Doug Carlston - Software People. An Insider’s Look at the Personal Computer Software Industry

MicroPro continued to be the top-grossing software publisher through 1983, but its flagship product, WordStar, had slipped from its position of dominance in the word-processing field, and its long-awaited upgraded version, WordStar 2000, was greeted with lukewarm reviews when it was released in 1985. The company weathered the shakeout, but not without major reductions in force.

Microsoft, the giant that grew from the hobbies of a couple of teenage hackers, was no longer a whiz-kid company. Microsoft BASIC continues to be the first computer language most microcomputer users learn, and MS-DOS, its operating system for IBM PCs, continues to dominate the operating-system market. But the company’s newest efforts have been greeted with less overwhelming success.

Microsoft’s display booth at the 1984-85 Winter Consumer Electronics Show was a collaborative effort with all of its Japanese partners—Kay Nishi’s ASCII/Microsoft Japan notable among them—to promote MSX, a standard operating system for a new kind of low-cost computer manufactured by more than twenty different Japanese vendors. The manufacturers were being coy about whether or not they were actually going to launch the product in the United States. After inflated claims of a new coup in the global operating-system wars, MSX had enjoyed only moderate success in Japan, and the jury was still out on Europe. More ominously, Microsoft’s long-awaited big project, a monumental production called Windows, touted to be an entirely new kind of operating system, was late and getting later—in the manner of VisiOn, the program that helped lead to the demise of VisiCorp.

The troubles of 1984 were not confined to the home computer software industry. They originated in the computer hardware shakeout that started in 1982. The problems extended across the software spectrum, from the tiny cottage companies to the corporate entrants to the much larger business software companies. During the summer of 1984, virtually every software stock listed on the public exchanges was down.

The troubles spread to the surrounding industries too. The computer magazines, with their advertising revenues dramatically diminished, experienced their own shakeout. And there were rumblings that even the venture capital community might not be immune. With so many firms chasing deals, more and more had to overbid on increasingly risky investments.

The first and most obvious fundamental reason for the software shakeout was a lack of business management skills on the part of many of the formerly hobbyist entrepreneurs, in particular their failure to remember the one thing required of all business ventures—that they return a profit. Many firms, however, feeling no need to be profitable today, since they were going to be phenomenally profitable tomorrow, spent money foolishly. For example, many entrepreneurs built organizations that were appropriate in size and structure for the large company they hoped to be rather than the small firms they actually were. And many got so wrapped up in this process of “building a company” that they stopped studying the market carefully and so lost the speed and flexibility that had made them successful in the first place.

One has only to look at the companies that ended up with huge inventories of product for dead machines or that failed to support the wildly successful machines when they came out. Or the companies that turned out the same kind of product year after year despite changing public tastes and increasing technical expectations. At Brøderbund we took pride in the fact that although we started as an Apple game company, Apple games constituted only 13 percent of our sales in 1984. Games for other machines accounted for 11 percent; productivity software accounted for 75 percent; educational programs were 1 percent of our sales.

If the lack of business skills that caught up with many of the hobbyists who founded software companies was one reason for the software shakeout, then the second reason—one that did in so many of the corporate industry participants— was an inability to develop and handle new sources of quality products. The reason for this shortcoming is a little hard to comprehend. Perhaps it was because hackers have such a huge advantage (in that they are more imaginative and more obsessed) over the more orthodox, properly trained programmers preferred by corporate publishers when it comes to dealing with very small computers. It’s a vicious circle—if you are going to work with hackers you have to be able to speak their language, but orthodox publishers don’t like hackers, partially because of their propensity for speaking their own language. Unfortunately, both the very best and the very worst programmers come from their ranks, but if you can’t communicate with them well enough to tell the difference, you are going to get into trouble very quickly.

The same inability to develop new sources of quality products also plagued many cottage-industry companies, some of which were founded by programmers who in turn seemed to have trouble working with new programmers. This difficulty probably stemmed from the fact that ex-hacker executives were at some level competing with these new programmers. This competition can be deadly, since hackers-turned-executives quickly lose many of their hacker skills, but many are loathe to admit that their hacker employees are more skilled than they are.

Developing new sources of product on a regular basis is critical to success because, in the final analysis, it is still the caliber of its product that determines a company’s success. Marketing is important. But the software consumer’s grapevine is phenomenally effective and can fairly quickly disrupt an aggressive marketing campaign for a faulty program.

The third reason why many companies failed is that they didn’t understand whom their market consisted of. I often had occasion to talk with people who were thinking about jumping into the computer software business. They would talk about the “installed base” of a certain type of machine, and I would nod sagely. “The Atari video game machine has the largest installed base,” they would say, and I would agree. “And among the computers, it looks like the Texas Instruments and Commodore VIC have the largest base,” they would continue.

They could not have chosen three less attractive software markets for 1984. The truth is that installed base isn’t a very important concept, as said earlier. What is important is that a particular kind of computer is selling well at the time that a company is selling software for it and that it is expected to continue selling well for a long time into the future! Indeed, these expectations are probably more important than actual rates of sale. For example, during the summer of 1984, Apple’s Macintosh was probably outselling IBM’s PCjr by only three or four to one. However, Mac software was outselling PCjr software twenty to one. The Mac was seen as a winner, but the PCjr was perceived to be a loser.

By contrast, many companies that entered the business late suffered from the “Future Computing Syndrome.” They sent scouts out to the thinktanks to find out what kinds of products could be expected to sell well in future years. The thinktanks told them, and they all went home and started working on identical products. The first ones to reach the market may have done reasonably well, but as more and more products flooded that market segment, all the companies suffered. The area of young children’s educational entertainment products (sometimes known as “edutainment”) was opened and then filled in this manner in 1983 and 1984. By the summer of 1984, it was a terrible market for most companies.

The fourth reason for the shakeout was the extreme availability ( not unavailability) of money. There was lots of money, tons of it, invested in just about every software idea under the sun. So we not only got too many competitors; we also got lots of companies with soft underbellies.