Skip to main content

Enterprise application integration, a software sector once seen as a potential hot spot, is faltering, turning "pure-play" EAI vendors into the walking wounded.

In recent weeks, three EAI vendors --







SeeBeyond Technology


-- warned of missed quarters; a fourth,



, is facing delisting from the


; and a fifth,



, is losing money as a result of weak sales and an expensive fight to defeat a hostile takeover.

What's more, these companies have been losing senior staff to an aggressive new player in the integration space --

BEA Systems


, whose core application server business is in danger of becoming commoditized.

Since BEA began moving into the EAI space last year, more than 90 staffers from pure-play companies, including top technology officers, engineering vice presidents and senior marketers, have joined the San Jose, Calif., company,

has learned.

To be sure, all technology companies churn through staff, and some of the former pure-play staffers joined BEA after being laid off. And there's not much the EAI companies (or anyone else) can do to end the slump in IT spending or avert a war.

Nevertheless, there is money to be made in the EAI space. Nearly every survey of IT wish lists ranks integration as a top priority.

But that isn't translating to higher sales for the companies that specialize in it. In the first quarter of 2002, the combined revenue of the five pure-play companies was $221.85 million. A year later, that had slipped by 11.7% to $195.96 million.

But war and IT doldrums alone don't explain the sagging fortunes of the sector. Simply put, EAI vendors have overpromised and underdelivered. "EAI doesn't work as advertised. It requires a lot more work than initially depicted," said Rob Tholemeier, an analyst with research house Ramberg, Whalen.

Even people within the integration industry admit that customers are all too often disappointed by the results. "We're back to old-fashioned IT budgets -- no more technology for technology's sake," says one technology executive who left a pure-play in frustration not long ago. "Customers really want solutions -- and they haven't been getting them."

Anecdotally, it is estimated that only one in four EAI projects are completed successfully. That's probably too low, but there's no doubt EAI is getting a bad name. "The whole sector is struggling. Integration is a tough issue and customers are tired of projects that never seem to end," said analyst Katherine Egbert of C.E. Unterberg, Towbin.

And as prospects shrink for the companies, the list of departures has lengthened. In recent weeks at Tibco, for example, five senior executives in engineering, marketing, operations and accounting have either left or told the company they will do so in short order. It isn't clear how many, if any, will land at BEA.

In a written statement, Tibco marketing chief George Ahn said his company's "attrition ratio is significantly below industry average. We are proud of the fact that we have the best and brightest employees among Silicon Valley companies and have in fact buttressed our staff with top-tier hires recently." He declined to elaborate.

Tibco, seen as a leader of the sector,

warned of weaker-than-expected financial results on March 5, followed by webMethods and SeeBeyond a month later (both Tibco and webMethods later reported revenue that was down year over year and below original estimates).

At least four former webMethods vice presidents (plus numerous other former staffers) are now at BEA, including former business development vice president Chet Kapoor, who now runs BEA's integration business.

Is BEA a serious threat to the pure-plays? "It's

the early days yet," said Egbert. "But BEA has an interesting approach." In the January quarter, Egbert noted, BEA logged just $20 million in integration-related revenue. But that's about half as much total revenue as SeeBeyond reports in a typical quarter, while BEA logged revenue of $249 million in the January quarter.

"The pure-plays missed a big opportunity," said Trip Chowdhry of Mercury Research. "They've stayed with proprietary platforms that add complexity to already complex problems."

BEA, he says, is moving toward an integration strategy based on standards (in particular J2EE-JAC) that are much simpler for developers to use, and consequently cheaper to deploy. "BEA will compete very successfully with the pure-plays," said Chowdhry, whose company does not have a banking relationship with BEA.

Of course, BEA may not succeed in solving the problem, either. Tholemeier, for one, says BEA's approach is, at bottom, no different than that of the pure-plays.

Maybe so. But new approaches, including the use of Web services and database-level (as opposed to application-level) integration show promise. And it's hard to believe that enterprises will give up the attempt to make disparate applications and databases work together with fewer hassles and less expense.

Whoever gets it right is likely to make serious money.