Market research : preparing correlation matrix from your qualitative data

One question we often get from prospects and customers regards qualitative market research : how should qualitative data be handled?

We discussed qualitative marketing techniques in several posts on this blog, but never discussed correlation matrix. Yet, this is an essential step to analyze qualitative data in a scientific way.

The scientific side of qualitative marketing has disappeared

Qualitative marketing is difficult, costly and takes time; three good reasons for managers to prefer cheaper techniques like surveys. Organizational pressure has slowly eliminated the most valuable elements from qualitative research. Essential analytical steps have been removed to propose low-cost to the market. As a result qualitative interviews have become the exception rather than the rule and were slowly replaced by focus groups. Worse yet, 99% of qualitative interviews are carried out without transcription, without coding. As a result, correlation matrixes can’t be realized.

Correlation matrix : a hierarchy of qualitative triggers

Based on the coding of the interview, a correlation matrix enables you to establish one the one hand a hierarchy of the most important triggers for the customers. On the other hand it enables you also to find out how those triggers work together.

Once you have prepared a coding guide, coded your interviews, you’re ready to analyze the occurrences of those codes. In the correlation matrix what interests us in particular are co-occurrences. In other words you need to find out the associations of some particular codes and how strong those associations are.

You’ll find below a typical example of such a correlation matrix (this example was produced using Maxqda).

What to do with a correlation matrix ?

A good correlation matrix will already give you an indication of the weights of each correlation. You can of course use those weights to draw conclusions.

A more powerful way to analyze your data is however to create a mind map. Start with the strongest correlation and gradually move towards the less strong ones. Draw the relationships between each variable and you’ll soon get a very visual and self-explaining map of your customers’ needs.