The Disappearing Middle: Liposuction for High-Profile Search Vendors

April 5, 2008

Last week I received a telephone call from a perky MBA working at a European investment bank. The caller exuded confidence about one of the major publicly traded vendors of enterprise search. She wanted to know if I as a contrarian would speak with her.

No problem. I enjoy being contrary–especially to perky young MBAs with exotic accent-tinged English.

I spoke with her. She was confident about her belief that a share price surge for one company was imminent or “coming round the mountain” as we say in Kentucky.

I gave her my opinion that her stallion was a donkey. Her favored company (which shall remain unidentified) had a poor track record in the money and technology races. Furthermore, no company–including her dark horse–was not going to change its performance record quickly, if at all. She thanked me and disconnected without so much as a “merci” with the cute ascending note native speakers add.

Now it’s two days after this call. I am trapped on an Air Canada flight. The seat back video flickers to life, and I see a documentary about weight loss. The hero (not a sandwich) weighed about 450 pounds. After months of effort, the hero ette was a mere slice of the person’s former self, weighing a lean, mean 200 pounds. Amazing what chemicals can do, I thought. I could see that the blubber around the middle was gone. I killed the annoying LCD and looked out at the frozen wasteland that makes northern Canada the inviting clime it is.

Two images jousted in my mind. A swarm of logos–companies in the search and retrieval business were more plentiful than flies in Canberra on a spring morning. The second image was the protagonist of the weight loss video grabbing clients and consuming their IT and search budgets. Inspired I grabbed a napkin and roughed out this table of search vendors dominating the 2001 to 2003 period.

  1. Autonomy and its IDOL engine and Bayesian core
  2. Convera, an amalgam of Excalibur Technologies scanning and OCR system and ConQuest Software supplemented by Semio clustering
  3. Endeca, a structured data system with MBA-energized work flow and eCommerce chutzpah
  4. Fast Search & Transfer, a Web search system retrofitted with enterprise functionality and re branded as the ESP (Enterprise Search Platform)
  5. Verity, the mid-1980s search system with its aging plumbing and a killer token-based security system.

In my opinion, each of these companies product lines in 2003 were stuffed full of moving parts. When opened for analysis, these systems were like plump Bierschinken. My recollection was that the top management in these five companies exuded in 2003 the confidence of a Gaius Terentius Varro before his introduction to Hannibal. I then converted my list into this table in order to see the differences in the market between 2003 and today.

2003 2008 Change
Autonomy Independent publicly-traded company Diversification away from search
Convera Broken up with parts sold to Autonomy and Fast; continues as a vertical search contractor Gone, no longer a factor
Endeca Independent, privately-held, recent infusions of investment capital from Intel and SAP In the SAP sphere of protection*
Fast Search Sold to Microsoft In the Microsoft sphere of protection*
Verity Sold to Autonomy Gone. Owned by Autonomy

* By sphere of protection I mean that a company can expect sales leads and some preferential treatment in certain procurements.

Inspired, I then sketched this diagram. It has four “force vectors” aiming at the vendors who have a high-profile brand as compared to a company like Exegy or Linguamatics, for example.

When I compared my table with the diagram above, I redrew the diagram eliminating the companies who were no longer stand alone players in the branded enterprise search sector. I came up with this representation:

What my diagrams allowed me to see is that the branded enterprise search sector and its remaining vendor (Autonomy with the Verity property), if I am on the right track, appears to be the center of considerable competitive attention and pressure from different sectors. In 2003, there were five vendors in the middle and no real alternatives for enterprise search. Today, we have a different stable of competitors in the race for search dominance.

The thought that occurred to me is that anyone competing with a company that is surrounded–in effect, under siege–needs to be aware of what I call “survival instinct”. There are many stories of warriors trapped, facing certain death. These warriors usually continue to fight, giving its enemies as much grief as possible before succumbing. The idea of surrender is often perceived as repugnant. Death in battle is honorable; life as a loser is not worth living to many ancient war fighters.

Then I had a Hegelian moment. One of those Thesis-Antithesis-Synthesis insights that whip college philosophy professors into an intelletual lather. My idea: Companies move into the branded search space, fight to death with one another, and then end up fighting to the death against the evil doers from all compass points. Today’s market, therefore, is going to be one roiling, nasty, eventful space for months, maybe years to come. Maybe the search wars will become as memorable as the Punic Wars?

Now, I shifted from Autonomy, Endeca, et al. My insight, if true, means that there are significant opportunities to make real money from search. Whre there were five competitors, now there is one. That means that capitalistic nature abhors a vacuum. The branded search segment will support four “replacements”, grabbing market share and visibility once held by those who have retired from the field or fallen into the orbit of far larger companies where their technology and image will be absorbed.

