Survey Design Tips

I received at least 10 emails from classmates at UFV asking to fill out surveys as part of their Bus 320 – Business Research Methods class. I think it’s great that marketing students are using tools like SurveyMonkey to do research, it’s all about practical learning, and these will be practical skills to have in the Marketing sphere. For now. Since this Fall 2009 I worked as a marketing research assistant for a financial institution, here are a few tips I picked up to share with you.

1. Sampling

It is critical that professors emphasize the importance of selecting a sample for the survey. While it depends on what kinds of surveys the students choose to do, it’s good not to pick your classmates. All of the survey invitations I received are from my classmates. There is nothing inherently wrong with this, however, there are many factors that can affect the results of the surveys. That’s why it needs to be as random and representative as possible. And picking your classmates isn’t as random as sending the invitations out to all students, students that you don’t know. Here is the difference in the real world. A financial institution had a marketing class at the local college conducting a market research survey. The financial institution intended to donate $5,000 dollars for the project and when the Marketing VP found out that the surveys were done by families and friends, the amount dropped to $500!

One may argue that it’s a lot harder to get emails from people you don’t know, but that’s part of Marketing. That is, to get people’s engagement and participation. Then the professors could also argue that this exercise is only to familiarize the students with using and executing the tools hence it doesn’t need to be done properly. But when done right, these data are valuable information. Surveys on cafeteria ratings, students’ employment status, feedback on the parking frenzy can all be used to improve students’ experiences on campus and help direct where the University wants to go. What’s a better way than asking students what they want to give them what they want?

2. Design

  • Before you ask a question, decide exactly how that question is going to help you in your research. For example, if you are surveying about the quality of the food in the cafeteria, how does a question like “what’s your year of study” help? It may be more revealing to ask for income level, or how often they eat at the cafeteria.
  • Before you ask a question, decide exactly what you are going to do with the answers. For example, if you ask “rate your experience eating at the cafeteria”, are you going to bring that data to the University and say “look, the salmon is of bad quality, there isn’t enough variety. Give students more options”? If you are not going to act on it, don’t ask.
  • Demographic questions should be placed at the end of the survey, when survey respondents are more comfortable with the process. If you ask right of the bat, it’s more likely that they will just shut off the survey
  • Keep it short and simple. Don’t explain – in my opinion, if you have a question that spans more than 3 rows, that’s bad survey design. The more direct your answers, the better the respond rate. “Was the coffee hot enough?” is better than “how was the coffee?”

3. Survey Tools

SurveyMonkey is good. It does everything a free survey tool should do. But you gotta admit its templates aren’t the best looking surveys out there. And it really turns me off, especially when the survey is long (I’m just too damn picky!) If you are serious bout doing it right, then you gotta do what I listed above – select the right sample, ask the right questions, and give some thoughts to the design. Gotta have these

  • Progress bar
  • Big fonts
  • Lots of white spaces
  • Images (that fit with the context of the questions)
  • A theme that doesn’t remind of a DOS program (which I think some of SurveyMonkey themes look like)

Again, SurveyMonkey is perfectly good if your prof just wants you to get a hang of designing a bare-bone survey. But if I were your prof, I’d make you sign up for SurveyGizmo, everyone of you. SurveyGizmo gives students their Enterprise package for free (commercially run for $159/month) and for good reason. It’s professionally done, corporate-level with robust and deep tools like logics and branching, A/B Test (it’s good to know what this is too cuz you’ve been tested a lot of times already – Google itself is one of them). And plainly, their templates are just better.

All of you have to do is to sign up with your student email, and follow them on Twitter (which is another thing you gotta sign up for, now!) and voila, the Enterprise package is for you for free. Why not???!!!

Leave a Reply

Your email address will not be published. Required fields are marked *