The Fall of the Barbarian Empire

Faced with both pricing and feature disadvantages, sales of Quattro, as well as WordPerfect, 1-2-3, and others, began to sag. Shocked out of his civilized demeanor, Kahn fell back on barbarian tactics. Borland launched an inept series of promotions designed to stop erosion in Quattro sales. The company accomplished the exact opposite instead, destroying the product's credibility.

The first disaster was the Quattro "WinDOS" bundle. The promotion was Kahn's idea, and he insisted on its execution over the strenuous objections of his marketing staff, who feared the promotion would puzzle the market. WinDOS included both the DOS and Windows versions of Quattro in the same box. Once launched, it quickly became clear that Borland's Cassandras were correct: WinDOS wreaked total confusion amongst prospective buyers. Some people thought that the WinDOS "product" was a hybrid of Windows and DOS. Some thought it was a DOS product that looked like Windows. Some thought it was a special Windows version of Quattro that ran under DOS. Some figured out the box contained the complete versions of both products.

In addition to being confusing, the promotion proved to be a sales killer. It turned out that someone who wanted the Windows version of the product had little interest in the DOS version, and vice versa. Many people took advantage of Borland's generosity to help friends and neighbors handle their spreadsheet needs with a leftover Quattro for Windows or DOS. As one observer noted, "WinDOS worked like a competitive upgrade's evil twin." The promotion was hastily canceled, and Kahn, in the spirit of "Le Roi can do no wrong," fired a few of the marketing personnel who had advised against the whole fiasco.10

WinDOS was followed by a series of price-slashing campaigns that finished what the earlier promotion had started. Although some of the early price cuts temporarily boosted sales, this soon stopped as the market wised up and tried to calculate just how low Borland would go. The answer turned out to be $29.95 for a product that a few months earlier had an SRP of $495.00. The pricing strategy reached its nadir when Borland actually ran full-page color ads announcing "Quattro Pro—That's like being offered a Lexus for the price of a Hyundai."11 (Apparently Borland didn't realize that the type of people who actually offer you a Lexus for the price of a Hyundai are often named "Vinnie" and have five Rolexes wrapped around their arms and a car trunk full of

10 In all fairness, Kahn apparently did have some second thoughts about the promotion before its launch. At the rollout of the WinDOS promotion, which took place in England at a swanky London hotel and featured large, billowing clouds of dry ice-generated fog, a giant stage-mounted vinyl WinDOS box that inflated on cue, enough flashing lights to restore the age of disco, and similar items of bad taste, a crowd of about 1,000 people consisting of key customers and journalists was kept waiting for more than an hour while Kahn and several Borland vice presidents considered revamping or canceling the whole idea. They didn't and thus was born one of high tech's most boneheaded marketing campaigns.

11 In my work as a marketing consultant I had used the Lexus/Hyundai comparison for years as an example of the value of credibility in positioning a product, and I was very excited to find out a major software company had actually been stupid enough to create an ad of this sort.

laptops they'll sell to you cheap.) Interest in buying Quattro turned to skepticism as people wondered why Borland had to sell it at a bargain-basement price or speculated that the company was looking to dump inventory before unloading the product.

Complicating things further was that as its woes with Quattro grew, Borland saw its clout with the U.S. channel sharply diminish. Quattro had often functioned as a loss leader for Borland, a lever it used to ensure distributors and resellers also carried generous quantities of the company's more profitable but lower-selling databases. As the spreadsheet's sales and profitability collapsed, distributors and resellers decided they needed to carry fewer Borland database products and wanted better terms on those they did stock, putting increasing pressure on Borland's bottom line.

In the meantime, Microsoft announced it was going to be shipping its long-awaited Windows database, Access, in November 1992, beating Borland's Paradox for Windows to market by 3 months. This was surprising, because Microsoft had pulled the plug on an earlier effort to develop a Windows-specific product and had started again from scratch. Borland, however, in the grip of its object-oriented fervor, had decided it wasn't going to ship Paradox for Windows until it was "right," and it wasn't right enough in 1992.

