Tuesday, April 04, 2006

Business Tip: Measuring Customer Satisfaction

If you ever worked in a customer service oriented business, you know the importance of measuring customer satisfaction (CSAT). By in large, most industries want happy customers to get repeat business and better word of mouth marketing. What's a company with no repeat business and bad word of mouth marketing? A bankrupt one. Whether your company is a clothing store, or a computer store -- you want to know what the customer thinks of you.

Proper steps of measuring customer satisfaction (CSAT):
1. Find your customer population.
2. Get a random sample.
3. Survey the random sample.

Find Your Customer Population
Let's take for example that you own an Internet service provider with 1 million customers. You want to measure whether your customers are happy with the service. To find the correct population, you need to remove yourself and other non-paying accounts (ie: employees).

Get A Random Sample
A random sample is a subset of the population that is non-bias and represent all possibilities. Let's take for example the ISP company that has 1 million customers. We need to get a sample size from it without choosing from a "bias" approach.

A common "bias" mistake that ISP companies use to measure CSAT is to survey only the customers that contact the help desk or customer service desk. This approach is flawed. The majority of the people that contact the help desk or customer service desk has problems and they are already not satisfied in one way, shape or form.

Survey the Random Sample
Surveying requires proper wording, and proper timing. Your wording and scoring should be easily understood and should be non-biased against or for the company.

One of the worst CSAT scales I've ever seen came from a former client where they measured CSAT surveys from 1 to 10. This subjective measurement does nothing for you. Many people consider "5" as adequate and "6" as good. A preferable approach is: "Bad, Fair, Neutral, Good, or Very Good".

Timing also is important. For instance, if you're doing phone surveys -- you really don't want to survey your customers when they are at work and they are not answering their home phone. Neither would you want to survey people when they are eating dinner or sleeping. For mailings, you would probably want to survey customers right after the peak seasons, where you want to measure your "true" customer base.

Surveys can be a great and positive tool for your company, especially if it is done right. Consult with a statistician or a math professor if you think you need to start surveying customers. A math wiz can tell you about other survey factors such as margin of errors and confidence levels. If you fail to consult a credible third party before you survey, you may fall into what I call the "survey trap" -- where you make decisions on flawed data.

On a personal experience, I've seen clients in the survey trap and I think to this day, they are still in it. They are constantly terminating contracts, employees, and other problems for the sake of percentages and numbers that show up on their computers. There are legal consequences for these issues as well -- terminating employees for bad survey data can cost millions in unemployment hearings, wrongful termination suits, and co-employment lawsuits [co-employment is when a company has some control in a contractor's business/employees].

No comments: