(FORTUNE Magazine) – A global hotel chain was stunned to discover a perverse consequence of its customer-centric Six Sigma quality initiative. Apparently guests were mildly pleased by the chain's sincere efforts to provide a hassle-free stay. But what really moved the customer-satisfaction needle was how well the hotel responded when something went wrong.

Guests who had experienced a problem that was quickly and politely resolved rated the hotel service higher than guests who had had no problems at all. What's more, more guests with happy resolution of their hassle said they were likely to recommend the hotel than did the trouble-free guests.

So the hotel's COO announced a new twist on Six Sigma: The company was going to cut back on problem prevention and spend more on its "recovery management" training. A few of the COO's underlings had joked about having their hotels deliberately foment customer snafus as a marketing technique. The joke was taken seriously: The chain is now trying to determine which kinds of problems may be appropriately experienced by which customers--as long as the appropriate recovery procedures are safely in place.

An ounce of prevention isn't always worth a pound of cure. Fixing your mistakes may prove a better business investment than preventing them. After all, it's customer perception--not Deming-trained statisticians--that ultimately determines product and service quality.

Everyone makes mistakes. Processes and procedures fail. That's life. But aside from catastrophic failure--the jet engine that explodes, the pacemaker with faulty circuitry, the lovely laptop whose battery bursts into flame when recharged--it's simply not clear whether most firms (and their customers) are better off investing more of themselves in preventing failure or in building on it. At software companies, that is called turning a bug into a feature.

If a painlessly honored warranty makes for more satisfied customers than a mere defect-free experience, then go with the greatest good for the greatest number. If the graceful recovery is economically faster, better, and cheaper than avoiding the original harm, then recoup with recovery. After all, few faces look happier than a frequent flier who receives--without asking--a polite apology and a free upgrade after a miserably late flight. Cheerfully comp dessert for the impatient patron whose entree is served 20 minutes late and the restaurant makes a customer for life.

No doubt designing creative recovery programs for customers and employees will become a lucrative consulting practice. Learning which people and personality types reward recovery and which simply want to be left alone can't help but be an important segmentation exercise.

The managerial problem is that most individuals and institutions are miserable failures at failing. By definition, quality organizations are so focused on preventing problems that when problems do materialize, they are woefully underinvested--both culturally and financially--in dealing with them in a cost-effective way. When you optimize around prevention, recovery is literally an afterthought.

Conversely, companies biased toward recovery have a different sociology of failure: We're so good at cleaning up the mess, they think, that hey, we'll fix the source of the problems next time around...maybe...if that's what people really want....

The obvious evolutionary end game is that "preventers" and "recoverers" will simply develop different domains of acceptable failure; preventers will get better at recovery, and recoverers (admittedly, the minority) will learn what problems need to be prevented.

That's not what's going to happen. As more and more organizations recognize that their paying customers and clients are their most important asset, they will make every effort to shift themselves into the recovery camp. Why? Because it's the surest way of maintaining interaction with the customer. People don't call you up to tell you everything is wonderful; they call you up because they want something else, something better, or something fixed. You can bet that all those much vaunted Six Sigma initiatives are going to mutate away from making "better" products and services and into programs that explore the most creative ways to increase customer dependency. Businesses will shift from enabling quality to quality enabling. People won't be our customers; they'll be our co-dependents. If we play it right, they'll be recovering with us.

Michael Schrage is co-director of MIT Media Lab's e-markets initiative and the author of Serious Play. He can be reached at michael_schrage@fortunemail.com.