Better Market Research
May 1, 2006 12:00 PM, By Don Kreski
The Customer Satisfaction Survey
For a larger image, click here
You see them in restaurants, airports, theaters, and department stores. Sometimes they irritate, sometimes amuse, but just as often, they’re ignored. Thousands of companies use customer satisfaction surveys, but do they do any good?
Susan Lewis, CEO of Lewis Sound & Video of Waukesha, Wis., says she feels they definitely do. “Our business comes in mainly from referrals,” she says, “and I need to know how a customer thinks of us at the end of the project.” The Lewis Sound survey, reproduced on this page, is short, simple, and effective.
Brief customer response cards like this are one of the most commonly used market research tools. Still, their value depends on how you use them after you get them back from your customers.
The wrong way to use a satisfaction survey
The first thing you need to know is that satisfaction surveys rarely or never give companies a scientifically measurable result. That’s not to say they can’t be extremely valuable, but you do need to think about what your results mean.
One problem is that the survey takers are largely self-selected. Managers rarely select a random sample of their customers, nor do they do their best to make sure every survey they issue is returned. Instead, they hand out the cards to everyone and accept whatever comes back. Under those conditions, it’s impossible to know what those responses represent.
We used to do this type of survey at one of the AV dealers I worked for. There were six or seven questions, each answered by checking a box on a five-point scale, where 1 meant very unsatisfied and 5 meant very satisfied. Our sales manager had the cards delivered to a secretary, who would record the results, average them, and report to him monthly.
He used the averages as the basis for a competition between departments. In March, sales might get 4.15 for overall satisfaction, service 3.92, and rental 4.35. In April, sales and service each might improve .2 or so while rental dropped to 3.99. People took these numbers very seriously, but since we didn’t control the sample we were measuring the differences had no real meaning.
There’s an even bigger, conceptual problem as well. It turns out you can’t take a meaningful average of this kind of scale.
Say you do a survey like this and 15 people return the cards. Five give your company a 1 for miserable service. Five give you a 3 for meets expectations and the remainder a 5 for exceptional service. What do you get for an average?
That’s right, a 3, and that meets expectations hides the fact that 1/3 of your customers are so unhappy they’ll never talk to you again. Wouldn’t it be nice to know the true proportions, so you can try to find out what you’re doing so right with some people and so wrong with others? The right way to use a satisfaction survey
If you truly want to learn from this kind of survey, start with these steps:
1. Forget about averages. Instead, report how many fall within certain ranges. With our example, it might be best to report 66 percent were satisfied or very satisfied, 33 percent very unsatisfied.
2. Leave lots of space for hand-written comments, and make sure those comments are part of any report.
Susan Lewis says that she limited her survey to five questions, with room for comments on each, plus room for overall comments. “I started out with checkboxes because that makes the card easily approachable. Someone can look at it, mark 5 boxes, and move on. I really wanted the comments, but if I had put the comment line without the boxes, most people would look at it and say, ‘Forget it, this takes too much energy.’”
3. Ask people to sign the cards and ask for a phone number. Then set up a system where unhappy customers get help and you have confirmation that they do.
“Most of our customers,” Lewis says, “love us and said wonderful things. But every once in a while they’d say something where I’d go ‘Ooh.’ I would call them at that point and say, I apologize, and what can I do for you?”
4. Consider calling the ones who are exceptionally happy as well, then reward staff members who served them.
5. Be very careful how you send and collect your surveys. I was good friends with many of the sales people at the AV dealer I worked for, as well as some of the drivers and service technicians. After we had been doing the survey cards for several months, I realized that many of my friends were throwing away certain surveys. If they knew a customer was unhappy and thought they might be blamed, they just didn’t hand out the cards.
It was also possible, given the system we used, that some of our people filled out the cards themselves. Security is an issue any time you reward or punish people based on customer satisfaction.
6. Consider an email survey or an online survey with an email invitation to participate. Email is faster and easier to administer, and it’s possible to make it quite secure.
What you have to gain
The unfortunate truth is a lot of these surveys are meaningless and are essentially ignored, even by the companies that issue them. You know that’s true if you’ve ever voiced a complaint, then written down your name and phone number, hoping someone would call.
Lewis, on the other hand, says, “They’re a kick to get back. We’d get them in the mail and it would give us a high.” Most AV integrators can expect to get a lot more positive comments than negative. On the other hand, nearly everyone can uncover flaws in their systems and perhaps find out if certain employees treat customers poorly. You’ll also find specific unhappy customers and have a chance to satisfy them before they go elsewhere.
In short, be prepared to learn. It’s fun to hear when you’re doing well, and it’s a good thing to uncover problems. They are the reason you’re not growing as fast as you want to. Fix them and your company will really take off.
Questions? You can reach Don Kreski at www.kreski.com/contact.html.