A glitch in Y2K

By Ross Kerber, Globe Staff, 6/9/99

Once Peritus Software Services Inc. seemed poised to profit handsomely from the computer problem known as the Y2K bug. Revenue doubled in 1997 on the company's boasts that its software could find and fix problems amid millions of lines of code.

But for Billerica-based Peritus and many other software firms hoping to turn anxiety into gold, Y2K simply hasn't panned out. Since last year Peritus shares have fallen 98 percent, hundreds of workers have been laid off, and two chief executives have quit.

And Peritus is not alone. While Y2K-related spending is estimated to total around $250 billion in 1998 and 1999, much of that money has flowed to large consulting companies that send in armies of expensive programmers to pore over corporate programs.

Big winners include Boston-based Keane Inc., for example, where Y2K orders helped push revenue to more than $1 billion for the first time last year. Sales at International Computerware Inc. in Marlborough were up 20 percent to $12 million last year for similar reasons.

Meanwhile, many others have struggled. In particular, companies that make fix-it programs known as software tools have fallen far short of expectations. In Phoenix, Viasoft Inc. will take a $10 million charge at the end of the year as it moves out of the Y2K market.

In Mountain View, Calif., declining sales of Y2K products helped drive Micro Focus into a merger that formed Merant PLC, analysts say. Shares in Pittsburgh-based SEEC Inc. have fallen about 50 percent since December on disappointing sales of Y2K tools.

At Peritus, executives are now studying a sale of all or part of the firm's assets. Even some successful tool makers, like closely held Data Integrity Inc. in Waltham, have adjusted their projections downward.

"The lesson here, both for the market analysts and the individual investor, is that companies can sometimes misinterpret the size of a market," said John Giordano, Peritus's current chief executive.

Like other tool makers, Peritus's products were meant to help programmers ensure software would work as designed through Jan. 1, 2000. By automating the search-and-replace process, most tools are advertised to provide a quicker, cheaper fix for Y2K problems.

The most common fix to make software Y2K compliant is the method known as "field expansion" in which programmers sort through code and extend data fields that contain dates from two digits to four. The longer data fields allow programs to express next year as "2000" rather than "00," which some systems might misinterpret as 1900.

Some tools enable programmers to find and expand date fields automatically; other tools employ techniques known as "windowing" or "encapsulation" to help computers determine the correct year.

Windowing programs generally insert logic commands that add or subtract years so that dates can be interpreted within a 100-year window: from 1953 to 2053, for example. Encapsulation tools determine the year by comparing dates with days of the week; the two pieces of data correspond on a 28-year cycle.

But managers responsible for implementing Y2K fixes often eschew tools and hire more programmers and consultants instead. Corporate Y2K chiefs consider the programmers more reliable and more accountable for fixes to specific sections of code.

Also, corporate customers often shun tools that introduce unfamiliar technologies, said Tom Maloof, chief information officer at Stone & Webster Engineering Corp. "Just because you're an information-technology company doesn't mean you will benefit from Y2K," he said of the tool makers.

The upshot is that for their most critical systems, companies often spend more on consultants to assure themselves the Y2K problem is licked. "You have to make a business decision," said Ralph Record, Y2K project manager of Central Maine Power in Augusta, which is spending $4.1 million to update code in a project begun in 1996. For billing and payroll tasks, Record uses cheaper windowing tools. But "for power delivery, you want to be sure it works," Record said.

Tool makers grouse their products aren't appreciated; some complain they've been shut out by bigger consulting shops who would rather hire more consultants, and charge for them, than buy complex software tools from outsiders.

In Waltham, Data Integrity expects revenue of about $25 million this year, up from $9.5 million in 1998, primarily because of sales related to a Y2K fix that many consider a windowing technique. Profits are also up, the company said, though it doesn't disclose by how much.

Results might have been better, said Data Integrity president Allen Burgess, except that he overestimated the interest of consulting companies in fixing Y2K problems quickly.

Typical of such shops is Computer Sciences Corp., Burgess said, the California consulting company that Data Integrity once hoped would buy its Y2K toolset.

But orders were slow, and last year Data Integrity obtained a Computer Sciences memo that stated the Waltham firm's product "appears to have some shortcomings."

A spokesman for Computer Sciences declined to comment, but Data Integrity calls the memo unfair and says it may sue.

"Billable hours is their business model, not providing a better fix," Burgess said.

Other organizations found it easier to replace entire computer networks rather than tweak aging software, said Alexander Arnold, an analyst with H.C. Wainwright.

"In terms of the work that needed to be done, and with a fixed labor pool to address it, it used to look like these automated tools would increase productivity to the point it was needed," Arnold said.

The lesson for software makers, said Kazim Isfahani, analyst at Giga Information Group Inc. in Cambridge, is that markets often don't become as big as forecast -- and they can become saturated and competitive very quickly.

"Tools alone didn't make people big money like they expected" in the Y2K market, Isfahani said. "You have some very big organizations making money, but they're providing more services" to attract corporate customers.