Last month, I suggested that customers would be better able to cope with a world of defective software if vendors published lists of their software's known bugs on their Web sites. But that suggestion is a coping strategy. In the long term, I want a world of low-defect software. How can consumers force vendors to create that world? I have two suggestions.
Publish Return Rates
First, consumers could encourage companies to post on their Web sites the return rates for their software packages. James Love, director of Ralph Nader's Consumer Project on Technology (http://www.cptech.org), suggested this option to me in an interview in 1998. He said, "You go out and buy some program like Big Happy Bear Goes to the Moon for your kid and install it on your system, and all of a sudden your computer's messed up. Maybe you wouldn't have bought the software if you could look on the vendor's Web site and see that 25 percent of the people who bought the program returned it."
This proposal would probably work only if legislation backed it up. Software vendors wouldn't like the law, but other industries already live with this kind of forced openness. For example, airlines must report the percentage of their delayed flights. Initially, the airlines weren't happy about the rule. But now, airlines with good percentages advertise their on-time record, and the percentage of delayed flights has fallen dramatically. Windows NT Magazine lives under a similar sunshine law; to ensure that advertisers know what they're paying for, magazines must regularly publish a report that details how many people purchase the magazine.
Boycott Unnecessary Upgrades
Second, consumers could send a clear signal to software vendors that we expect high-quality products by refraining from purchasing an upgrade until we're sure that the upgrade's code is clean and reliable. NT users could lead the rest of the software industry in this campaign; many NT users are already careful about upgrade decisions. Some companies still use NT Server 3.51 because they aren't convinced that NT 4.0 is more reliable than the earlier OS. In addition, many IS shops that currently use NT Server 4.0 didn't upgrade until Microsoft released Service Pack 3 (SP3).
PC veterans might recall MS-DOS 4.0, which had one—and only one—useful feature: It could format hard disks larger than 32MB. Otherwise, the OS offered little that MS-DOS 3.3 didn't offer, and MS-DOS 4.0 had some bizarre quirks. Few companies upgraded to MS-DOS 4.0, so Microsoft scrambled to build a better DOS. MS-DOS 5.0 was the result. The software's good value brought Microsoft customers back into the DOS-upgrading fold.
Like the users who refused to upgrade to MS-DOS 4.0, make the business case for every software upgrade you consider. Think about buying an upgrade in the same way that you'd think about constructing a new building or buying a delivery truck; all three purchases are investments. Businesses generally apply standard business analysis tools to investment decisions. They examine each investment's costs and expected revenues (or savings). They consider the net present value of each investment. When you view software upgrades in this context, you can justify few upgrades on the basis of their new features. But you can justify many upgrades that improve the software's quality. If you consider a software upgrade and decide against it, write a brief, polite letter to the vendor explaining why you made that decision.
You've probably heard a lot about total cost of ownership (TCO) in the past 2 years, but you might not have considered that bug-free, or at least low-defect, software could do more to lower the cost of software ownership than any other factor. Good market information from vendors would help customers make more informed software-buying decisions. And if software consumers put more thought into buying new software and upgrades, vendors might choose to focus more attention on products' quality—a choice that would ultimately benefit vendors and buyers alike.