OPINION11 December 2014

Five things you should never do in a segmentation…


Anne Coveney identifies the five worst crimes committed in the conception, implementation and use of segmentation studies.


Once, the primary function of a segmentation study was to inform communications development – but today these studies are far more likely to guide the development of products and propositions.  With segmentations increasingly relied upon to guide bigger and more strategic decisions, it’s more critical than ever that we get them right.  

In spite of this, however, poorly conceived and badly implemented segmentations are rife.  At best, these studies are consigned to a drawer. At worst, they’re responsible for a raft of misguided decisions. Ultimately, the client company has no option but to commission another segmentation study, often within the calendar year, always with another agency. 

1. Allowing tunnel vision

It’s counter-intuitive to think behavioural or attitudinal or needs-based… Good segmentations use every piece of the puzzle. If you fail to cover every aspect of the market within your survey, you’ll never segment accurately or with confidence.

2. Paying lip service to qual

Qualitative research is essential for identifying what’s genuinely important to customers; the alternative is a segmentation based on internal beliefs, hypotheses and assumptions.  Even qual plus quant isn’t enough; integrated research teams make or break the study. Quant execs need to hear first-hand what consumers are saying while qual execs need to ensure that the questionnaire is sufficiently nuanced.

3. Not living in the real world

It may be the greatest, most academically superior piece of research ever done, but if the marketing team can’t use your segmentation then it’s of no value. Segmentations should be an iterative and collaborative process, using the client’s knowledge and input at every phase of the process.

4. Naming your segments

Names may help segments to ‘come alive’ but they are highly likely to be counter-productive.  Segments are rarely black and white, so unless the name is a stroke of genius (which comes along rarely) then it will immediately imbue the segment with characteristics that don’t exist – and overlook many that do.

5. Aiming too high (and encouraging the client to do likewise)

Good segmentations are incredibly powerful but they will never be a panacea, able to answer every marketing question that arises. Managing the client’s expectations is critical. The good news is, you can underpin the segmentation’s success by planning (and budgeting) to embed it within the business, ensuring it lives on and becomes an indispensable tool.

Segmentations are expensive and (should be) infrequent; let’s get it right first time.

Anne Coveney is head of quantitative research at Optimisa Research



9 years ago

Great advice Anne

Like Report

9 years ago

And if I may add one more item, you cannot assume that each person fits perfectly into one segment. People will often fit a bit better into one segment but also fit somewhat in several other segments. People may also shift segments over time. So, think about segmentation as a philosophy to guide you in your thinking, not a carved in stone text book.

Like Report

9 years ago

Very useful, grounded in experience suggestions. I commend the author.

Like Report

9 years ago

With respect to point 1--It is not just attitudes vs. behavior or even both attitudes AND behavior. A good segmentation also provides insights into how habitual the behavior is; to what extent there is dissonance in the attitudes, and what types of situations/occasions are apt to be disruptive enough to produce either behavioral or attitudinal change.

Like Report