The Displayr Guide to Customer Feedback Survey Analysis

Welcome to our complete guide to customer feedback surveys!

If you’re looking to better understand your customers and make data-driven decisions based on their needs, then you’ve come to the right place. We’ll take you through the entire process, from designing your survey to interpreting the results.

An Overview of Customer Feedback Surveys

Let’s start with a simple question: what is customer feedback?

Customer feedback can come in many different forms. A retail company could prompt customers for a rating after their product is purchased, and also scan review aggregators. A tech company could ask for a simple thumbs up/thumbs down reaction from users on their website, and follow-up with a longer survey.

These are all important and relevant forms of customer feedback. However, this guide will focus specifically on customer feedback acquired through surveys.

What is a customer feedback survey?

A customer feedback survey can take many forms. It can be as simple as a single Yes/No question, and as complex as a multi-level questionnaire. The size and structure of your survey should depend on your product, your user-base, and the kinds of insights you’re seeking.

Customer feedback questions can be divided into two broad categories:

  • Open-ended questions: those that allow users to answer in their own words. This usually produces more detailed responses but can be difficult to analyze in large numbers.
  • Closed-ended questions: those that have a restricted response format, like multiple choice or a 0-10 scale. The most common types of closed questions are Net Promoter Score (NPS) questions and Customer Satisfaction (CSAT) questions.


Why customer feedback surveys are important

Customer feedback can inform your decision making in many ways. It can identify groups that are vulnerable to churn, pinpoint parts of your product that require improvement, and even unearth some fresh ideas you would never have thought of.

Acquiring a new customer is almost always more expensive than retaining an existing one, so it makes sense for everyone to make customer retention a key priority. By carefully and regularly analyzing your customer feedback data, you can better understand the needs of your customers and increase brand loyalty and retention.

Survey design

Before diving in, you need to ask yourself two questions:

  • How many questions should I ask?
  • What questions should I be asking?

There are no definitive answers to these questions, but there are always a few factors you need to consider.

Asking the right questions

Right out of the gate, there are two closed questions you should be considering:

  • Net Promoter Score: On a scale of 0-10, how likely are you to recommend this product/service/company?
  • Customer Satisfaction: On a scale of 0-10, how satisfied are you with this product/service/company?

These two basic questions have a long history in market research, and together they can form the backbone of your customer-relationship strategy. Other closed-ended questions regarding specific attributes of your products and comparisons to your competitors can enrich the overall feedback you receive.

Open-ended questions allow users to elaborate on a response they gave to a closed-ended question. For example, you can start with a closed-ended question like, “On a scale of 1-10, how would you rate this product?” and follow up with an open-ended question like  “What did you like about the product”.

The combination of closed-ended and open-ended questions can paint a complete picture of how a user feels about your company, product, or service.

Limit the number of questions in your survey

It is tempting to ask as many questions as possible in a survey. After all, more information is always better. But asking too many questions in a survey could drastically lower your response rates since customers are less likely to complete your surveys.

As a general rule, you should aim for under 10 questions in any survey. If a question isn’t essential, then it should be omitted.

Ask well-defined questions

Do not require the customer to do any heavy lifting. The questions they are asked should not be ambiguous or open to interpretation. Your customer should immediately know what is being asked of them.

Also, try to ask laddering questions. If you ask how satisfied a user is with a product, you can continue by asking questions about specific components of the product. Grouping related questions will remove any ambiguity.

Common Do’s and Don’ts

Do Ask… Don’t Ask…
The same questions regularly Leading questions
Both closed and open questions Non-essential questions
Demographic questions Too many questions

Sending out your survey

Once you’ve designed your survey, it’s time to release it into the wild!

Here are some questions to consider:

  • Who should you survey?
  • When should you survey?
  • How often should you survey?

How to collect feedback from customers

The first thing you need is a contact list of users and customers. Hopefully, you have a collection of email addresses from users and customers that you can use to send out your survey. As long as you can get hold of them, there are many ways to contact your customers. Using a Customer-Relationship-Management (CRM) system to issue your survey is a popular approach, and there are also many third-party options to choose from. Many companies also have a built-in pop-up survey that prompts users on their website.

When to survey

A survey can be issued directly after a transaction is made, or sent out to users at regular intervals. A survey sent after a transaction should try to gauge a customer’s first impression and perhaps ask how they found out about the product. A regular survey should ask questions that can be used to track customer feedback metrics over time.

How frequently you should survey depends on your product and customers. If your product is constantly changing, then you want to continuously track the response to each change. And if your customer-base is rapidly growing, then you should be eager to find out how the new users are responding.

Who to survey

Though the terms are often used interchangeably, there is an important distinction between a “user” and a “customer.” A customer is already paying for your product, while a user may be undergoing a free trial or using a freemium option. It is vital to survey both groups. Retaining paying customers and convincing non-paying users to become paying users are both crucial to your success.

Survey sample sizes

Major studies usually involve a sample size of at least 100 participants. But of course, that is not always possible in the real world. When it comes to surveys, more is always better. A larger sample size delivers stronger conclusions and produces more statistically significant results. You can maximize your sample sizes by making sure you are always collecting email addresses from users, customers, and potential customers. In the meantime, try to design a survey that maximizes response rates.

Increasing response rates

It can be notoriously difficult to convince a customer to fill out a survey. They can be time-consuming, confusing, and complicated. You can improve response rates by making your survey as convenient as possible for your customers. Restrict the number of questions, remove any ambiguity, and make the survey feel intuitive.

