The following post originally appeared on the Harvard Business Review Blog Network.

If you're serious about creating a customer-driven culture, then you will certainly tie incentive compensation to customer feedback. Right?

Wrong.

Everyone knows incentive compensation can really focus people's attention. Done right, it can reinforce a company's values. It can spur creativity and accelerate actions designed to meet important goals.

When it's done wrong, however, the unintended consequences can be devastating—which is exactly what happens when some companies try to link compensation to customer feedback. Here are three big risks:

Extreme focus on the score. You can't tie compensation to feedback unless you have a method of quantifying that feedback. This can take the form of a Net Promoter score or some other kind of customer-satisfaction score. Enterprise Rent-A-Car, for instance, defines its Enterprise Service Quality index, or ESQi, as the percentage of customers who check the top box in a survey, indicating that they were completely satisfied with their rental. It's the basis for their goals.

But with incentive compensation, you get exactly—and only—what you pay for. Once compensation depends on improving a particular score, people tend to focus on the metric rather than on what it tells you about what customers want or need. Rather than focusing their efforts on improving the company's products or customer experience, employees tend to devote more time and energy to explaining or criticizing the metric itself. They will find creative ways to attack your data collection methods. They will invent innovative ways to explain away poor performance ("Oh, problems with our vendors affected our customers and depressed our scores," "The market was down," or "The weather was bad."). They may complain that they can't control all the variables that affect the scores, so they shouldn't be held accountable for them.

Gaming. The creativity that should be going toward finding innovative ways to wow customers often finds a different and more dangerous outlet. At metrics-driven companies, an extreme focus on customer feedback scores can eventually lead to gaming the system. Some of the more innovative frontline employees may "bury" customers who might give negative feedback by recording inaccurate contact information or marking them as opt-outs from the surveys. Some branches at Enterprise, for example, were rumored to fudge a digit or two on the phone numbers of unhappy customers, so that surveyors making follow-up calls couldn't make contact. (Enterprise terms such practices speeding and regards them as grounds for dismissal.) At other companies, employees ask only happy customers to answer the survey. While shopping one day, I noticed a clerk circling the survey invitation at the bottom of the receipt with a Sharpie pen and telling customers, "It would really help me if you would fill out our survey. If you mention me by name, I'll get an award." She did this only for some of the customers at her register, however. Guess which ones.

A far less creative form of gaming—and a common practice at many companies—is outright begging. Most American consumers (and, I'm told, many Europeans) have experienced the quintessential form of begging for scores: the car dealership that posts a sign showing the satisfaction survey with a big circle around the top score. I was really disappointed recently when, after an otherwise outstanding purchase, a luxury car salesman showed me the survey I would be receiving from "corporate." He circled two particular questions. "Mr. Markey, if even one customer gives me less than a 10 on either of these two questions, I can lose my status as a top seller, and I will lose access to the best cars. I will be out of the running for a bonus, and it will make my job harder next year."

Disaffection and disengagement. If you don't have a stable, reliable metric before you begin working it into your compensation system, get ready for a real firestorm. You're likely to run into two critical issues:

• Unstable scores. Problems with sample selection, changes in the mix of customers, differences in response rates, and inadequate sample sizes can lead to scores that fluctuate wildly or are otherwise unreliable. That will infuriate anyone whose pay is tied to the scores.

• Poor understanding of links between scores and behaviors. When you begin to quantify customer feedback, employees and managers may not know how to move them in the right direction. Until your team has some sense of control and influence over the outcomes, a compensation system based on a new score may seem unfair or arbitrary. It can lead to anger, finger-pointing, and defensiveness, rather than to a focus on delighting customers.

Enterprise experienced many of these problems in the early days of ESQi. Branches that received low scores, recalls CEO Andy Taylor, "ripped the measurement, the survey questionnaire, and the sampling technique behind it." Over time, however, the company refined its methods, learned how to prevent gaming, and helped people understand the close links between ESQi scores and business performance. As a result, people in the organization now trust ESQi as a fair indicator of their performance.

Ultimately, most companies should work customer feedback scores into their incentive compensation systems. From executives all the way down to customer-facing employees, that can help keep everyone focused on how to earn the loyalty of customers. An incentive compensation plan that places high value on earning customer loyalty also sends a great message to the organization about what the company values. But before you take this step, you'll need to meet five essential preconditions, which I'll discuss in my next post.

This series of posts from Rob Markey highlights the ideas in a new book by Fred Reichheld and Rob, The Ultimate Question 2.0: How Net Promoter Companies Thrive in a Customer-Driven World (HBR Press).