Gene is a principal in Progressive Performance Software. He can be contacted via the web at http://www.stgtech.com/ and via e-mail at [email protected].
The most oft-cited examples of "superior" technology losing out in the marketplace are Betamax, Macintosh, and the Dvorak keyboard. The economic explanation for this triumvirate is the theory of path dependence, the keystone of "increasing return economics." The theory is that in the high-tech world, as opposed to other manufacturing industries, a company's profit margin often increases with each additional customer. Increasing return economics claims that because of a putative size advantage, the early market leader will be able to crush late arrivals by slashing prices, running aggressive ad campaigns, throwing its weight around with suppliers and partners, and so on. This leads to path dependence, because although the late arrivals may have better products, they won't stand a chance in the market. The veracity of path dependence theory has great relevance to us as programmers, as it is likely to affect our choices regarding the APIs, programming languages, and operating systems that we work with.
Much as in the AI debate, where it is useful to differentiate between the claims of "weak" AI and "strong" AI, the notion of path dependence can be divided into a weak and a strong case. The weak case says little more than that the future is, to some extent, dependent on the past. The strong case goes much further, and contends that we often get stuck using inferior products, as the first product to market can win out against significantly superior competitive products. Strong path dependence means that it is only through coincidence that the market winner is the best technology. If strong path dependence is a common phenomenon, then certain conclusions are suggested: it is generally futile to compete against standard holders in the market, and it is useful to have the government intervene in the standard-setting process.
A few examples illustrate how commonplace the idea is that Dvorak, Betamax, and Macintosh were superior products that the market capriciously rejected. (A search of Alta Vista will turn up about as many examples as any reader would care to see.)
- Jared Diamond, in Discover
Magazine, writes, "The infinitely superior Dvorak keyboard is
named for August Dvorak... QWERTY's saga illustrates a much broader
phenomenon: how commitment shapes the history of technology and
culture, often selecting which innovations become entrenched and which
are rejected."
- In the Babson Free
Press, Business Editor Daniel J. Keefe, wondering what will
become of Apple, writes: "This product, that we know today as VHS,
ended up becoming the standard for home video, despite its lesser
quality. We can draw some interesting parallels here."
[5]
- Wayne Huang of VRoom
Studios suggests that Windows users are suffering from a serious
psychological disability, and asks, "What would be the psychology
behind choosing Windows computers in the face of such overwhelming
evidence of the superiority of the Macintosh computer in everything
from stability to graphics and being the platform of choice for
creating web content?" [3]
Some of the leading proponents of strong path dependence feel that a new branch of economics is necessary to explain this phenomena. However, like most new branches of economics, this branch does not stem from a failure of classical models to fit new phenomena, but from a failure to understand how to apply the classical models correctly. The empirical evidence necessitating a new branch seems to be shaky, at best.
The Evidence
No one doubts that there are markets that have a natural tendency to move toward a state where a single producer, or type of product, dominates. The VCR market is an excellent example -- people want to exchange tapes, studios want to release movies in one format, and video stores have no desire to stock two of everything so that they can offer VHS and Betamax. However, the issue is not if this occurs, but whether it leads to poor choices for standards, and whether a group of industry experts, operating with a government mandate, could make better choices.
Economists call the effect a sale has on parties other than the buyer and seller an "externality." Pollution is a classic example: it may be the people down the river who pay the price for the waste a factory produces, rather than the manufacturer or the customer. An example of a "standards externality" would be the effect that the next sale of Windows NT has on all of the previous purchasers. Even though they did not participate in the sale, they gain a benefit from having one more customer in their "camp." Developers are more likely to write for their chosen platform, there is a broader base of users from which to hire employees, and so on. Classical economic theory tells us that the "... expected effect of a 'standards externality' is on the amount of the standard-using activity, not on choice of standard or the mix of standards." [8] In other words, there will be a higher percentage of consumers using the leading product if there is a widespread benefit to having a standard -- and this is just what consumers want! However, the existence of bias toward a standard should not, per classical theory, affect which product is the leader.
Examining the three popular cases put forward by the proponents of strong path dependence will show how this theory works out in practice.
