Late accounting move rocks tech firm, stock falls. (Up Front).

SeeBeyond Technology Corp. became mindful of the groundswell for increased scrutiny of corporate accounting -- three weeks too late.

The Monrovia-based company, which had become the No. 2 player in the market for software that allows different applications to talk to each other, reported last month it would fall $2 million shy of its previously projected revenues of $44 million for the first quarter ended March 31.

After the announcement, Chief Executive Jim Demetriades said that auditors from Ernst & Young advised the company not to record another $2 million in revenues that were booked, but not collected, in the first quarter.

"Post-Enron, what we're learning is that auditors go through things very carefully," Demetriades said. "If they don't want us to record it until we get paid, we won't."

As a result, SeeBeyond reported quarterly revenue of $40.4 million, and Wall Street reaction was severe. The stock fell more than 50 percent, from $6.58 to $3.15, on April 21. The stock closed May 1 at $3.14.

This, despite the fact that SeeBeyond reported net income for the quarter of $2.9 million (3 cents per diluted share), improving on a loss of $4.4 million (6 cents a share) for the like period a year earlier -- and the company's second consecutive positive quarter. Revenues were $40.4 million versus $49.4 million in the first quarter of 2001.

"The pre-announcement was not that dramatic, but when the reported numbers were even lower, in some ways it's a bit dicey," said Erick Brethenoux, vice president of equity research at Lazard Freres & Co. LLC.

Not to Demetriades, who said he learned a lesson by making an announcement based on unaudited books. "There's just no question visibility remains clouded, but we're optimistic recovery will eventually come about," he said.

(Demetriades, one of the founders, has decreased his stake in the company from 40 percent to 30 percent between September 2001 and February, selling 7.3 million shares at prices ranging from $1.40 to $9.01 per share. He said the sales were made in response to a margin call and in an effort to diversify his portfolio.)

Setback after rise

The lower revenues come after a streak of positive news.

After raising $15 million in a private placement with institutional investors in December, and another $70 million in a secondary offering in February, company executives felt confident enough to commit to a 12-year, $25 million lease for 120,000 square feet of office space.

But analysts, wary of the shortfall in announced revenues and mindful of what they consider a shaky market in the company's niche product, known as "middleware," are now less than optimistic.

Jaime Roca, an analyst at Precursor Group in Washington, D.C., said middleware and an emerging integration field known as Web services are converging. The Web services market will likely be the domain of BEA Systems, IBM Corp., Microsoft Corp., Oracle Corp. and Sun Microsystems Inc.

Middleware acts as a conduit or translator between disparate programs used in unison by a business. A small-scale analogy is the interface between a personal digital assistant and PC. The larger the business, the more complicated the task of communicating among these programs.

"Middleware players, especially the small players in the market, might feel some pressure because it's the Big 5 that are leading the race' Roca said.

There will be a battle, and the question is whether SeeBeyond can regain investor confidence and continue to grow. Brethenoux said prospects are not good if the company defers revenues to bolster future quarters.

Demetriades denied that the E&Y advice was aimed at fortifying the numbers going forward. "We don't know that we'll be able to account for that revenue in the next quarter," he said. "We get to account for it as the customer deploys it."

Thomas Berquist, an analyst at Goldman Sachs, said profitability might be difficult to maintain because tough times are ahead for the industry. Software stocks, as a class, are down 30 percent to 40 percent in 2002 after a strong end to 2001.

"The question that people have is about the health of the integration space overall," he said. "Many of the vendors have been struggling as the economy has been struggling, and SeeBeyond had the weakest balance sheet in the field" prior to raising the $85 million.

Building a war chest

Demetriades said the recent cash infusions have established a $100 million-plus balance in the bank, which should allay any fears on the part of customers about the company's long-term viability.

Berquist wasn't so sure. "With the weakness reported in the March quarter, people are worried that they've lost their momentum:' he said.

Brethenoux, of Lazard Freres, said the cash would help SeeBeyond, so long as it remains in the bank and doesn't end up being spent to shore up for lagging sales.

"They have to keep a decent treasure chest to be able to say that 'Even in tough times, we go into the cash we have and we'll be around,'" he said.

