While customer surveys have become a standard part of nearly every transaction, most surveys fail to engage; even worse, they fail to collect reliable data. Wondering how to improve customer satisfaction surveys and get yours to stand out? Read on for the biggest survey mistakes and solutions to common survey problems.
“Quality is never an accident; it is always the result of high intention, sincere effort, intelligent direction, and skillful execution,” said WWII Marine and Medal of Honor recipient William A. Foster.
When it comes to surveys, quality is paramount—because only high-quality surveys measure customer and employee perceptions accurately. With a solid read on perceptions, you’ll be able to sharpen your company’s strategy.
I hope you don’t stop reading here, but if you do, here are four key points I want you to takeaway:
- Forget aimless surveys
- Stop survey biases
- End nonsense data
- Mine customers’ comments
Forget Aimless Surveys
Before you write a single question, gather your team and get clear on what the goal of your survey is. Are you looking to improve the quality of your customer service? Cut down on shipping mistakes? Understand how your products are used?
Here’s an example of what NOT to do: Alaska Airlines once sent me a 94-question survey that required at least 30 minutes to complete! Surveys should be short and focused, with every question working in a single direction.
And don’t forget to ask how you will use the survey data within the year. If you can’t act on the data, it’s not the best use of your resources—or your customers’ time!
Stop Survey Biases
To achieve quality output, you need quality input. For surveys, this means asking neutral questions that will result in factual data.
Biases can occur intentionally or unintentionally. They can be subtle and pop up in dozens of ways. If you are looking for advice on improving customer satisfaction surveys, eliminating bias is one of the most critical areas to consider.
1.  Leading Constructs: These questions assume your respondents feel a particular way. For example, don’t ask, “How delighted were you with the customer service you received today?” This assumes your customers are at least somewhat delighted. Maybe they weren’t delighted at all! |
 Solution: Instead of asking how satisfied/happy/delighted, etc. your customer is, ask, “Please rate the customer service you received today.” This is a neutral construct. |
| 2. Social Pressure: Good surveys work hard to create an environment where customers can answer objectively. Why? Because we all feel social pressure to respond positively if the person we’re rating is present or quid pro quo is implied in any way. |
| Solution: Text or email your survey an hour or two after the interaction has completely ended. Asking questions while the server is standing at the table or while the rep is on the phone irritates customers and results in bogus data. |
| 3. Double-Barreled Questions: These types of questions combine two issues into one. For example, “Was your customer service representative polite and proactive?” asks about two distinct topics. Questions like these lead to flawed data and obscure the next steps you should take. |
| Solution: Make sure your questions ask about one issue at a time. |
| 4. Insufficient Survey Sampling: This is when your survey data only comes from certain types of customers and doesn’t reflect your customer base at large. Examples of customers who might be overrepresented in your data are those with free time on their hands or an axe to grind. |
| Solution: Use methods outside surveys to ensure your entire range of customer populations is evenly reflected in your response data. |
| 5. Overly Small Samples: Sometimes survey data is misrepresentative simply because you don’t have enough responses to give you reliable facts. For most populations, you need 370 responses to achieve statistically significant results. However, your customer base consists of many populations, so it’s best to work with our sample size calculator to determine how many responses you need from each group. |
| Solution: Use incentives to increase your survey response rate. Also, take a hard look at your email, text, or web invites. See if they can be better crafted and more persuasive. |
End Nonsense Data
When clients start thinking about improving customer satisfaction surveys, they often overlook the importance of good survey usability. Surveys that are too difficult or overly time-consuming cause customers to stop before they’ve barely begun.
| 1. Too Many (Or Too Few) Text Fields: If your survey doesn’t have enough text fields for customers to expand on their answers, you lose valuable information. On the other hand, if you ask too many text questions, customers are overwhelmed and exhausted. |
| Solution: Interview your test users about your proposed open-ends. Do your questions feel ponderous? Unclear? Are there too many? Are they in the right places so users can elaborate where they want to? |
| 2. No Logic Gating: Customers don’t have unlimited energy and time. The longer your survey, the lower its completion rate. |
| Solution: Wireframe your survey logic. Then, critically review your wireframe with a team prepared to poke holes and weaknesses in your logic. Your objective is to maximize question relevance and to be as concise as possible. |
| 3. Generic Questions: Avoid asking the same questions that Hilton Hotels, Delta Airlines, and Enterprise Rental Car all ask. For example, the “How likely are you to recommend…” question might have a place in your survey, but don’t necessarily assume that. Your company is unique. To some extent, your questions should be too. |
| Solution: Put on your creative cap and think like a customer. Find ways to show customers you’re listening and care about their insights. You want your customers to consider your questions, not roll their eyes and say to themselves, “Well, I’ve seen that 100 times before.” |
| 4. Vague Questions: These questions contain confusing language, bad grammar, or anything that makes your survey burdensome or unclear. |
| Solution: Hire an editor to proof your questions — and, as stated earlier, test your survey with users. |
| 5. Jargon: Maybe you know what you’re talking about when you say things like you’re going to “leverage our core competencies this year,” but don’t assume your customers know your shorthand and internal ways of talking. |
| Solution: Avoid acronyms, branded phrases, or any internal lingo. In a word, use plain language—because if your questions are misunderstood, you’ll end up with false data. |
Mine Customers’ Comments
Text answers are a gold mine of valuable information. But to glean actionable insights from open-ended text, whether you use AI or not, your text needs to be tagged (and then cross-coded and quantified.)
