It's one of the great mysteries of our time. How can products dominate their category when people widely despise them? Why is it not enough for a developer of software, or a provider of a service, to say "It works better!" — and get people's business with that simple statement?

A few years ago, I ran across an observation by Thomas Landauer in a book entitled "The Trouble with Computers." Don't be offended, fellow users of software, but he compares us to pigeons in a behaviorist experiment. As I summarized in a March 27, 1995 column in PC Week,

"Computers are addictive," according to Thomas Landauer, formerly a director of research at Bellcore and now a professor of psychology at the University of Colorado.

Landauer compares the frustrated end user to a pigeon pecking at a lighted key to get a pellet of food. In a famous experiment by behavioral scientist B.F. Skinner, pellets were originally given for every correct peck, then gradually reduced to random rewards, averaging one per 100 pecks. When Skinner stopped giving food entirely, Landauer reports, "the bird pecked the key 10,000 times before Skinner gave up."

In his forthcoming book, The Trouble With Computers: Usefulness, Usability, and Productivity, Landauer attributes users' tolerance for badly designed software to a similar addictive behavior. "Sometimes a very simple error defeats you. You try again. Nope. You think you see why. You try again. Nope. You think of another thing to try. You try something you've tried before, just in case the computer wasn't paying attention. Far into the night, you finally succeed. What a wonderful feeling. Thus is a psychological addiction born."

Now I find new evidence that people don't automatically seek out what works best. Writing this week in the journal Science, researchers report that

Near misses activate the same reward signals in the brain as a win. Slot machine makers capitalize on the near-miss effect. Researchers have found that they program their games to tease players with near misses about 30% of the time–a number previous studies have found optimal for getting gamblers to keep coming back…

A team led by Luke Clark of the University of Cambridge in the United Kingdom developed a simplified slot machine game on a computer…In a gambler's brain, a near miss triggered a similar pattern of brain activity as a win…

Clark says that the rewards of a near miss may have ancient origins. In skill-dependent tasks, such as hunting, people do have some control over the outcome, and trying again after a near miss could bring home the bison, he explains. "The healthy brain is looking for ways in which it can control the environment, and gambling games harness that natural system."

It's like something out of Evil Geniuses in a Nutshell: our DNA wires us to believe that if we're getting close, another try will bring us still closer. This may work for throwing spears, but it rarely works when software just isn't written very well.

What's the impact on developers? It's not enough to eliminate the pain: you also have to compete by offering superior delight plus a genuine opportunity (not just an illusion) for mastery and control.

You have to provide features that are just really cool, as well as everything that users might reasonably want — and you have to give the users extensive options that let them tailor the user experience to personal taste, and give them a strong sense of being invested in your application.