Iteration

It is a commonly accepted truth about software development that the way to get good interaction is to iterate. The devotion to usability testing at most universities and many large software-development companies particularly Microsoft led to the spread of this idea. And, yes, iteration is an important element of good design: Keep working on it until it's right. However, many product developers have interpreted this to mean that you can dispense with design and merely iterate across random thrusts in the dark.

In 1986, Microsoft rushed version one of Windows to market, and it was so pathetic, it deservedly became the laughingstock of the industry. Six months later, Microsoft shipped version 1.03, which fixed some bugs. A year later, Microsoft shipped 1.1, and then version 2.0.^[2] Each iteration of the product tried to solve the problems created by the previous version. Finally, four years after the first version shipped, Microsoft shipped Windows 3.0, and the industry stopped laughing. Few companies in the industry have pockets deep enough, or the tenacity, to ignore public humiliation for four years to finally get it right. One side effect of this is that the industry sees its de facto leader staggering blindly about until it does get it right, and the industry makes the obvious assumption that that is the correct way to do things.

^[2] Microsoft's version-numbering logic is nonexistent. There were at least four major releases of Windows before Windows 3.0. Windows 3.1 was a dramatically different and improved version, with many major changes, and it clearly should have been called Windows 4.0. I'm sure that Microsoft marketing people called it 3.1 instead because they didn't want to squander the market equity already earned by "version three."

But shipping all of those interim versions was very expensive. If Microsoft could have arrived at the quality of Windows 3.0 without shipping a couple of those intermediate releases, it could have saved millions in development and support dollars, earning additional millions in sales much earlier in the product's life (not to mention saving their customers billions of dollars and many headaches). Accepting as true that multiple versions are inevitable is an extremely expensive capitulation of good sense.

Microsoft's strategy is based on simple attrition. In military terms, attrition means that you might be evenly matched with your enemy in quality or even somewhat inferior but you have so many soldiers and guns that you merely trade down until your opponent cannot field any more regiments. In software terms, it means shipping a bad product a real dancing bear then listening to your clients moan and complain. You tweak what they dislike and ship an updated version. After three or four versions, the overt pain suffered by the users subsides and the quality of the product reaches some acceptable minimum, aided by broad functionality, and does not improve thereafter. Iteration never creates great products.

The attrition strategy is not only expensive and time-consuming, but it is a hateful one because it is abusive of people who use computer technology. Unfortunately, it is working pretty well for Microsoft. Time after time, it has shipped half-baked, ill-conceived, poorly built, undesigned products to the sneers and derision of industry observers, both partial and impartial. But while the industry pundits jeer, Microsoft continues to support its first efforts with 2nd, 3rd, 4th, 5th, and 11th versions. Such products as Windows, ActiveX, Word, Access, Windows NT, and many others have eventually emerged as Goliaths in their respective markets.

The attrition strategy only works if you have a rock-solid brand name, lots of time, the nerves of a poker player, and vast quantities of money. So far, no other players in the computer industry have exhibited those qualities in equal measure to Microsoft.

The real problem with Microsoft's spectacular commercial success is that many smaller companies attempt to emulate its success by emulating its attrition strategy. This is often quite unsuccessful in the long term, as Web browser maker Netscape has shown, but it continues the legacy of abusing end users.

It is quite possible to beat the attrition player, but not by using a matching strategy. After all, regardless of who you are, Microsoft has more money than you do. Instead, you must strike Microsoft hard where it is weakest in its development process, which puts programming in front of interaction design. Microsoft is doubly handicapped in that it has many people at the company with the title of "designer" who do design-related things. As shown by the excerpts from Fred Moody's book in Chapter 8, "An Obsolete Culture," the Microsoft culture has already made a place at the table for ineffective, after-the-fact design. Any company willing to do real interaction design can beat Microsoft.