Betamax vs. VHS
Sony, the developer of the Betamax format, and Matsushita, one of the VHS developers, chose to give different weight to the importance of ease of transportability (which meant small tape size) and recording time (which increases with tape size). Sony thought that consumers would want a paperback-sized cassette, even though this limited recording time to one hour. Matsushita, on the other hand, opted for a larger tape and a two-hour recording time. Later, when Betamax raised its recording time to two hours (by slowing down the recording speed), VHS raised its to four. Otherwise, the two technologies were nearly identical.
Liebowitz and Margolis point out that:
Not only did the market not get stuck on the Beta path, it was able to make the switch to the slightly better VHS path. Notice that this is anything but path dependence. Even though Beta got there first [by two years!], VHS was able to overtake Beta very quickly. This, of course, is the exact opposite of the predictions of path dependence, which implies that the first product to reach the market is likely to win the race even if it is inferior to later rivals. For most consumers, VHS offered a better set of performance features. The market outcome, in which VHS prevailed, is exactly what they wanted. [8]
QWERTY vs. Dvorak
The superiority of the Dvorak keyboard to the standard QWERTY layout is usually taken for granted. Yet this myth is built around a single study, which was performed by the U.S. Navy during World War II. The study, it turns out, was conducted by none other than keyboard inventor August Dvorak himself, who was the Navy's top expert in the analysis of time and motion studies during World War II. The study used inadequate controls, and allows no real comparison between the two training groups. [9] Later studies have failed to support Dvorak's results.
The apocryphal story on the QWERTY keyboard is that it was deliberately designed to slow down typists, to prevent key jams. However, the keyboard design actually had to win many typing-speed contests to eventually triumph in the market. "The QWERTY keyboard, it turns out, is about as good a design as the Dvorak keyboard, and was better than most competing designs that existed in the late 1800s when there were many keyboard designs maneuvering for a place in the market." [8] If this were not the case, and Dvorak's claims of large efficiency increases with a few days training were true, then a single corporation with a large typing pool could internalize the costs of switching technologies, and lead the market away from the "QWERTY trap" all by itself.
Notice that the new "ergonomic" keyboards, from Microsoft and others, also belie the notion that we are locked-in to a particular keyboard choice. Even though they use the same key order as the traditional QWERTY keyboard, the ergonomic models use a very different shape for the keyboard as a whole. Yet, despite the re-training necessary to use them, they are apparently viable in the marketplace. And mice are, in fact, a replacement for keyboards on many input tasks, and have become ubiquitous despite their radically different operation.
Macintosh vs. Wintel
Calling the Macintosh superior to "Wintel" computers ignores many dimensions of what users want from their machines: expandability, a good price/performance ratio, a wide choice of peripherals, a wealth of software, and so on.
Whether or not a technology is superior often depends on who is answering the question. Macintosh always had its strong adherents among the visually oriented professionals in graphic design, and in education, where ease of learning was paramount. However, it never won the hearts or minds of the general business community, which continually opted for the superior price/performance and "power-user" features of DOS over the ease of learning of the Macintosh.
Nor did it win over the techies, who preferred the ease of programming of DOS (e.g., the Macintosh did not even support a scripting language for many years after its release) and the ability to get directly to the machine with assembly language, allowing for high-performance graphics games and so on. Apple failed to realize that a large portion of the early PC adopters were tinkerers, who wanted to be able to get inside their machine, trying to hook a piece of lab equipment to the bus, for example. The early Mac was simply not to be tinkered with. The case was not even designed to be opened by the consumer. The first developer of a hard drive for the Mac did so with no assistance from Apple, and despite the fact that Apple engineers said it was impossible. This was a crucial mistake, for it is the tinkerers who develop the peripherals and applications that a platform needs to succeed. Tom Steinert-Threlkeld sums up Apple's failure: "Open up your architecture to all comers and win -- or keep it closed, like the Macintosh, and lose." [12]
What Does This Mean to Software Professionals?
If strong path dependence is so uncommon that the proponents of the theory can't find any good examples, then there are several important conclusions we can draw.
Listen to what the market is telling you.
If your "superior" product is failing, perhaps it's not actually superior in some important dimension. Analyzing why Apple lost out to Microsoft in the battle for the desktop, Virginia I. Postrel, editor of Reason Magazine, points out that Apple was "...too hung up on the beauty of their product to understand that consumers actually cared about many other things: price, plenty of software, and compatibility with other systems. Quality is not one-dimensional." [11]
To compete against a market leader, specialize.
