A Tricky Tack For Borland

One thing Philippe Kahn knows how to do is sail. Several years back, he and his crew won the Pacific Cup Race from San Francisco to Hawaii in record time. Typical of Kahn, he won by heading off in his own direction, first sailing south to avoid a high-pressure system where the wind would be weak rather than heading west like the other competitors.

Today, Kahn's software crew, Borland International Inc., needs its own mighty maneuver to set it ahead once again. For the past year, the No.5 maker of personal-computer programs has been veering straight toward the high-pressure system that is Microsoft Corp. Seriously late in bringing to market new products for Microsoft's Windows, Borland has been forced into price wars with the software giant, heavily discounting its new spreadsheet and data-base programs. "The last thing you want to do is sail right behind the leader," says Goldman, Sachs & Co. analyst Rick G. Sherlund. "The right strategy is to sail away, and maybe get a puff of wind that Microsoft doesn't get."

Kahn believes that even though the course he chose has resulted in some short-term setbacks in earnings and market share, it will soon position Borland to compete better with Microsoft. The key, according to the 41-year-old founder, chairman, and CEO, is a big push this fall into "the client-server" software. The move will be aided by Borland's expertise in "object-oriented" programming techniques--a way of assembling software out of reusable modules, or objects. In the short term, the shift to objects has caused costly delays in the critical new Windows products, but by moving early to the new technology, Kahn asserts, he'll wind up ahead of the pack later on.

For now, the immediate pain overshadows the potential gain. Because the first two Windows products created from objects, the Quattro Pro spreadsheet and the Paradox data base, were more than a year late when they shipped last October and January, respectively, Borland lost market share: Paradox dropped from 30% to 22% last year. But favorable reviews for the new Windows version helped Paradox pop up to 25% in June, says market researcher InfoCorp. Quattro Pro, also praised by reviewers, has been stuck at 10% for the past two years, says InfoCorp, though Borland claims that market share has grown in recent months.

Rebuilding market share is only half of Kahn's plan. The other half involves a new division within the $464 million company dedicated to the fast-growing, high-margin business of client-server computing, in which powerful "server" computers coordinate the actions of PC networks. Headed by former Microsoft executive Robert H. Dickerson, the group is developing a slew of new software, including programs to link data kept by individuals on PCs to the corporate data bases stored on servers--minicomputers or souped-up PCs. "We're back in a big way," says Kahn. "You'll see our [new] stuff unfold this fall, and everyone will be surprised."

CATCH-UP. Customers and investors have heard such talk before. And since so many of Borland's products have been so late, many observers aren't holding their breath this time around. Besides the delays with Paradox and Quattro Pro, Borland two years ago promised a new version of the dBase package, which it got in its 1991 buyout of Ashton-Tate Corp. The first hit data base for PCs, dBase still has a loyal worldwide following among programmers. Now, Borland says dBase for Windows will be out by early next year. Meanwhile, Microsoft began shipping a Windows version of FoxPro, a dBase clone, last January.

In the meantime, Borland is mounting a slow recovery. After posting losses of $160 million over two fiscal years because of product delays and restructuring charges from the Ashton-Tate buyout, earnings are rebounding slightly. In the three months ended June 30, Borland earned $6.2 million on revenues of $123.4 million, compared with net income of $1.7 million on sales of $114.8 million last year. Those numbers beat most analysts' estimates, but Wall Street is taking nothing on faith. Borland's stock is trading around 20, near its 52-week low (chart). "Borland has failed to live up to its promises," says Robertson Stephens & Co. analyst Peter Rodgers. "They're clearly not in a leadership position anymore."

And despite a favorable reception for its new packages, Borland's immediate future is still clouded by software price wars. Data-base programs that sold for more than $500 two years ago now sell for around $100. That price deflation--in the only market that Borland has come to dominate--is largely because of Microsoft, which has shipped about 1 million copies of its $99 Access package since November.

