I don’t agree. I think the GOOG has a solid platform. Granted the company’s approach to customer service is not much different from the math club laughing at a football player’s effort to solve a ninth-grade algebra problem. But the GOOG has about 9,000 Google Search Appliance licensees, and that makes it a pretty big dog in the search kennel.

My conclusion is that the mid-market may be gestating the real threat to incumbent leaders like Autonomy, Endeca, and Fast Search & Transfer. More than that, the mid-market battles are ludi for search gladiators. A strong mid-market vendor could rise up to challenge IBM, Oracle, or SAP.

Observations

The mid-market is an interesting space for search and information access. The companies have enough money to license, fail, and regroup. The mid market firms, however, cannot afford to fail in search too often. Without cost controls and tools to extract value for their information, the mid market companies are vulnerable to attack. A search vendor able to make sales in the mid market is preparing itself to move into the battle field when the largest 2,000 or 3,000 organizations fight large-scale battles.

How vulnerable are high-priced search systems that allow the upstarts to snatch sales from them? I think that the branded search engines have to win in the mid-market to survive. At this time, the balance of power may be shifting from well known vendors of search and retrieval to more nimble, more technically adept up and coming vendors.

Stephen Arnold, May 6, 2008

Comments

Copyright © 2008 Beyond Search · ArnoldIT.com · Disclaimer · Revolution News theme by Brian Gardner · Log in