Preparing your survey data

Consider yourself extremely fortunate if your data is ready to analyze without any data preparation. Chances are there are a few issues that need to be taken care of before diving into the analysis process.

Dealing with missing data

Missing data is a common problem in real-world data analysis. Perhaps a question doesn’t apply to a subset of customers (how old is your child?), or maybe a question makes some customers uncomfortable (What is your income?). Before determining how to deal with missing values, we must first determine why they are missing in the first place.

A common approach is to simply exclude missing values from the analysis. This is certainly the easiest solution, but it’s rarely the best one. Statistical methods can be used to predict missing values based on other responses in the dataset.

Cleaning messy data

There’s a good chance your data set will contain some formatting issues. Perhaps a numerical score is stored as text, and maybe your text contains HTML tags. Analysis software can sometimes be smart enough to automatically make those changes for you, but it’s always good to always make sure before proceeding.

Recoding data

There are cases when recoding your data before beginning your analysis can save a lot of time. For example, recoding to your NPS data to a (-100, 0, 100) format makes calculating scores a breeze.

Analyzing your survey data

Finally, it’s time to analyze! This is where the magic happens. It’s where data is transformed into information, and information into insight.

Calculating NPS and Customer Satisfaction

Customer Satisfaction and Net Promoter Score are two of the easiest metrics to calculate. The Top 2 Box score is the commonly used measurement for customer satisfaction, though a simple mean or median can still be informative.

The Net Promoter Score follows a more rigid formula:

Percentage of Promoters – Percentage of Detractors


  • Detractors: respondents who gave a score between 0-6
  • Passives: respondents who gave a score between 7-8
  • Promoters: respondents who gave a score between 9-10

Customer Satisfaction and NPS are great starting points for your customer feedback analysis.

Industry benchmarks

Before rushing to judgment, take a moment to research the benchmarks for your industry. You can gauge where you stand by comparing your NPS results to the industry averages.

Industry Average NPS
Healthcare 62
Retail 54
Hospitality, Travel, & Restaurants 53
Manufacturing 51
Automotive & Transport 49
Financial Services 46
Construction & Engineering 45
Customer Services 42
Insurance 42
IT & Software 41
Banking 37
Media 34
IT Services 33
Telecommunications 24

Data modeling and algorithms

Let’s take a step beyond NPS and customer satisfaction. There is a whole world of data modeling techniques and algorithms for us to venture into. Regression modeling can be used to identify the drivers of customer satisfaction and cluster analysis can show how feedback varies across different market segments.

There’s also the issue of statistical significance. You may find that customer satisfaction scores are increasing or decreasing, but you need significance testing to separate the real changes from the noise.

Sentiment analysis and Topic Modeling

When it comes to open-ended questions, conventional data algorithms can only do so much. When manually curating and coding responses are not feasible, there are text processing algorithms that can help categorize feedback. Sentiment analysis can measure the tone of your feedback and topic modeling techniques can identify groups of words that commonly appear in customer responses. These algorithms have serious limitations and can rarely replace human judgement. For example, they are awful at detecting sarcasm and can mislabel unusual phrases.

Data visualizations

Data visualizations do two things: they present your findings in a visually appealing way, and they help you better understand your own data.

There are many ways to visualize your results. A scatterplot is a great way to demonstrate how satisfaction scores have evolved over time, and a bar chart can be used to compare those scores across different age groups. A correlation chart can depict the main drivers of satisfaction, and a pie chart can show the make-up of satisfied customers. To find out what works with your data, you’ll need to get your hands dirty and try them for yourself.

Monitoring customer satisfaction over time

Customer feedback data is especially useful when tracked at regular intervals over time. By monitoring your NPS and Customer Satisfaction scores, you can attribute changes in those scores to specific changes in your product, company, or brand. It also gives you more data to work with. A single survey could prove to be an anomaly, but you can be confident that a series of surveys over a prolonged period will be fairly accurate.

Insights and Actions

So you’ve made sense of your customer feedback data. Now it’s time to communicate your results and take actions based on your findings.

Sharing your results

The adage “Show, don’t tell” applies. It’s always better to show how satisfaction scores have changed over time rather than to simply state it. Make full use of hard numbers and data visualizations to present your findings.

Be mindful of your target audience. A report for customers and the general public is intended to showcase your results and let customers know that their voice is being heard. A report for management should be tailored to driving future success.

Taking action

The insights you gather from your customer feedback surveys should inform the way you move forward. You should identify groups within your customer-base that are vulnerable to churn and find the sources of their dissatisfaction. You should also find groups with high rates of satisfaction and search for potential customers with similar traits.

If a large subset of your respondents gave an NPS of 8, then they are just a nudge away from being promoters. On the other hand, respondents with an NPS of 7 could be on the verge of being detractors. You may want to specifically target these customers to improve your Net Promoter Score in the future.

Customer Feedback Survey Content

Get started with Displayr today

Not ready to chat to an expert? Download Displayr for free

Book a demo!
Customer Feedback Surveys
Customer Feedback Surveys
Customer Feedback Surveys
Customer Feedback Surveys
Customer Feedback Surveys
Customer Feedback Surveys
Customer Feedback Surveys
Customer Feedback Surveys
Customer Feedback Surveys
Customer Feedback Surveys
Customer Feedback Surveys
Customer Feedback Surveys

We created Displayr so that smart people can do their own data science: explore, model, visualize, and report. One app for everything. Easy for novices, powerful for experts.

Tim Bock, CEO