When it comes to creating a new survey, where do you like to start? Do you start with a laundry list of questions and then filter from there to the end result? Do you start with a potential list of respondents, and then write the survey questions based on the respondents? Now…where should you start?
Start with purpose
I’ve been in many meetings where the survey project started with the end result (we want a bunch of slides that show us pie charts and slices of the data based on these particular demographics), the audience (how about we start with our customers, then open it up from there), or even the laundry list of questions that the customer wanted to have answered from surveys (we’d like to get a net promoter score to start a feedback loop, but we also want to get satisfaction ratings for customer service, the overall customer experience – how about we ask what color of the product they’ve liked best?). I’ve also sat in meetings where the approach was simply, “We don’t really know what we want to ask, we just want to do a survey.”
The best place to start when approaching any kind of survey project is with the purpose of the survey itself. Are you (or your client) wanting to measure how your brand rates against other brands? Are you wanting to get a baseline customer satisfaction score? How about trying to decide what features are most important to your potential customers?
It’s all right to be thinking about your respondent audience, your list of questions, and even what you hope the responses will show. Just be certain you come back to the actual intent of your surveys, or, as I like to put it, the one burning question you want to have answered from your survey.
Follow with relevant questions
To me, creating the questions for a survey once you’ve defined your purpose can still be difficult to do because there is always more information that you want to gather from your respondent audience. It doesn’t matter how well a research report has been written, I find myself always walking away with more questions that I wish had been asked in the survey, or hoping there is going to be a follow-up survey to dive deeper into a particular aspect of the report that I found compelling, but just want more data.
The trick is not to follow that train of thought in the early stages of writing the survey.
How can you avoid that pitfall? Go back to your original purpose for the survey. This can be difficult to do. As much as I love to teach that you come up with that primary purpose and stick to it, I know it’s difficult. For example, you could state that your one burning question is: “Do our customers like us?” But there are so many questions you could ask, from asking about the product offering (is it enough, is it too much, is it too little, is it too difficult to navigate, is it too…) to the entire customer experience (were you greeted when you entered the store, were you asked if you needed help, could you find someone to answer your questions when you needed someone), answering the seemingly simple question, “Do our customers like us?” can end up being a 45-minute survey experience!
The remedy? Identify things like key performance indicators, and possible narrow down the original question you came up with in step one. For the example, “Do our customers like us?” you could easily narrow that down to something like, “What do our customers like most about us?” or even, “How do our customers like us compared to our competition?”
Personalize the experience
Once you have decided on the survey questions, it’s time for you to plug it in, add any needed logic, and talk about personalization. If all you have is the first name of a potential respondent, it still is worth your while to include that in the survey invitation at a minimum. Everyone expects personalization in any kind of customer interaction, from a newsletter addressed to them to a survey invitation with their name on it.
Top it off with customization
Customizing your survey should be the very last thing you do. Make sure that everything else is working before you start to customize the survey experience. Customizing should be the icing on the cake, rather than a primary focus. Content first, then you can spend time making it look pretty. This reduces the risk of having a really nice-looking, but poorly-functioning survey. It also reduces the risk of finding that the error in your surveys was from the programming, and not the CSS formatting you’d just applied. This same rule applies to many other types of content but is especially true for a survey. In my own opinion, respondents are far more interested in a survey that is straightforward and doesn’t feel like it’s wasting their time than they are in a survey that looks cool but doesn’t make sense or is terribly long and boring.