The company will fill the increased office space, Demetriades said. Despite the bleak outlook for sales, SeeBeyond will increase staff by as much as 15 percent worldwide, 20 percent in Monrovia.

Brethenoux conceded that SeeBeyond has better than a fighting chance in the middleware market if the economy rights itself. But he warned that Demetriades should be casting about for smaller companies with high-end technologies that SeeBeyond can acquire to keep it competitive with WebMethods Inc. and TIBCO Software Inc.

WebMethods posted a $15.7 million loss in its first quarter ended March 31, an improvement over the like period a year earlier, when it lost $46.8 million. TIBCO reported net income of $3.4 million on revenues of $74 million in its first quarter ended Feb. 28.

John Ederer, an analyst at Pacific Growth Equities Inc., remains upbeat about SeeBeyond's chances of competing once the company's performance helps people forget the first quarter's results. "They ended up doing a lot of damage by pre-announcing one number and reporting another:' Ederer said.

[GRAPH OMITTED]

SeeBeyond Technology Corp.

YEAR (Dec. 31)                    2001     2000

Revenue (millions)              $185.9   $114.8
Operating Expenses (millions)   $203.3   $156.7
Operating Loss (millions)       (17.4)   (41.9)
Net Loss (millions)             (17.4)   (42.7)
Loss Per Share                 $(0.24)  $(0.69)

SUMMARY

Business: Software manufacturer

Headquarters: Monrovia

CEO: Jim Demetriades

Market Cap: $300.9 million

Dividend Yield: N/A *

Total Liabilities: $55.1 million

P/E Ratio: N/A

Long-Term Debt: $1.4 million

* SeeBeyond Technology Corp. does not pay dividends.

RELATED ARTICLE: Avoiding 'Law of Averages'

SINCE the stock market has just suffered two consecutive losing years, many people figure it is due to rise in 2002.

But since the stock market enjoyed two consecutive decades of outsized gains in the 1980s and '90s, many people figure it is due for a subpar overall showing in the '00s.

Confused? So it goes when you try to plot an investment course by what is known commonly as "the law of averages," or in more sophisticated circles as reversion to the mean.

If your goal is to set up a workable investment plan, though, law-of-averages thinking may do you more harm than good. It can lead you into a mathematical trap aptly called "the gambler's fallacy." Even if the markets truly are subject to mean reversion -- which is a matter of some dispute -- trying to analyze them on that basis just muddles the issue.

To see the gambler's fallacy in action, suppose that Stagger Lee and Billy are flipping a coin in a dark alley. The coin comes up heads four times in a row. Considering the long odds against five identical outcomes, Stagger Lee now bets everything, including a hat he has just acquired, on tails. He loses when another head turns up.

If Stagger Lee subsequently shoots Billy, as happens in the song, well, he may be justified in suspecting that the coin was not quite legit. Still, he ought to save some of his anger for whoever taught him math. Our Stagger Lee fell prey to the "erroneous belief in mean reversion when out- comes are in fact independent," to borrow the words of Frank Schmid, senior economist at the Federal Reserve Bank of St. Louis.

The probability of a tail on the fifth flip was no more than 50-50; the outcome of previous flips had no influence of any kind. The coin certainly did n't "remember" them unless it was rigged, in which case the odds favored another heads.

Without any particular scientific proof to back me up, I'm inclined to view the stock market as highly, but not completely, efficient -- difficult, but not impossible, to exploit using analytical skill.

As for mean regression, I'm ready to buy the idea that prices and perhaps other numerical values in investing tend to regress toward equilibrium points, creating a pendulum effect

Sad to say, when we try to put this idea to practical use we find we don't know what the mean values are. We can figure historical market returns, such as 7.43 percent a year for stocks, after adjusting for inflation, from 1925 to 2001. This is not enough information, though, to tell us what a "normal" return for stocks ought to be.

To further complicate the problem, whatever mean regression does occur in the stock market follows no predictable time schedule, In the 1990s, law-of-averagers drove themselves batty trying to figure out when an obviously overvalued market would stop going up.

Mean regression, in short, is a troublesome basis for investment decisions. If Stagger Lee had only known that, he might still have his Stetson hat.

Chet Currier, Bloomberg News