Video: Why do you need to reduce the number of questions in surveys to ONE?

In the last weeks, we’ve been home more than usual, more people started to communicate through digital sources, and more brands did the same. Having more interactions with customers equals more opportunities to collect feedback from them, and hopefully, make the best use of this feedback to design better experiences. Unfortunately, and I would say, incredibly, many brands still use outdated methods to collect feedback from customers. What is an outdated method that I see daily, and that you should stop now? I would say, it is sending surveys that contain a lot of questions.

Last week I received three survey requests, one from an airline, one from a telco/communication company, and another from an HR tool. The one from the airline was the craziest one, with 5 pages of questions. Seriously, what’s the problem with you? The one from the HR tool contained 16 questions, therefore doable in approx. 8 minutes. Obviously too much.

First of all, just for fun, you can put yourself on a scale of self-evaluation to understand how outdated your methodologies are today. If you send out surveys you are already a bit outdated, because today there are so many ways to collect feedback that is voluntarily generated, that if you do it well, you don’t need to ask for more. However, going back to our scale, If you send our surveys with more than 5 questions, you are somewhat outdated, if you send out surveys with 10 questions or more, you have some problems. But if you send out surveys with 3 or 4 pages of questions you are a dinosaur with some problems. Which problems?

There are 3 immense problems that arise if you keep sending surveys with lots of questions. Ten questions are already way too much! 

1. People don’t respond. What’s the purpose of dedicating my time to fill in a survey with so many questions? To give you an idea, the number of people that leave a review on e-commerce, if asked at the right moment, is around 10 percent. If you send them a survey with 10 questions this percentage drops to less than 1%. Long story short, if you send out surveys with many questions you are losing 88% of customer feedback.
2. Problem number one was already enough to change your perspective? Then let me take it from there, and add that that 1% of feedback that you may collect is not as valuable as you believe. Why? Because if you ask many questions you are adding bias to the responses. It’s almost impossible not to add bias, even if you are an expert in creating surveys. Let me give you an example. In the airline survey, one question was related to the catering, to what I ate on board: “how would you rate the quality of the snacks/meal?”. What if I was sleeping, and I didn’t eat. What if I was not hungry, or what if I ate the snack, but its quality was absolutely irrelevant for my satisfaction? If you add questions, you set yourself for failing to interpret customers’ opinions.
3. If you ask questions, you are filling the time of the respondents with things you, and your stakeholders, want to learn, not with what respondents want to say. This is absolutely unacceptable from a customer experience standpoint. It’s like you volunteer to help, and you are obliged to do it in a way that is not what you would do. Customers’ time and feedback are gold, and they need to be valued, and not abused. Furthermore, you also prevent yourself from learning what you don’t know, which is even worse!

• You can ask two questions only, one closed and one open-ended. The closed one can be a score from 1 to 5 or one to ten. The open-ended question should basically explain the “why” behind the score. So a good copy can be: “Could you tell us more about your experience?”
• Use a tool to process the data. If you collect unstructured feedback you end up having the problem of analyzing them. If you collect dozens you can still analyze them yourself by just reading through them, but if you collect hundreds or thousands of feedback you would need a tool to process this data. Text analytics capabilities of the most famous survey collection tools are not sufficient, and you would need a specialized natural language processing software.

For more interesting content, click here.