Editor's note: David Lavinsky is a research analyst with the Customer Satisfaction Strategies Division ofFindlSVP, New York City.

As companies strive to increase their profitability, they are placing a greater emphasis on customer satisfaction measurement (CSM) studies. These studies can do just that: increase a company's bottom line by showing them how to both retain existing customers and attract new ones. Unfortunately, methodological flaws prevent many CSM studies from achieving this goal.

A good CSM study identifies the key issues that drive the customer in choosing a supplier and identifies the purchaser's loyalty rationale. However, this is not the goal of the study. The goal is to allow a company to optimally allocate its resources to maximize customer satisfaction. Fortunately, the two go hand in hand; the company reallocates its resources to the areas deemed most critical by the study.

Too many CSM studies are being conducted without this goal in sight, and with most, you can never achieve it, because these CSM studies measure satisfaction based on expectations. Expectations vary from issue to issue, so such studies only allow you to monitor customer satisfac­tion. They do not allow you to act on the results to maximize satisfaction in the future. The following ex­ample will help explain the downfalls in measuring satis­faction based on expectations.

Let's say a company determines that its customers find two main issues important when choosing a supplier: product quality and customer service. As a result, the questionnaire for the company's CSM asks the following question, "For the following two issues, please tell us whether Company A (1) exceeds your expectations, (2) meets your expectations, or (3) does not meet your expec­tations."

Suppose the company receives a "1" for product quality and a "3" for customer service from a customer. Further­more, suppose that the company had the same questionnaire last year and received a "3" for both issues from that customer.

The only result of this survey would be a low quality comparison of satisfaction versus the last study. It won't help to optimize the allocation of resources because expectations are a relative measurement and thus do not stay constant from issue to issue. Thus we cannot tell if the customer is more satisfied with one issue over another. Even though the company received a higher rating for product quality than customer service it does not mean that the customer is more satisfied with product quality. Maybe the customer didn't care much about product quality and expected a horrible product only to find that it was fair; this would have exceeded their expectations. On the other hand, the company's customer service could have been excellent, but since the customer knew this and expected even more, they responded "does not meet expectations" even though they were quite satisfied.

Another downfall of measuring satisfaction with ex­pectations in CSM studies is that they also fail to stay constant from competitor to competitor, even within the same issues. Sticking with the previ­ous example, let's say that Company A also asks the customer to rate Company B on product quality. If the customer expected a higher level of quality from Company A than Company B, the expectations scale would once again render a com­parison impossible. The customer could give a lower expectations rat­ing to Company A, even if they were more satisfied with A than B. This would happen if they expected even more from Company A than Company B. (Another methodological downfall of many CSMs is that they fail to gather insight into the competition. It is extremely diffi­cult to increase your profitability and/or market share if your are not aware of your competitive situa­tion.)

Fortunately there is a way to achieve the goal of CSMs. It entails substitut­ing questions measuring satisfaction of the key issues based on expecta­tions with questions asking the customer's per­ceptual and im­portance ratings of them. Instead of asking if the customer's ex­pectations were met on product quality and cus­tomer service, ask two ques­tions. First, how important is each issue to the customer? Answers may include (1) Not impor­tant, (2) Somewhat important, (3) Very important, or (4) Critical. Second, how well does Company A perform with regards to each is­sue? (1) Very poor, (2) Poor, (3) Fair, (4) Good or (5) Excellent. (Note that both importance and per­ceptions scales can vary based on the specific project.

As opposed to the expectations questions, the importance and perceptions questioning provides a comparable measurement for each issue. Thus the company could de­termine whether its customers are more satisfied with product quality or customer service. It could also determine which issue is more im­portant to its customers. Further­more, it could accurately compare its performance to its competitors.

Such a questionnaire would lead to the goal of optimally allocating the company's resources to maxi­mize customer satisfaction. (Obvi­ously, the actual questionnaire would address more than just two issues.) This goal could be achieved in two ways. The first method would involve constructing a quadrant analysis (see chart). This analysis would map each issue on a grid with importance and perceptions as the axes. The company could deter­mine its key strengths (issues with high importance and high percep­tions) and its key weaknesses (is­sues with high importance and low perceptions). The company could furthermore reallocate its resources from the overemphasized issues (is­sues with low importance and high perceptions) to the areas of weak­ness. Issues low in both importance and perceptions would be monitored over time.

A second method to optimize the allocation of resources with the re­sults of the importance and satis­faction questions is to construct a mathematical maximization model. This simply consists of multiplying the importance of each issue by the maximum satisfaction rating ("5" for "excellent" would be the maxi­mum score in this scenario). The sum of these products would give you a figure representing 100 per­cent customer satisfaction. Since each issue would be given a differ­ent weight according to its impor­tance, you could then determine (by plugging in estimations of increased perceptions if increased expendi­tures were made in each area) how to best allocate your resources to come closest to this 100 percent goal.

Expectations measurement is only useful in CSM studies when it is solicited by focused questions that are mutually exclusive (not to be compared with other questions). Such questions could include "How often do you expect a visit from your sales rep?" or "How long of a lead time do you expect between ordering the product and delivery?" However, measuring satisfaction based on expectations only allows you to determine if the satisfaction of your core customers is increas­ing or decreasing over time. Since this methodology fails to allow meaningful comparisons across is­sues and competitors, it does not allow companies to proactively im­prove their customer satisfaction in the future. By utilizing importance and perceptions ratings you can determine where your resources will be most useful in ensuring the sat­isfaction of your customers.