How To Create A Survey For Customer Satisfaction

There's a world of difference between well-designed and poorly-designed customer surveys. A great survey design doesn't just look appealing, it gets you the insights you need.

Kicking off your optimum survey design

When you want to survey customers, where's the best place to start? Our helpful list of 10 customer survey design rules begins long before survey questions are created. In fact, there are many considerations from demographics to timing. All help you improve survey response rates and the quality of data collected.

Check out our customer survey design infographic

In this post, we've detailed 10 areas of guidance for perfecting your customer survey design. These are summarised below in a handy infographic to keep close at hand. Use it whenever you need a structure (and inspiration) for creating your next customer survey.

The 10 customer survey design rules you need to follow

Let's look in detail at each aspect of customer surveys. Our 10 rules revolve around why, who, what, where, how and when – more or less in that order!

Define the Survey Purpose

So this is the 'why'. First, you need to clearly define exactly why you want to survey your customers. What are your goals from the survey exercise? What do you aim to learn from the survey? How will you apply the collected data to influence your business decisions?

These questions have big implications for survey design. For example, if your goals are very broad and wide-ranging, you'll need a survey programme to match. Or if you hadn't previously considered what to do with the results – now's the time to decide.

Identify the Target Population

Now we're getting onto 'who'. Here's where you define your target demographic. What characteristics do they share? For example, customers who have bought a certain item in the last 6 months. Or customers who have lodged more than 3 support tickets since a certain date. Other demographic splits could be along location, gender or age lines.

Once you have this picture, you can build an idea of what survey format or timing is appropriate. Likewise for language and other cultural considerations. We'll cover these in more detail in the other rules listed here. Also, you should consider segmenting your target survey population if it's particularly broad or large.

Ensure Adequate Sample Size

The number of customers your survey needs to be representative of the population you're targeting.

So first, work out how many customers exist in the total target population. Then work back to figure out how accurate your survey results need to be. Aim for a margin-of-error of around 5% (1% to 10% is typical). This indicates the likelihood that the answers derived from the sample set reflect the answers of the total target audience. Aim for a confidence (or 'significance') level of 95% (90% to 99% is typical). This is a measure of how likely it is to get the same survey results from another sample taken from the same target audience. Low confidence levels below 90% are not recommended.

This process should help determine what response rate you require. The anticipated response rate determines the number of surveys that need to be sent in order to achieve the required number of responses. The rest of the survey design rules in this list help optimize response rate.

Keep Customer Feedback Surveys Short

Who wants to fill out lengthy customer surveys? By designing your survey for brevity, you're doing your customers a favor. Making it as easy as possible to give you really valuable feedback.

So aim to keep the number of questions to as few as possible, ideally just one. One simple question is often the most effective, so only add more if there's a strong case for it. And keep the survey questions short too. The shorter they are, the more direct and easy to interpret. Also try to keep the number of selectable responses to as few as possible. Lots of response options can give you survey data that's difficult to draw conclusions from.

See what a 1 click survey looks like:

Be Clear In Your Survey Design Wording

As we've established in rule 4, customers appreciate surveys that are easy to complete. That necessitates making them easy to understand. You need to remove the potential for misinterpretation, confusion and ambiguity.

So always be clear and specific, exploring one simple idea at a time. Be aware that binary 'yes/no' questions will not provide options that many respondents want to give. So leave scope for nuance.

Avoid all forms of bias. For example, asking a 'leading' question that might skew results rather than getting an honest response. Another kind of response bias is self-selection bias, which is where results skew toward the experience of participants. Higher response rates lower the impact of self-selection bias.

Try getting into the habit of using words rather than numbers in survey response options. Using 'extremely likely', 'slightly likely', etc. will provide more representative answers than numbers.

You should also consider anticipated responses when composing questions. Is the respondent required to think or remember something? Will the question make the respondent feel uncomfortable? Where appropriate, provide the facility for respondents to enter text responses.

Lastly, avoid personal questions. If strictly necessary, ask such questions at the end.

Incentivise Responses

The use of small incentives can encourage survey responses in certain groups. The emphasis here is on small, rather than large, incentives. This is because incentives can, in some cases, increase self-selection bias. Even so, they are a potent consideration in the survey design arsenal.

Such incentives could include discounts, access to premium content or other privileges. You could even enter respondents into a free prize draw.

Select the Right Format

There are many examples of eliciting customer feedback. Keep an open mind on your options rather than just doing the same for each survey. The key question here is, what is the best format for your survey audience?

Straight away, we at Customer Thermometer are jumping up and down with email surveys including those embedded in email signatures. But you have more options to choose from. A tablet-based or even paper questionnaire could make sense in a trade show environment, for example. Or what about quick online polls, personal phone interviews and in-person focus groups?

Whatever format you choose, just consider whether it's right for your audience. Will it provide the requisite response rate? There's nothing stopping you using multiple formats, so long as you can collect the data accurately.

There are so many options when it comes to fun surveys, have a look around our icon library for ideas.

Test Your Survey

You won't know if your customer survey design works until it encounters customers. So rather than just let it loose, do some testing before live deployment. There are 3 main levels of test.

The first is to send the survey to yourself. Did it appear correctly and could you interact with it as planned? Second is to send it to a test group (e.g. friends and colleagues). Again, you're looking to verify question clarity and ensure survey presentation and format is optimal for the target audience. Third is to carry out a practice run using real respondents from the sample set.

A testing checklist might be helpful here so you can ascertain various aspects. These include the reliability of the survey and whether it returns consistent results. You should make adjustments based upon the test results, and send out when testing is complete.

Send yourself an example of a 1 click survey:

Choose the Right Time

Appropriate survey timing is essential and must be precisely calibrated for the respondent profile. There are some general rules about survey timings detailed here. There may be additional governing factors per audience segment. Some organizations collect responses over time to gather a broad range of feedback from various people (1 week minimum is recommended).

The most meaningful surveys are carried out soon after the customer experience the survey relates to. For example, asking customers about the arrival of shipments is most applicable directly following the shipment arrival. Also, consider surveying respondents at various stages of their customer journey to assess how their satisfaction changes.

It is bad practice to hassle customers for responses. However, most people will happily accept being reminded once and once only – this typically boosts response rates too.

Respond to Respondents

This last rule is commonly overlooked. It goes to the heart of behaving as a customer-centric organization. It truly reveals whether you treat customers as people or as numbers.

All it takes is a little politeness and consideration. By simply acknowledging and thanking every survey respondent you affirm their effort in helping your business. They didn't need to answer your survey, so respond in kind.

Your survey is likely to elicit negative feedback from some quarters. It's crucial that you identify and respond to negative feedback immediately. Always make personal responses to any feedback at the extreme ends of the spectrum. In other words, personally contacting the most negative respondents to address their concerns, and personally thanking positive respondents for any praise.

Start Designing A Survey Today

These 10 rules of customer survey design provide a framework to ensure excellent response rates and the information you need to improve customer experience. You need to put them into practice in order for them to stick.

With a free Customer Thermometer account, you can apply these rules right away. Designing a beautiful survey is easy in the guided setup. Get started today: