Data analysis, simple and complex: a primer

Market research and data analysis go hand-in-hand. After all, what good is collecting data without knowing how to use the data you’ve collected? However, knowing just what kinds of data analysis you should be conducting to achieve the decisions you need to make can be tricky. Let’s look at some data analysis tools and how they can be used to help you make better business decisions.

Simple analysis

Let’s start with the simple analysis: response frequency charts. This is the most common analytical method used, in part because it’s quick, it’s easy, and it’s often all that is needed to get some top-level data to support or refute actions being proposed. For example, a response frequency chart could be used for determining how often a particular product is being used, or whether an audience likes a particular color design over another.

For these questions, you don’t need super in-depth analytics to give you answers to help you make decisions. Instead, you just need a quick count to show what preferences people had when presented with a set of options. And using the built-in infographics tool with QuestionPro, you can easily turn those frequency charts into something that is more visually appealing for your audience.

Another type of simple analysis is looking at text responses. The simplest (and, at times, funnest) method for doing text analytics is to look at a word cloud. The word cloud shows frequencies of words used in text responses. You could think of this as a response frequency chart (like those bar or pie charts), only for words. In QuestionPro, you can tweak the word clouds to remove certain words (such as “the”) to refine the view of what words were used most frequently.

Advanced analyses: GAP and TURF

Now, let’s look at some of the advanced analytics available and what decisions they can help you make.

A GAP analysis can sound self-explanatory, but it’s really quite handy. Using a side-by-side matrix measuring two attributes for a list of items (such as importance and satisfaction of a few different key metrics for your organization), you can then look at a GAP analysis to determine where the, well, gaps are between what is important to your customers and what their satisfaction is for those items they deem most important.

Perhaps your website has a fantastic user interface that your clients are extremely satisfied with, but it turns out the user interface isn’t as important as the way your products are displayed on the website, which your clients say could use some help. Using a gap analysis, you can easily see these types of discrepancies on a simple quadrant.

A Total Unduplicated Reach and Frequency (TURF) analysis is not-so-self-explanatory, but is extremely valuable for questions such as, “If we add a product to this line-up, how much market share might we gain?” or “If we only have X budget, how could we best use that budget to improve our customers’ experience?” Using a simple matrix question asking satisfaction on a list of metrics, you might see that your respondents are unsatisfied with a few items. Using the basic frequency chart, you might decide to tackle all of those items to increase customer satisfaction.

However, using a TURF analysis, you can actually zero in on which of those items (or what combination of items) would actually reach the most customers and have the highest likelihood of increasing customer satisfaction. Then, you could enter your budget, assign how much you’d be willing or able to spend on each item, and do some price modeling to see where your budget would be best spent.

Already, you’re making better business decisions with better data analysis!

More advanced analysis: conjoint analyses

Are you ready for the big-time? Do you want to know how your customers value different attributes about your products so you can determine what to focus on for the next product launch? Enter conjoint analysis. Using conjoint, you can determine what is driving your customers’ purchasing behaviors. Now, we aren’t just talking about how often respondents would select choice A over choice B; we’re looking at what makes them choose choice A over choice B.

Two ways to do conjoint analysis are to create a maximum differential set of questions and to create a discrete pairing set of questions. Using MaxDiff, respondents are given sets of attributes from which to select one that matches one end of a measurement scale and then select one that matches with the opposite end of the scale (such as “love it” and “hate it” or opposite ends of agreement).

This type of question has been found to avoid scale bias, since there is no middle-ground, only extreme opposites. Using discrete conjoint, you can set up attributes and groups of attributes that are combined randomly to present two options to the respondent. For example, you might set up three products, each with three levels of three attributes (price, model, and size). The respondent then sees a certain set of pairings (product A with attributes A1, B3, and C2; product C with attributes A2, B2, and C3; etc.) from which to select the one they would most likely purchase if presented with those options in a store. From these two types of questions, you can determine what attributes your customers are valuing.

Learn more about conjoint: May 28, 11AM Pacific

Later this month, we’ll be co-hosting a special webinar discussing conjoint analysis. We’ll run through more in-depth on what it is and when it can be used, hear from someone who has used conjoint analysis and how it was used to make business decisions, then show you how to go about creating a survey to conduct this analysis and run the analysis in the reporting tools. We’re excited to get to show you this new tool and how it’s used!