The companies who fill the void in the middle may come from one or more of the four sectors marked on these napkin diagrams. Let’s look at each of these categories, keeping in mind that more detail appears in my just-published study Beyond Search: What to Do When Your Search System Doesn’t Work:

Up-and-Coming Vendors

I have a list of more than 150 vendors providing search and retrieval to the enterprise. Some of these are spin outs from university research labs and have little more than a Ph.D. or two and a Web site. Others are companies with fresh, innovative, and very modern technology. Others are bundles of older technology cobbled together from the ashes of failed search vendors. Yahoo, in fact, uses one of these refurbished technologies for its online system. The idea that crossed my mind is that where there were five branded players, there are now four openings. I think that of these 150 up-and-comers, there will be several who become the next well-known search brands. Ride one of these ponies in the search derby, come in first, win big.

Text and Content Processing Vendors

Organizations with clunky key word search systems want to find cheaper, better, faster ways to connect their employees to needed information. The old joke advising a buyer to pick two from the better-faster-cheaper triad won’t hold. Today’s buyers want all three. This is the equivalent of having cake, eating it, and getting a death-by-chocolate dessert. This sector is quite dynamic, and several companies from the dozens and dozens offering these types of systems could morph into a full-fledged vendor of branded enterprise search. In case you missed the news, acquisitions have been proceeding apace in this sector. Check my Web log posting about SAS Institute’s plans for Teragram here. My instinct suggests there’s money to be made betting on these horses in the search race.

Open Source

Each passing day, Lucene and Flax look more and more appetizing to organizations on a budget. Lucene works, and it scales. Any 20-something IT professional can get it installed, configured and working in a day or two, maybe less. In 2003, open source or “free” enterprise search was incomprehensible. Today, super power IBM uses Lucene in its OmniFind search solution, a version of which is free with Yahoo logos here. With commercial companies offering versions of Lucene without charge with for-fee support available, I anticipate more activity in the open source sector because the open source runners are wild horses and wildly unpredictable.

Stepping Back

I sat looking at my diagrams and notes. Years ago, I wrote an essay about the problem of trying to survive in the middle of the pack. That essay is not in my files, but I recall that its core argument was that a vendor who priced in the middle of the pack almost certainly faced a tough revenue battle. The idea was that the “middle” was a transition between a product that would sell in large volume at a competitive price. The problem with being a low-priced vendor was the threat of a no-cost alternative. In search Google and Lucene have significant advantages because their product appears to be “free”. There are costs, but these costs are easy to disguise in a budget.

The high-end is the land of milk and money. Vendors can charge a premium for the perceived value of their product. Vendors in the middle need to find a way to charge enough to pay the bills, fund research and development, and make a profit. That’s tough to do in the middle with average pricing. There’s a reason some vendors charge $250,000 to $500,000 for a search system via license fees, support, and customization. People grouse about the service Google provides for a $10,000 Google Search Appliance. Exactly what type of personal support will $10,000 buy in today’s economic arena? Answer: not much.

The middle is a treacherous place. Organizations pass through it on the road to high volume low margin success or low volume high margin riches. Getting stuck in the middle is a death sentence for some.

I can, based on my Air Canada flight analysis, offer some observations for your consideration and comment.

  1. Turmoil in the search and content processing sector is going to be with us for the foreseeable future. Mergers, acquisitions, bloody marketing battles, and brazen hyperbole to make a sale are the features of this landscape.
  2. New winners will emerge. If you can pick the winner from the hundreds of stallions in this race, you can make a great deal of money. Forget new markets like mobile search. The tried and true enterprise markets will generate great wealth for two or three vendors. (No, I don’t know who will win. I keep track of the search stallions. I don’t bet on them.)
  3. The enterprise application super vendors are going to put more and more pressure on the search vendors who fill the now vacant sneakers of Convera, Endeca, Fast Search, and Verity. The vendors with “get out of jail free” cards are Endeca and Fast Search. Endeca may end up in the gravitational pull of SAP and eventually be sucked into that company. Microsoft already has Fast Search, so its future is effectively in the control of Redmond, not Oslo. Verity is Autonomy. So is it conceivable that a company such as Autonomy may find itself alone in the exposed center. Could this be a digital Thermopylae?
  4. The text mining sector may follow the same path as the big five search vendors of 2003; that is, a few emerge as the dominant vendors, then attrition kicks in. Text mining is on the fast track to commoditization in this scenario.
  5. Open source may never win over the Fortune 50, but it’s going to get a hard look from every start up with a computer geek on the team. Why pay for pain when you can get search angst for free? This argument makes sense to the 20-something entrepreneur.

My Air Canada flight landed, skidding on the tundra. I tucked my napkins in my pocket and thought about a big cheeseburger. The table, the diagrams, and the turmoil in the search market put an edge of my appetite. I wondered where the person in the get-slim video got his hamburgers in the airport.

Stephen Arnold, April 5, 2008


Comments are closed.

  • Archives

  • Recent Posts

  • Meta