What is Consumer Segmentation?
Consumer segmentation is a process of organizing consumers into distinct groups based on needs, personal values, product preferences, financial value, and/or demographics such that those consumers in the same segment are similar to each other but distinct from consumers in other segments.
Segmentation research can fail for reasons other than bad data and/or poor grouping math. It can fail because you did not have a specific purpose in mind, assuming that segmentation approaches are fungible. They are not.
Let’s take two polar opposite business needs where the segmentation that succeeds for one purpose will utterly fail for the other.
Segment people for category expansion
Some marketers, especially those who are category leaders, are focused on category expansion. One way of expanding a category is to link a need or set of values to it that are not yet well associated as a usage case. A great example is how Church and Dwight encouraged a broader range of uses for baking soda. Putting a box of Arm and Hammer in the refrigerator is a classic success story of category expansion.
Which consumers would most care about that? Homemakers who are focused on keeping a clean home or those who work late and eat out a lot? You can imagine a segmentation study that would reveal a consumer segment of interest and what messages might be motivating to them regardless of whether they currently buy the product category.
Segmentation for ad responsiveness
I have done an extensive amount of research on this subject. My approach has been to find segments that are mathematically engineered to repeatably produce superior ROAS (Return on ad spending), defined as the incremental absolute quantity of sales caused by advertising, divided by incremental advertising spending. Typically, a ROAS of $2 or greater is considered to be good. One universally observed pattern is that non-category buyers will produce considerably lower ROAS than category buyers. Personally, I have never seen a counter-example. So, in an age of addressable marketing, when trying to drive ROAS, direct ad impressions away from non-category buyers.
The implication is that a segmentation study for the purpose of category expansion is useless and even damaging for ad responsiveness, and vice versa.
If your goal is to segment for ad responsiveness, there are a few rules that have mathematical properties that are as close to a guarantee as you will ever get in marketing.
- Target consumers who have a 20-80% probability of choosing your brand given a category purchase. In the work I did with the MMA, Neustar, and Numerator, we called this segment the Movable Middle. The math behind it is from calculus…take the first derivative of the logit function and you will find that change is maximized at p = .5. I have conducted 5 studies since we developed the theory and in 5 out of 5 cases, the Movable Middle returned 2-13 times the ROAS of those not in the Movable Middle.
- Target consumers who are close to an upcoming purchase. This is the principle of “recency”. In the Persuadables white paper research I did with Viant (a leading AdTech firm) and NCS, we found that delivering ad impressions to consumers who are probabilistically close to an upcoming purchase doubled ROAS.
- Target your own heavy and medium buyers. This principle came from the Persuadables research, but the Movable Middle research explained why we got this result; most of your customers (even heavy buyers) are not that loyal to you so you they are mostly in the Movable Middle and you are fighting for their purchases. This finding that heavier buyers are still in the Movable Middle is more than empirical. It comes simulations based on the Beta probability distribution that describes the percent of category buyers who have a given probability of buying your brand.
There are different segmentation approaches for other marketing goals, too. For example, shopper marketing teams will want to segment consumers based on their shopping styles, how they navigate stores, and trip type (e.g. stock up vs. fill in) and turn these insights into shopper and category management programs.
My advice: be careful to match your segmentation approach to the need and you will have a winning formula.