Keep online surveys short

13.Nov.09
by Dirk Knemeyer

Online surveys are one of the most commonly-used feedback mechanisms for businesses. And it’s no wonder: they are cheap to create, deploy, tabulate, report and share. They provide a degree of insight into how customers think and feel about your company, products and/or services. While getting valuable and insightful results is much more likely when the survey is being administered and interpreted by a trained professional, it is an unfortunate reality that most companies let untrained marketing or communications professionals attempt to create surveys. I’m writing today to try and help those intrepid souls.

While there are a variety of simple tips to help amateurs create better online surveys, the single most important by far is: keep it short. Most people will respond to a short, quick survey. Those same people will not respond to a long and arduous survey. Despite that, most customer satisfaction surveys are looooooong. For example, a hotel chain I stayed at recently sent me a 21 (!) page survey. Here’s how the transaction went:

  1. Get an email asking me to complete their survey
  2. Click on the link and intend to complete it
  3. Arrive at the page and see there are only five questions on it
  4. Heartened, fill the page out and click “Next”
  5. Get an alert that I’m now on page “2 of 21”
  6. Disgusted, immediately close the browser and return to my life

This type of abandonment is the typical response from people who are pre-disposed to provide feedback only to discover a long survey. Not only do you not get any data from an abandoned survey, the negative feelings toward your company for trying to subject the consumer to a painful feedback process certainly impacts your brand. Those “soft” perception hits are perhaps the biggest risk when you prove difficult to do business with, even on something as innocuous as an online survey.

My real-life example of a 21 pager is an extreme; in reality you want your survey to be as short as possible. If it has only three questions, they will almost certainly all get answered. Anything more than that, with each question, you are increasing the likelihood of the responder disengaging. So, what’s the right number of questions? It depends. That’s where research professionals can add a lot to the creation process. As a general rule of thumb, the more important your company, products and/or services are to the lives of the people you are asking for feedback, the longer the survey can be and still get complete information. But it’s an art, not a science.

Acknowledging that this is not an absolute, to give you a rule of thumb to work with, I will not write an online survey with more than 10 questions or fields. Under any circumstance. My goal is to figure out what the most important things we need to know are, and then optimize the survey to get that information as completely as possible. There are lots of reasons for that:

  1. I want people to see my company as easy to work with
  2. I know that getting lots of respondents flattens out issues of selection bias and getting only certain types of customers - who may have more limited patterns of perception and response
  3. I know that less data makes for easier analysis and understanding from stakeholders
  4. I know that less questions get better and more thoughtful answers than longer surveys, where people switch onto autopilot

Any one of these are reasons to prefer short surveys; all of them together make it a no-brainer. Under no circumstance are there more than 10 things that “have to” be responded to. There just aren’t. So why not make the survey simple?

To be clear, I am not claiming that online surveys should “never” have more than 10 questions. But it is an excellent and easy shorthand way to start producing more viable online surveys. The less questions you have, the more people will complete the survey. The more they will pay attention and not glaze over questions. The simpler the data set and contingencies are to draw conclusions from. And, if the data you get indicates you need to ask more questions, well it is certainly easy enough to create another short survey! The bottom line is getting valuable feedback that helps you improve some aspect of your business, for the benefit of the customer experience. Remember: your survey is part of the very customer experience ecosystem that you are striving to improve!

Topics: Design, user research, Blog