Done right, customer surveys can provide valuable insights into your customer base. Done wrong, they can alienate those customers and damage your company’s reputation for being easy to do business with. This article examines ways to ensure your surveys provide the best results.
Surveys are a great way to learn about your customers, but they can be labor intensive. And if poorly designed, they can produce useless or even dangerous answers.
Get the most out of your investment by following these four tips for creating effective customer surveys.
Segment: A good survey always groups responses into customer segments. Different types of customers want different things, and understanding these distinctions is vital. This information is lost if you lump everyone together.
Let’s say your survey shows that the average customer wants weekly deliveries, occasionally needs technical advice and uses both will-call and delivery services. It’s possible that most customers fit this pattern. It’s also possible that your customer base consists of two distinct groups. Service contractors buy small orders almost daily from your counter; project contractors place big delivery orders monthly or so. Averaging the responses from these two distinct groups would produce a very distorted picture of your market. If you instituted weekly deliveries, you’d only be satisfying a mythical average customer that doesn’t exist. It wouldn’t work for service folks, and it would be unaffordable overkill for the project guys. Oldsmobile went out of business trying to satisfy the average car buyer. The brand became a boring compromise that fit almost no one very well.
You can segment surveys several ways. One is to control distribution so you know which segment is responding. You can create different versions for large and small companies or for MRO and OEM accounts. If you use a kiosk at the counter or enclose a survey with invoices, you know responses will be from current customers. If the survey is online, you can include information in the link to indicate which email form or website page they came from.
A second way to segment is to put a self-segmentation question in the survey itself. Have them check the box for the type of business that most closely describes them. Don’t forget to give them an “other” option to fill in.
The third approach is to use the pattern of survey responses to define groupings of customers. If respondents tend to answer several questions the same way, you may have identified a segment.
Fair warning: In my experience, it’s pretty rare to get this level of clarity and consistency from a survey. A more successful approach is to develop a tentative set of segments with qualitative methods (e.g., customer interviews) and then use surveys to validate them and estimate their prevalence.
Keep it short. Seriously. Everyone starts with good intentions to make their surveys brief but somehow end up with 47 multipart questions. When the marketing group puts together the annual customer survey, every other department sees it as their only opportunity to insert their pet questions.
A good survey is designed to help you definitively answer one or two specific strategic questions. If you can’t articulate what you want to learn from the survey in a single sentence, you won’t be able to resist all the trimmings.
Long surveys are self-defeating. They kill response rates and produce unreliable data. By the third page, respondents are mindlessly clicking boxes so they can get back to work. It’s 2017; we all have the attention span of a gnat.
If respondents care about you, one to two minutes is a good maximum time for completion. If they don’t (i.e., they are prospects or transactional customers), that should be 20 to 30 seconds. It’s also good to
have the entire survey fit on one page so they can immediately see how much time it will take. If you resort to multiple pages, progress bars and “you’re almost done” messages, the respondents are probably checking out before they’re done.
Ask tangible questions. Different types of customer feedback have different purposes. Interviews and focus groups are great for getting qualitative insight and developing hypotheses about your market. Surveys are good for quantifying and confirming or invalidating those hypotheses.
Avoid questions that are ambiguous or open to interpretation. Words like “service” and “quality” can have very different meanings depending on the person and situation. The more concrete and specific the question, the easier it is for respondents to provide accurate and useful information. Many distributors like to use the Net Promoter Score because it measures how customers feel by asking a very tangible question: Would you recommend us to a friend?
Consider these three examples for measuring satisfaction with your fill rates:
“On a scale of 1 to 5, how happy are you with our service?” This question is OK, but responses will depend on mood and could relate to everything from the friendliness of inside sales to your returns policy.
“Compared with our competitors, how likely are we to have the items you want in stock?” Still fairly subjective.
“In the past 30 days, how many times did we not have what you wanted in stock?” Very specific.
One of my pet peeves is the “service matrix” question, where the customer is asked about the importance of various services (price, delivery, product range, courtesy, etc.) and how well the supplier performs on each. Inevitably the respondents mark every item as “very important” and supplier performance is either average or, if they’ve recently had a bad experience, horrible across the board.
Measure actions, not just words. Leading firms increasingly use behavioral surveys in addition to the traditional structured question and answer format. A behavioral survey collects information about customers based on how they act rather than what they say. The most familiar example is tracking activity and navigation on a website or app. This can tell you what new products generate the most interest, how many customers accessed information from mobile devices and which customers are most likely to ask for technical information and support.
Sophisticated online sellers continuously run experiments to gauge customer sensitivity to different elements of their value proposition, including price, payment terms and delivery. They offer limited time specials and measure the impact on click-throughs, conversions and sales. When combined with segmentation as described above, these behavioral metrics can provide great insight.
Behavioral surveys can’t measure everything, but they are generally more reliable. They capture a bigger sample and aren’t affected by respondents’ honesty or memory – or tolerance for taking surveys!