Segmentation is the ultimate form
of data reduction
Successful segmentations are the result of strong planning, good communication and the use of best-in-class statistical models.
We place people into groups based on their data and use their group membership to make predictions on how they are likely, on average, to respond to a set of underlying questions with a particular pattern.
These models are often used to understand the characteristics of groups in the population with very different needs, attitudes and behaviours.
We have experience of running many hundreds of segmentation studies and are power users of Latent Class Analysis, considered one of the most robust and flexible methods for identifying distinct groups with different patterns.
It is also the best method for mitigating “response style” effects to ensure the segmentation focuses on genuine differences rather than differences in rating scale use. It uses powerful underlying models which enable delivery of either a classic segmentation model, where we allocate cases directly into a “bucket”, or a segmentation model built up, layer by layer, a dimension at a time.
The success of a segmentation depends on getting the right balance of content in the survey, which is measured in a way that will work well for this type of model. Of course there are many other practical considerations such as the size of segments, the number of segments, and whether the segments are actionable for the client. We can advise clients on both the soft and hard issues, tapping into our experiences of handling many 100s of successful projects over the years.
Please contact us for further information.
Case Studies
-
Fashion Retail Segmentation
A leading men’s retailer wanted to understand the various groups that made up male shoppers…
-
RSA – The Power of Small
We analysed a segmentation of Micro-businesses to determine different tribes of businesses. The six tribes…
-
Stakeholder Segmentations
Examples of Stakeholder Segmentations: MP’s NGO Business Analysts and Journalists for a leading Media group…
-
Medical Treatment Segmentations
Our experience covers Segmentations among medical specialists in hospitals and research centres, focusing on preferences…
-
Political Segmentations
We analysed a series of segmentations used to type voters by their attitudes, behaviours and…
-
Working Practices
We have used Segmentation in the past among Dialysis Nurses to understand differences in their…
Key Factors
- Good communication and planning essential to success
- Must get the content balance right / clearly defined
- Need robust statistical tools and a well thought out approach
- Need a pragmatic approach to number of segments, segment size etc.
- Control for unwanted patterns e.g. response style