"... a single-standard equilibrium, if it is achieved, is more readily displaced by an alternative if preferences differ. This suggests that product strategies leading to strong allegiances of some group of customers are likely to be effective in the face of an incumbent standard." [8] Linux is following this path with great success, appealing to the tinkerer market that GUIs tend to ignore. Ellen Ullman did an excellent piece in Salon Magazine on the satisfaction she derived from deleting NT and installing Linux on one of her machines. [13]
There may not be safety in numbers.
Many developers might prefer working on a platform other than Microsoft's, but stick with Windows for fear of being crushed by the "Microsoft juggernaut." Similarly, many IS managers have sought safety in being extremely conservative with their technological choices. Ten years ago, this conservatism found expression in the idea that you couldn't get fired for going with IBM. Today, Microsoft is substituted for IBM. (Doesn't that say something?) The idea of path dependence is implicit in this attitude -- no one can replace Very Big, Inc., as its lead is too large. The absence of any evidence for strong path dependence should give these developers and managers pause. Those who bet on OS/2, which was developed and backed by both of the "can't-get-fired-for-going-with" companies, learned this lesson the hard way.
You can't rest on your monopolistic laurels.
Full disclosure: I am not a big Microsoft fan. I have worked on UNIX systems for a decade, and hope to continue to do so for the foreseeable future. Nor am I a big fan of most current GUIs -- I discussed their limitations in an article for Software Development several years ago. [1]
However, we should give credit where due. Liebowitz and Margolis remark that, "Since high technology changes so frequently, a firm that achieved monopoly with one technology will not be able to hold on to its lead unless it is extremely resourceful." [8]
Far from stifling innovation, as Arthur contends will happen, Microsoft has had to innovate constantly to maintain its position. Just three years ago, "Microsoft Bob" was, according to Redmond, the "easy-going software that everyone will use." But it turned out that it was actually Web browsers that everyone would use, and Microsoft quickly adopted this interface as its own. Giving the browser and the OS the same interface is a major innovation, a step that no other OS has undertaken. Or consider Rich Text Format (RTF), Microsoft's erstwhile standard for portable, formatted text. Although Microsoft was blind-sided by the success of the Web, it quickly adopted HTML as its new standard, and RTF has been left by the wayside. Microsoft may have wanted to stifle innovation in these cases; it was simply unable to do so. And whatever happened to that network software Microsoft sold in the late 1980s? Can anyone even remember the name of it?
Related web sites
[1] Callahan, Gene. "Excessive Realism in GUI Design: Helpful or Harmful?" Software Development, November 1994.
[2] Diamond, Jared. "The Curse of QWERTY." Discover Magazine, April 1997. (The hyperlink here is to the archive search engine. Enter "QWERTY" in the search field, and this article will be returned.)
[3] Huang, Wayne. "Psychology and OS Platform Preferences." Vroom Studios, May 5, 1998.
[4] Kedrosky, Paul. "The More You Sell, the More You Sell." Wired Magazine, October 1995.
[5] Keefe, Daniel J. "What Ever Happened to Apple Computer?." Babson Free Press, February 17, 1998.
[6] Kelly, Kevin and Reiss, Spencer. "One Huge Computer." Wired Magazine, August 1998.
[7] Liebowitz, Stan and Margolis, Stephen E. "The Fable of the Keys." Journal of Law and Economics, April 1990.
[8] Liebowitz, Stan and Margolis, Stephen E. "Should Technology Choice be a Concern of Antitrust Policy?" Harvard Journal of Law and Technology, Summer 1996.
[9] Liebowitz, Stan and Margolis, Stephen E. "Typing Errors?" Reason Magazine, June, 1996.
[10] Liebowitz, Stan and Margolis, Stephen E. "Path Dependence and Economic Evolution." Jobs & Capital Volume VI, Number 4 / Fall 1997.
[11] Postrel, Virginia I. "Creative Insecurity." Reason Magazine, January, 1998.
[12] Steinert-Threlkeld, Tom. "Linux: The back door is open." Inter@ctive Week Online, September 11, 1998.
[13] Ullman, Ellen. "The dumbing-down of programming." Salon Magazine, May 12, 1998.
These op/eds do not necessarily reflect the opinions of the author's employer or of Dr. Dobb's Journal. If you have comments, questions, or would like to contribute your own opinions, please contact us at [email protected].
