Microsoft Access didn’t exist in the MS-DOS era. Microsoft wasn’t even a player in databases until relatively late in the MS-DOS era. The dominant database during most of the DOS era was a company called Ashton-Tate. They had a product called dBase II that ran on CP/M machines and dominated that era. It ported over to the IBM PC and dominated there as well. Ashton-Tate’s dBase II and its successor, dBase III, became the de facto standard for databases on microcomputers. On October 12, 1991, Borland acquired Ashton-Tate for $439 million. It was a colossal mistake.
The crowded DOS database market
Borland already had a database program called Paradox. Between Paradox and dBase, Borland expected to control 70 percent of the MS-DOS database market.
But dBase had a problem. Another company, Fox Software, had a clone of dBase. It was cheaper and worked better. Ashton-Tate sued Fox claiming copyright violation. Ironically this was on the same grounds as Lotus had sued Borland, for creating a Lotus 1-2-3 clone it called Quattro. (Cuatro is the number four in Spanish.) Part of the agreement with the department of justice was that if Quattro was deemed legal, Borland would have to quickly resolve Ashton-Tate’s lawsuit with Fox.
Borland’s CEO, Philippe Kahn, said at the time that Borland intended to make the dBase commands “an open standard for the benefit of all software users.”
What went wrong
This was a noble move on Borland’s part. The suit wasn’t resolved until 1996, but Khan effectively said he didn’t have a problem with dBase clones. And in doing so, he created a problem for himself.
The original dBase II was written in Z-80 and 8088 assembly language. To create dBase III, Ashton-Tate used a conversion program to convert the assembly language code into C. This made the program slower, a problem that solved itself as PCs grew faster. But the translated code was hard to maintain. The next major version, dBase IV, was late, buggy, and slow. That was also true of dBase III, but Ashton-Tate had quickly released revised versions. This time it took two years. Impatient customers started switching to clones like Fox and Clipper.
Ashton-Tate’s market share dropped from 63% in 1988 to 43% in 1989. In its final year as an independent company, it lost $40 million and laid off 400 of its 1,800 employees.
Borland paid almost half a billion dollars for damaged goods.
In March 1992, Microsoft paid $170 million to acquire Fox, putting Microsoft in the database market. Meanwhile, Microsoft and Ashton-Tate had been cooperating on technology to make desktop database software use a SQL backend. This partnership didn’t get far, but it evolved into what became Microsoft Access. Microsoft had its database of the present in Fox, and its database of the future in Access.
The shrinking software market and how it bit Borland
Prior to 1991, a no fewer than four major database products fought for desktop market share: Borland’s own Paradox, Ashton-Tate’s dBase, Fox, and Clipper. Today, that market is dominated by Microsoft Access, Microsoft Access, Microsoft Access, and Microsoft Access.
The usual trope that other software makers were slow to adopt Windows didn’t apply here. Borland was quick to move its software to Windows and even had its own compilers for Windows that competed successfully with Microsoft.
There was something larger going on. In the 1980s, companies had few people who understood the growing personal computer phenomenon and so most technical people were given free rein to purchase whatever software they thought they needed. Borland had done an excellent job marketing to those with a highly technical bent.
I even remember an article in Personal Computing magazine titled “Status Software.” The only trace I can find of it today is a reference in a 1986 issue stating it was published in 1985. The idea was to buy a certain collection of software titles to display in your office so you’d look smart, and another selection of titles that you would actually use most of the time and probably hide from view.
By the mid-1990s, companies were beginning to ask what the return was on the investment they had made in this loosely controlled PC software buying spree. Company executives were starting to ask questions that were hard for technically minded staff to answer, and so corporate standards began to be created. This required new kinds of marketing and support materials from software vendors, but Borland remained focused on the technical side of its products.
This created an opening for Microsoft. Fox provided a way to run legacy dBase code on Windows for those who were inclined. And for those wanting to modernize to a client-server model, Microsoft had Access.
Microsoft Office and the end of the line for Borland and dBase
Bundling Access into Microsoft Office sealed the deal. Office had the best spreadsheet and no worse than the second best word processor and presentation software, and the three worked well together. Fledgling corporate IT departments didn’t want to support three different office suites. Microsoft usually won by default, which usually brought Access along for the ride. Meanwhile, dBase didn’t exactly go away, but was no longer a growth market.
By October 1994, Borland was about ready to give up. It didn’t have a word processor or presentation graphics program, having discontinued Ashton-Tate’s offerings in those spaces. It sold Quattro Pro to Novell for $140 in cash and included the rights for Novell to sell up to a million copies of Paradox. Borland planned to refocus on software development tools.
In retrospect, Borland didn’t get a lot for its $439 million purchase. It would have been better off buying Fox instead, letting Ashton-Tate bleed so it could buy it for less, or staying the course with Paradox. The two larger software firms, Lotus and Microsoft, weren’t in any hurry to buy Ashton-Tate either.
David Farquhar is a computer security professional, entrepreneur, and author. He has written professionally about computers since 1991, so he was writing about retro computers when they were still new. He has been working in IT professionally since 1994 and has specialized in vulnerability management since 2013. He holds Security+ and CISSP certifications. Today he blogs five times a week, mostly about retro computers and retro gaming covering the time period from 1975 to 2000.
I did a deep drive into Ashton-Tate history for my own blog. Personally, my favorite detail about the company is that there never was an “Ashton.” The co-founders were George Tate and Howard Lashlee. It was an advertising executive who came up with the name Ashton-Tate, which I’ve always assumed was just because he wanted a name that started with “A” so the company would be at the top of any alphabetical list. (See Accolade, Acclaim, etc.)
As far as why the company failed, you could go back to 1984. George Tate died suddenly. A few months later, the CEO abruptly resigned. (Lashlee was never involved much with the database part of the business.) The new CEO, Ed Esber, was perfectly qualified, but he quickly got himself into an arms race with Jim Manzi over at Lotus to see who could try and expand their fiefdoms via acquisitions the quickest. Ultimately, both companies lost that race.