Microsoft also informed everyone that the new database would have a $99.95 introductory price and would be part of the Microsoft Office suite. Borland immediately cried foul, but it was hard to feel much sympathy for the company. After all, wasn't Paradox supposed to be an "end-user" product? That's how Microsoft was positioning Access, with the Fox line serving as its high-end "developer" product, and $695.00 for a starter DBMS seemed a bit dear.

People also remembered that before Borland purchased Ashton-Tate it had launched a competitive upgrade program against its former rival, offering dBASE users a "lite" version of Paradox for $149.95. It seemed only fair that if the software barbarians were willing to swing the price sword, they should be prepared to defend against it.

Access 1.0 met its November ship date and was greeted with generally decent reviews. The product lacked certain features prized by developers, but it was surprisingly stable for a first release, it intelligently copied many of the ease-of-use features of Paradox, and the price was certainly right. It also incorporated a specialized version of Microsoft's Visual

Basic, allowing developers to leverage their existing skills when developing applications with the new product.

The release of Access could have triggered a positioning war within Microsoft a la what was happening at Borland, but Microsoft was able to finesse the situation. The acquisition of FoxPro had been very friendly, and former Fox employees, flush with Microsoft stock options and cash, were treated well and, in the main, were happy to cooperate with their Microsoft colleagues and their new database. For example, they helped integrate Fox's heralded "Rushmore" technology, a system of binary indexing that speeded up data querying and data retrieval, into Access, a move that helped make the product more competitive with Paradox. And though over the years Access received the lion's share of Microsoft's marketing and PR attention, a situation that engendered some resentment amongst Fox acolytes, FoxPro was regularly updated and improved. This helped soothe wounded feelings and, just as important, kept the dBASE community (or at least Fox's portion of it) from looking for other alternatives to buy and recommend.

January 1993 saw the long-awaited arrival of Paradox for Windows 1.0 at an introductory price of $129.95, Borland's counter to Microsoft's pricing gambit. Unlike with Access, initial reactions were decidedly mixed. Paradox for Windows had serious memory management issues and many bugs (if you've read Chapter 5, this should sound familiar). Features that the developers had requested and hoped would be in the product, such as macros, referential integrity checking, and a data dictionary, were missing. Reviews of the product were polite (Philippe Kahn wasn't Ed Esber) but not enthusiastic, and many urged buyers to wait until the next release. Sales of the product were disappointing.

From the standpoint of Paradox developers, there was a bigger problem. Paradox for Windows was indeed very object oriented. In fact, it was so object oriented that it was a completely new system. Developing applications in it meant learning a new scripting language and starting from scratch. Existing applications couldn't be ported to the Windows version but had to be totally rewritten. Borland now had three mutually incompatible development platforms: dBASE, Paradox for DOS, and Paradox for Windows. Only Paradox for DOS had any claim to the title "best of breed," but the DOS market was rapidly shrinking.

Speaking of DOS, although it was shipping Paradox for Windows, Borland also released a new version of Paradox for DOS. It was a solid improvement over the previous product in many ways, but it still lacked many key features that had been long requested. It did have Turbo Vision, though, an object-oriented new interface that required you rewrite a great deal of your previous applications. And it had a new object-oriented version of the old script language. Then Borland had also along the way introduced ObjectVision, which was sort of a database but really wasn't. Some developers began to think that all this object stuff was getting out of hand.

Faced with this dilemma, the development community hesitated. If they were going to have to rewrite and relearn everything anyway, perhaps it made sense to take a closer look at Access and even FoxPro. Microsoft was certainly selling a great many units of both and, unlike Borland, seemed to be in good financial shape. This is an important consideration to developers, who don't want to see the companies that provide their development tools go out of business and leave them and their applications stranded. Many did take a look, and some decided that Microsoft was the place to be.

