Behavioral Science and Segmentation (Part 1 of 6)

Editor's Note: I was thrilled when Adam Warner approached us about writing this series on closing the gap between what people say they do and what they actually do, otherwise known as the say-do-gap. Each sequential article will look at a traditional research method (e.g., a brand tracker or a concept test) and show how researchers can overlay behavioral science to the method to get a deeper understanding of the why, as well as the what, uncover unconscious drivers and motivations, and close that say-do-gap. Bring on the learning!

---

Behavioral Science is often used to answer big complex questions; how do we define value? What role does sustainability play in our purchasing decisions? But to only look at behavioral science through this lens misses the true value of the approach.

Research participants don’t always tell us the truth. That’s not because they’re making an active or conscious decision to withhold, but more likely because we’re terrible at articulating our behavior and motivations. This creates a gap between reported behavior and actual behavior that can lead to poor outcomes for brands, products, and services.

However, there is an answer; behavioral science. In this series, I’ll be sharing how you can apply behavioral science techniques to traditional research methods and close the gap between claimed and actual.

This week we look at segmentation

Commercial research often begins with the question, who do we need to speak to? Whether it’s demographic, cultural or attitudinal, a segmentation must answer: who are my customers, how do I recognize them, and what do they do?

Segmentations uncover the structure of the whole market. They group potential buyers with all their diversity and complexity into a smaller number of categories, each of which shares enough similar characteristics to be described simply. You need to know your audience to develop the right products and brand strategy to reach them.

Clearly, a vital piece of work at the foundation of product development, marketing, and sales. But a segmentation is only as good as the data it is based on. And the data underlying most segmentations is ripe with half-truths. When someone asks us about our behaviors and attitudes we tend to flatter ourselves. We will usually claim more interesting and “moral” behavior with much more frequency than is the case.

An example of this can be found in reported behavior on recycling. In a Kantar report on sustainability during Cop 27, they found that while 60% of us claim we also recycle hard plastics, in reality, the figure is around 15%. Additionally, researchers are people too. When looking at segmentation data we tend to find recognizable personalities and mistake statistical patterns for real people.

We forget that a segment is a collection of different individuals with some shared attributes, and look for a single persona that represents them. Although the persona is just an example, we assume every customer looks just like it. Respondents mostly have good intentions.

They want to tell you the truth about how they use products and brands. But they also have powerful intuitive biases. They want you to like them – so they exaggerate how often they use your brand. They want to feel like a good person – so they report unrealistic levels of behaviours like recycling and healthy eating.

Segmentations can end up representing a simplistic frequency distribution on product usage, with no more insight than “some people buy from us more than others”. At the opposite pole, they can involve a complicated algorithm that is hard to understand and has little meaning to marketers who are trying to communicate to the segments.

The segments may become a required break in future research, but do they inspire action? Whichever way it goes, the segmentation will provide little value in guiding future research or marketing strategy.

What to do?

With the right research design, you can close the say-do gap and get a deeper understanding of your consumers. With a behavioral science approach, you can base your segmentation not just on explicit questions, but also incorporate implicit measures, such as IRT, to tap directly into emotions and unconscious responses. You can also use gamification to build a heuristic segmentation that measures consumers’ decision-making styles, guiding how marketers can influence each customer group towards making a purchase.

You can adjust your statistical analysis to adapt your data to include different kinds of measures – conscious, unconscious and behavioral – and create a richer understanding of consumer segments. This allows you to turn the segments into something actionable: useful instructions for the marketing team to whom they are speaking too.

Once all that is done, a narrative qualitative overlay will allow you to enrich those segments. It will allow you to better cut through the tendency to gravitate and create personalities familiar to us. Instead, you will get to understand the common stories shared by each segment, providing empathy and clear language to help talk to them better.

By injecting behavioral science into your segmentation work you maximize the value of that segmentation across all key stakeholders in your organization, minimize the risk of new product launches and marketing campaigns, and ultimately close the gap between claimed and actual.