The point of running a market research survey is, of course, to help you understand the market for your product or service.
This understanding depends on a deep and balanced analysis of your survey data. Below, I explain the three major kinds of survey data analysis.
Toplines are the overall percentage answers to each of your survey questions. Simple as that. Your survey platform should be able to present those immediately after your survey is closed (or even while it’s still fielding). Be sure you’re filtering out bad responses you identified while cleaning your data. Most survey platforms should make filtering easy.
Toplines alone may be all you’re interested in seeing. But most likely, you’re going to notice a few things that raise evidence of certain people responding differently than others to your product (for example, you may find that 15% of your respondents have the same willingness to pay, but at a price-point five times higher than the other 85%). Think deeply about this possibility, then segment your data accordingly (see below).
Segmenting your data means slicing and dicing it in strategic ways. If you’d like to see how men and women answered your survey differently, you segment the data based on gender. If you’d like to see how rich and poor people answered differently, segment on the question where you asked about rich and poor.
Typically, your survey platform should make segmentations easy. Play around with your “Reporting” or “Analysis” tab to find out how respondents’ answers to individual questions (like gender) can be used to segment your data later on.
The possibilities are many when it comes to segmenting your data, but a few limitations exist. First, remember that when you segment your dataset into two groups, the sample sizes drop. For example, if your survey was fielded to 200 people—50% male, 50% female—your findings regarding only females or only males are based on a sample size of just 100 people. If you segment those groups even further (say, women under age 30), you’re dealing with an even smaller sample size, meaning your findings aren’t as robust.
But that doesn’t mean segmenting two, three, or even four times isn’t worthwhile. If you find that women under age 30, for example, seem to be highly interested in your product, you can go back and gather more responses from that particular segment to make sure you have enough responses from this group to make an informed, defensible decision.
This is a complex analysis that will probably require outside help. But the insights gleaned from this kind of analysis are often exactly what business owners have in mind when they envision what they’ll learn from a market research survey.
A cluster analysis divides your respondents into groups, depending on their combination of answers to your survey questions. Statisticians define this as, “grouping a set of respondents in such a way that respondents in the same group are more similar to each other than to those in other groups.”
This means you’ll be able to isolate groups of respondents who exhibit the most interest in your product concept, then find similarities between them—perhaps in where they shop or what other products they own—that help you target them later on. You’ll learn, in other words, the “type” of person most likely to buy your product.
Yes, this may sound like segmentations, which we discussed above. But a good cluster analysis will help you identify what segmentations to run, rather than you doing trial-and-error and cutting your dataset every which way in order to discover which cuts reveal meaningful information.
These are just three ways to analyze your data. I recommend you start here, but the possibilities are almost endless when it comes to what you can learn from a well-designed survey.
If you need help with your market research survey project, reach out to me! I’d love to make sure you get maximum value from your survey data.