Later that year, Borland shipped a new release of Paradox for Windows that fixed the bugs, but the learning curve and rewrite issues remained. The financial news at Borland had only gotten worse in the meantime. More developers left the Borland fold.

In the meantime, Microsoft released its first version of FoxPro for Windows in January 1993 to excellent reviews. The migration from dBASE to FoxPro and other clones accelerated. In March 1993, Borland released dBASE IV 2.0. It didn't include the compiler, but you could buy it separately. It was a decent product, but DOS was becoming increasingly irrelevant. No one appreciated that little game with the compiler, either. And where was dBASE for Windows? More dBASE developers slipped away from Borland's embrace. They were joined by yet more Paradox developers who were convinced it was time to leave a sinking ship.

In 1994, Borland finally shipped its first version of dBASE for Windows. Demonstrating that the company could be every bit as obtuse as Ashton-Tate at its finest, the product lacked a compiler. It was also difficult to use and master. Sales of Paradox, dBASE, and Quattro continued to collapse. By this time, the trickle of dBASE and Paradox developers leaving Borland had become a hemorrhage.

Desperate to regain its footing, Borland bought a Windows dBASE clone from a small company called Arago and released it not long after as dBASE for Windows. Unlike its first dBASE for Windows, this was a solid, high-performance piece of code with all the features the developers wanted. Incredibly enough, it even included a compiler. At this point, Borland had two incompatible versions of dBASE for Windows, dBASE for DOS, Paradox for DOS, and Paradox for Windows to market and support. Those who weren't confused were indifferent, as fewer and fewer people were buying any Borland database. The company was on the brink of extinction, and few people wanted to invest in development tools from a firm in that condition.

By this time the Borland barbarians were restive, and in 1995 a tribal revolt (in part sparked by Phillipe Kahn awarding himself a handsome "performance" bonus of several million dollars of increasingly scarce Borland cash) led to Kahn being kicked out of the company he had founded and of which he remained chief stockholder. Borland exited the business applications market as quickly as it could, selling off Quattro12 and Paradox to Novell. Corel then proceeded to buy WordPerfect and PresentationPerfect from Novell, combined them with its Borland acquisitions in its own office bundle, and then demonstrated how to lose millions of dollars by attempting to sell a suite of second-tier products directly against Microsoft.

Borland then decided to change its name to Inprise. Half the market thought the company was called Imprise, and when they found out they were wrong, people often looked quite comical as they attempted to reprogram their vocal cords to say the word correctly. After some time had passed, everyone got tired of a name that made you simulate a speech impediment and Inprise went back to being Borland.

The company also decided to return to its development and language roots. Despite being given up for dead, Borland successfully morphed hoary Turbo Pascal into a spiffy new development application called Delphi and crawled back from the edge of the abyss. In 2001, Borland's revenue rose 16 percent to $221 million, and profits climbed 11 percent over the prior year to $23 million. Borland ended 2001 with nearly $300 million in cash. But by 2006 the company was again in financial trouble

12 Kahn drove the sale of Quattro to Novell.

and announced it was leaving the software development market for good, divesting itself of TurboPascal's various progeny. The Borland roller coaster has not yet stopped rolling.

Philippe the Barbarian was dead, but Philippe Kahn survived and went off to found Starfish Software, a company that combined the long-forgotten Sidekick program with wireless technology. Motorola purchased Starfish Software in 1998 for $400 million. Kahn then went off and founded a new start-up called LightSurf Technologies, which VeriSign bought for $270 million. He then founded FullPower Technologies. Reportedly, his table manners these days are impeccable.

Money Attraction Secrets

Money Attraction Secrets

Discovering The Laws Of The Right Financial Blueprint. I bet you're scared, angry and maybe even confused. These are perfectly rational and appropriate reactions to the worldwide credit crisis that erupted in 2008 and sends shudders through every home in the United States.

Get My Free Ebook


Post a comment