Clients sometimes ask why they can’t just read the comments to learn what customers say. There are a few reasons why reading text doesn’t glean the same insights as text that‘s been tagged:
- Our short-term memories can’t store more than about seven comments, so it’s impossible to quantify or build a visualization from the comments
- Reading text is an entirely subjective experience; there’s no way to achieve replicability in your findings.
Here’s how it works: begin with statistically valid samples. Then, build a coding framework. Next, use several analysts to apply that framework and use cross-coding to measure replicability.
When your findings are replicable, you can prove your conclusions are verifiable. And, of course, replicability builds trust in your findings.
Customers’ comments give you actionable insights you can’t get from rating questions alone.
In addition, with text, you may learn about issues you never thought to ask about, which you can add to your next survey making it more valuable.
Keep in mind that AI is an excellent option for large datasets, but you still need to train and test it—and not just once—because customers’ text, just like your marketplace, company, and products, is constantly evolving.
Also consider that if your dataset is small or specialized, AI will probably only work in a limited way; use a research team to tag your data instead.
More Tips for How to Improve Your Customer Satisfaction Surveys
Maybe you’ve eliminated the most significant errors from your survey, but it’s still not delivering the results you need. Keep reading for advice on how to improve customer satisfaction surveys.
Let Customers Answer Anonymously
One of the first questions on your survey should always ask respondents if they would like to (or need to) answer anonymously. Of course, it’s human nature to want to know who’s behind the ratings, but customers often don’t want to be bothered with potential retaliation or follow-up calls probing for more information.
In fact, when given the option, around half of respondents will generally choose to take a survey anonymously. And every customer’s feedback is valuable, even if it’s anonymous.
Craft Your Email Invites with Care
The email (or text) invitation is your respondents’ first impression of your survey. Unfortunately, all too often, the invite is an afterthought. And yet, how you frame your survey impacts how many customers will bother to double-click on its link.
As a best practice, have your company’s CEO write the first email to survey respondents, letting them know they will hear from an independent survey consultant company. This signals two things to your respondents:
- You are taking the survey seriously
- They can answer honestly
Write Questions That Echo How Your Customers Speak
When was the last time someone told you they “somewhat agree” OR “neither agree nor disagree” with your opinion? This has probably never happened because no one talks like that. So don’t give your customers awkwardly phrased survey answers.
Instead, try to write your surveys as though you’re talking in person. Rating scales written with familiar words are easier for users and tend to collect more accurate data because customers know what you’re talking about.
Let Customers Opt Out of Irrelevant Questions
Don’t force your customers to answer questions that don’t apply to them.
For example, if you ask a customer how often they go to the gym when they don’t belong to a gym, they are forced to give a false answer.
Instead, give your respondents an opt-out answer like “None of the Above,” or ensure the question isn’t required. Even better, with sound logic gating, customers will only see questions that pertain to them.
Ask “Answerable” Questions
There are some questions that customers can’t answer. We’ve seen surveys ask respondents to “Rate the blank space to text balancing on our website.” These kinds of questions don’t belong in customer surveys; instead, they should be answered by a design team.
Follow Up with Everyone (or everyone you can!)
Follow up with every customer or employee who took the time to take your survey.
Even better, surprise them with a small gift like a $25 gift card to Starbucks, thanking them for their effort. For customers who give their feedback anonymously, there is not much you can do regarding follow-up. Still, the end of your survey should always have a clear note expressing genuine appreciation.
Prioritize Analysis
A customer survey aims to learn what your customers think daily, monthly, or yearly and determine what to change. If there’s no team in charge of finding the story in the data, your effort (and your customers’ time) goes to waste. The solution is to make analysis a priority.
- Solution: Don’t call hunches facts. Sometimes, signals in the data are just random chances and are not factual findings. Test your data carefully to make sure you aren’t drawing conclusions that the evidence doesn’t
- Solution: Never ignore variance. You fail to get a nuanced picture if your analysis only looks at averages and not the distribution of scores. Segment your data to see which customers score low or high, for what products or services, and under what circumstances.
- Solution: Present data clearly. Make sure your graphics support and contextualize your data accurately. For example, don’t use a pie chart when your answers don’t add up to a whole, and never compare dissimilar factors in a single chart.
Great Survey Questions to Ask
The best questions to ask are the ones that are right for your company. Nevertheless, examples of questions can be an inspiration for improving your customer satisfaction surveys.
- Comparison Questions: “Who is our closest competitor?” Then, follow this question with, “What are areas in which <COMPETITOR> performs better than Company XYZ?” Your customers are constantly comparing your company to your competitors, and you should be too.
- Ranking Questions: These questions typically include two columns. The left column lists key factors or departments. Your customers select the issues that matter most to them by dragging those selections to the right column in priority order.
- Identify Room for Improvement: “What could XYZ do to make your experiences easier or better in some way?” Always look for ways to improve.
Customer Surveys: From Aimless to Actionable
High-quality surveys are never an accident; they result from careful planning, writing, and execution.
Are you struggling with how to improve customer satisfaction surveys? Want to make sure your insights are actionable?
At Interaction Metrics, we offer a free 45-minute customer experience consulting session. It’s the perfect way to fuel your thinking. If you want to discuss customer feedback, we guarantee you’ll come away with at least a few ideas to improve your own surveys.
