How to conduct market segmentation research

How to conduct market segmentation research

There are two main ways to conduct a market segmentation study: a priori segmentation and post-hoc segmentation.

What is a priori market segmentation?

A priori translates from the Latin as ‘from the former’, and relates to an argument or line of reasoning that is deductive, based on logic rather than observation or experience. However, in the context of market segmentation, it means grouping people according to predetermined factors.

For example, FMCG brands may find that grouping customers according to how often they use their products is the most meaningful way of addressing their needs. Someone who drinks Coke every day and at every meal is likely to have an entirely different relationship with the brand to someone who has one can a week as a treat.

Demographic differences such as age can also be helpful a priori methods of segmenting a market. For example, marketing children’s clothes to parents with toddlers requires one approach; the same brand appealing directly to teenagers will need a very different marketing strategy. Similarly, vitamins and health supplements are often explicitly targeted at specific age groups or life stages.

How to conduct a priori segmentation research

Most segmentation research uses quantitative surveys, which are usually carried out online nowadays. For some types of a priori segmentation, you don’t need to conduct a separate survey or to use any complex analysis. If the segmentation variable can be asked directly of the respondent, such as age, gender or even frequency of use, you can create segments from a simple crosstab of that variable, and you can use existing research such as brand tracking studies or customer satisfaction studies to create your segments.

However, you may want to conduct additional analysis to get a more detailed picture. For example, if you want to divide your customers up according to how loyal they are to your brand, you could use one simple measure of loyalty, such as the NPS. However, you may decide that it is a bit more complicated than that. You can then conduct a survey designed specifically to understand loyalty and use multivariate analysis techniques to create your segments.

What is post-hoc market segmentation?

Post hoc means ‘after the event’. In the context of market segmentation research, this means analysing research data to derive segments, without having decided on them in advance.

As an aside, not to be confused with post-hoc fallacy. Sometimes the term ‘post hoc’ is used as shorthand for ‘post hoc, ergo propter hoc,’ a Latin phrase meaning ‘after this, therefore because of this.’ The phrase expresses the logical fallacy of assuming that one thing caused another merely because the first thing preceded the other.  For example, if the doorbell rings, and immediately afterwards, your microwave dinner pings, the fallacy would be to assume that the doorbell had caused your dinner to finish cooking. Not only is this ridiculous, but it also has nothing to do with market segmentation.

When to use post-hoc market segmentation

You would be more likely to use post-hoc market segmentation when grouping people according to their attitudes, to a particular product or category, their opinions about a particular issue or their motivations towards a particular behaviour. The resulting segments are typically more descriptive and richer than those created using a priori techniques.

For example, a supermarket chain may want to understand how to group customers in a more detailed and nuanced way, according to a variety of variables, not just how often they shop or how much they spend.

How to conduct post-hoc segmentation research

The first step is to design a questionnaire that will cover all the areas that will be useful in differentiating customer types, and in meeting your objective for the segmentation.

Exploratory qualitative research

We would always recommend starting with some qualitative research – focus groups or depth interviews – to guide questionnaire design. Using our supermarket example, we might suggest an initial qualitative stage, interviewing customers of different ages, family structures, incomes and genders. Having done so, we can be sure that the survey design hasn’t missed anything and uses language that will resonate with customers.

Analysis

Once the survey data is collected, we can conduct analysis to create your segments. We use techniques such as factor analysis and cluster analysis to determine how different variables relate to each other and how customers are grouped according to multiple dimensions.

The goal of the analysis is to create segments that contain people who are as similar as possible within each segment, and as different as possible to people in other segments, be it for a study of local markets or global market research. Segments also need to be useful – so they must be capable of being targeted in the real world, using demographic identifiers or other ways of finding segment members.

As such, segmentation is as much an art as it is a science. It is not unusual for the analysis to create several different options for how to divide the data. In such a case, you would look at how useful each option is, the relative size of the segments and the degree to which they all differ.

Number of segments

Most studies yield somewhere between four and eight segments, with five or six being ideal. Any less, and the segments are likely to be too big, with too many differences to make the exercise useful; any more and the segments might be too small and only slightly different from each other. Having more than eight segments, which all need separate marketing strategies, can also be unmanageable in a practical sense.

Sample size

It is critically important that the sample size for the survey is large enough to conduct the segmentation and to enable analysis within the segment. Best practice is to have segments of at least n= 200. If you have six segments, that is a minimum of 1,200 respondents – although you actually need more as segment sizes are likely to differ.

Naming the segments

Once you have derived your segments, it is important to name them so that you and your team can easily understand what they mean. Using the supermarket example, you might have a segment that always buys the own-brand budget items wherever possible and is very price sensitive. Rather than just calling them ‘segment one’ or describing them, if you call them ‘low-end budgeters’, it creates a mental picture. Similarly, you may have another segment that shops every day for a small amount of goods and likes to browse. These could be the ‘basket browsers’.

In summary

Market segmentation research ranges from simple analysis of existing data to more complex analysis requiring dedicated qualitative and survey research. Either way, a unique segmentation that aligns with your business objectives will help you to understand your customers and prospects, uncover more opportunities to serve your market and meet customers’ needs in a way that feels personal to each of them.

We are specialists in customer segmentation research and delivery in both B2B and consumer markets.