An even bigger problem for Borland, though, could be the trend toward selling software "suites." These bundles of three or four different programs go for one discounted price, usually around $800. Microsoft now gets about 70% of its applications-software revenues by selling suites, estimates Sherlund. To respond, Borland linked up with WordPerfect Corp. in April. Borland now sells WordPerfect's word-processing program bundled with Borland's data base and spreadsheet. But getting programs from different suppliers to work as well together as packages from a single supplier is tricky. "I believe it's impossible," says Robert K. Weiler, a senior vice-president at Lotus Development Corp., which also sells suites.

Meanwhile, Lotus is sticking it to Borland on the legal front. The copyright suit it filed against Borland in 1990 has largely gone Lotus' way. The Cambridge (Mass.) company claimed that a 1-2-3 compatibility feature in Quattro Pro infringed on Lotus copyrights, and a Boston judge agreed in July, 1992. Then, a few weeks ago, the judge cleared the way for a trial to determine damages, which Lotus believes are in the range of hundreds of millions of dollars. Borland vehemently disagrees and will try to appeal the decision before any damages are determined or paid.

CROWDED FIELD. Borland's real future, however, will be determined not in the courtroom but in the marketplace, where it must work to become a big player in client-server software. Data-base suppliers such as Oracle, Sybase, and Informix have a huge headstart over Borland and have enjoyed record profits for the past year or so. In addition, startups such as Gupta Corp. and Powersoft Corp. have carved out niches in data-base software and programming tools that help companies set up small client-server networks--the same market Borland is eyeing. Those companies floated two of this year's most successful initial public offerings.

Dickerson is the one who will have to deal with these rivals. "When the smoke clears," he boasts, "the Guptas and the Powersofts of the world are gonna be toast." He'll start with what he calls "glue" programs that will allow Paradox and dBase to exchange data with Interbase, Borland's own server data base, as well as with those from Oracle and Sybase. Then, a whole series of products is planned, including new programming tools and versions of Interbase for Microsoft's Windows NT operating system and Novell Inc.'s pervasive NetWare networking software, according to sources close to the company. These products would let companies link all of their employees into one central inventory or accounting system without using an expensive mainframe. Kahn says client-server programs should bring in hundreds of millions of dollars in revenue in a few years vs. $25 million now.

He believes that object-oriented programming will give Borland a special edge. Kahn says he can now quickly add new features by simply adding new modules to any of his packages. Rivals are also adopting object technology, though most are doing it more slowly and quietly. They raise an interesting question: If Borland's object technology is so good, why hasn't there been a bigger payoff? Steven P. Jobs, whose NeXT Computer Inc. has developed an object-oriented operating system for desktop computers, claims it's because Borland's software is not built on such an operating system. "You can't do it on top of Windows," argues Jobs. "Borland just has a thin veneer" of object technology.

Some critics think Borland's problem is much simpler and lower-tech. They think that Kahn, who had cultivated an image of his company as a lean and aggressive bunch of "barbarians," simply became complacent. "When he starts believing all the press that he's successful, that's when the mistakes are made," says Brad Silverberg, a Microsoft vice-president who used to work for Kahn at Borland. Wall Street analysts cite Kahn's decision in 1991 to go ahead with plans to build a new, $100 million headquarters despite mounting pressure on profits, as an example of his complacency. Kahn says it was the only way to consolidate the company's 1,900 employees, now spread among 11 buildings in Scotts Valley, near the California coastal town of Santa Cruz.

The snazzy new headquarters befits a software superpower. But there are many who doubt that Borland will ever achieve that status. Influential analysts such as Rodgers of Robertson Stephens believe Borland missed its chance and is now only valuable as a potential acquisition by a financially strong software company. Kahn denies an assertion by Lotus that he approached that company about a merger and claims he rebuffed entreaties by Computer Associates International Inc. Kahn says he's steadfast about keeping Borland independent. He seems to be counting on one amazing gust of wind.