Think you know segmentation?

Why Segmentation?

Segmentation is used in market research to facilitate better decisions, enhance profitability, and guide your sales and R&D.

  1. The more knowledge gained about a target market, the easier to influence the customer to distinguish a product, service or brand.
  2. Determining what drives customer behavior in a specific product of service category.
  3. Informs what is valued by a specific customer type or profile, therefore tailoring the advertising to what appeals to that group.
  4. Pinpoint unmet needs in profitable markets.
  5. Reveal comprehensive patterns of consumer behavior, to better characterize different groups (i.e., demographics)
  6. Identifying brand positioning within different needs segments.

Four Methods of Segmentation

Although there may not be commonly-discussed definitions, there are at least four very different kinds of quant segmentation, which researchers should be cognitive of when planning their project.

A priori segmentation is an approach that differentiates the market based on predetermined variables. This method utilizes supervised learning tools to develop statistical profiles. This data gives a forecast into key marketing variables, such as brand usage.

Post hoc segmentation applies characteristics or variables to differentiate consumers. This is achieved through information gathered from consumer surveys, customer data or other sources.

Key driver segmentation looks at one or more key variables and run analyses based on correlation, regression or other methods that give insight as to why one variable is superior or lesser, or drives up or goes down.

Brand segmentation analyses behavioral trends observed in consumer purchases, image and perceptions to distinguish clusters of brands.

Let’s look at these in more detail.

A priori Segmentation

In latin, a priori means “from the former.” In other words, a priori segmentation refers to knowledge naturally held by the researcher before the data was collected, as opposed to knowledge gained through testing or experiences.

This traditional method infers the use of earlier knowledge, hunches, assumptions, or organizational custom to classify markets based on predetermined factors, such as demographics, lifestyle practices, and behavioral norms. This approach asks the research to first determine the segmentation variables (such as age or income), and then customers are classified afterwards.

The benefit of a priori is that sophisticated statistical techniques are not required to segment respondents; instead, relying on simple cross tabulations and filtering. This is often sufficient if the researcher already feels confident about the organization’s previously defined segments, and are looking to fill in those segments with new response data describing the segments themselves.  

Example Case Use

In one example of a priori market segmentation, a client in the water heater industry ran a research study that included men and women of age 18 years and older. They were asked questions about their preferences with a new energy efficient water heater system, evaluating the concept. Respondents also evaluated the concept of the system based on likes / dislikes.

Per the previous client requirements, respondents were segmented by age and family size. These segments were predetermined, so there was no need to determine if these segments were indeed unique. The client then revealed that in a primary segment, only 21% liked the concept due to economic, environmental or other factors. Using a priori allowed the client to plug this insight into their existing marketing and advertising strategies (which are based on the segmentations that were already defined).

Post hoc segmentation

Post hoc means “after this”, and refers to the reasoning, discussion, or explanation that occurs after something has already happened. In research, a post hoc segmentation would define segments based on how the market view their world. This might be referred to as attitudinal, psychographic, or behavioral segmentation. There is no predetermined target group in post hoc segmentation (unlike a priori). The goal is to identify customer groupings with market significance that is leveraged in marketing strategies.

Post hoc segmentation classifies consumers into fairly homogeneous groups, using data from surveys, customer data, or other sources. After data collection, statistical techniques such as K-means, Hierarchical or TwoStep Cluster Analysis can be applied, which establish segments that harbor unique attitudes or behaviors from the others.

Example Case Use

One case regarding the use of this type of segmentation can be applied to purchasing fashion accessories. Your client is launching a new line of purses, and does not have any preconceived notions of whom to target. Therefore, the researchers gathers survey responses from women of different psychographics and demographic characteristics. From there, the researcher might look for logical groupings (aka segments) of consumers who are seeking the same color, willing to pay similar price points, or require purses of the same carrying capacity.

To determine the post hoc segmentation, the researcher might run a K-Means or Hierarchical Cluster analysis on the gathered data, to establish that women who want black, large purses are willing to pay $250+ for their accessory; while women who want slightly smaller grey or brown purses are only willing to pay up to $100.

Read how post hoc was applied in the restaurant industry.

Key Driver Analysis Segmentation

Key driver analysis (KDA) is a statistical technique that determines the relationship between potential drivers (independent variables) and customer behavior (the dependent variable). Basically, it seeks to determine the extent to which different drivers influence a specific consumer choice, such as interest, purchase intention, brand preference, or overall satisfaction.

Key driver segmentation applies multiple variable regression to assess driver performance in order to reveal correlations. These key drivers are variables that significantly affect the outcome in the market.

This method of segmentation can be used in quantitative research, and can be useful in determining what motivates a consumer’s behavior. Also, this helps to inform which service aspects are most significant to a business’ customers.

Example Case Use

  Women were polled about hair growth products in a recent study; specifically, they were focused on oil treatments. Respondents rated oil treatment brands they were familiar with, on a range of attributes, as well as future purchase interest. Driver segmentation was implemented, with purchase interest acting as the dependent variable, and attribute ratings acting as the independent variables. For example, one finding found that oil treatments that used a very specific ingredient were more likely to drive purchase interest across health conscious consumers.

Learn how a healthcare device manufacturer used software for key driver analysis.

Brand segmentation

The focal point of brand segmentation is what differentiates the brands. Quite similar to post hoc segmentation, this technique also utilizes customer information achieved from surveys, or actual buying patterns from customer records.

Consumer data is then used to segment brands within a product/service category that are most alike compared to other brands within the same category. This gives insight into the strengths and weaknesses of a brand in comparison to competitors. Also, to highlight the standout characteristics of a brand image that appeal to consumers most. Finally, it informs what options are available for expanding the product line or introducing entirely new products.

Example Case Use

Here, we will look at an example of this approach in action.

Men purchasing colognes were surveyed on a range of topics, including usage of long-lasting fragrances.  Brand image of the leading long-lasting fragrance brands was also asked. Hierarchical cluster analysis and correspondence analysis revealed two brand groups; one appeal segment linked to attributes of unique smell, attention drawn and suits me; the other was a functional segment attributed to longevity, high quality and good ingredients.

Women were polled about hair growth products in a recent study; specifically, they were focused on oil treatments. Respondents rated oil treatment brands they were familiar with, on a range of attributes, as well as future purchase interest. Driver segmentation was implemented, with purchase interest acting as the dependent variable, and attribute ratings acting as the independent variables. For example, one finding found that oil treatments that used a very specific ingredient were more likely to drive purchase interest across health conscious consumers.