Repertoire analysis

Using repertoire analysis to estimate purchase probability within purchasing repertoires

What is it?

•When consumers have a range of choices to make, it is normal for them to select a subset of the available choices from which to buy. The choices may be brands/ outlets. We refer to this as their repertoire set

•When the purchases are relatively frequent and it is easy to select different brands or outlets then repertoire analysis can be useful

•Each consumer will have a different repertoire set based on personal preferences, and it can be made of a single option, a dominant option and a range of alternatives, or a range of options with no stated preference

•Most analysis treats all customers the same irrespective of their repertoire set, which may overstate the penetration of brands. We weight consumer’s affinity to each brand based on their personal repertoire set

•From this we can estimate the probability of what brand the consumer will buy next in that category, and profile the consumer into affinity segments


•The reasons for brand selection may be quite complex and include behavioural (habit), psychological (brand commitment) and external factors (availability, positioning)

•We are of the belief that all these factors will be discounted into recent behaviour and the same factors will persist into the near future. Any factor which changes the purchase pattern will cause that persons repertoire set to change. The model is thus dynamic

•We can analyse a category and:
• estimate the probability of purchase for each brand
• split the category market into 4 probability segments for each brand
• profile the consumer in each segment
• develop brand sets for different demographic/geographic profiles

How does it work?

•The analysis requires a set of questions which ascertain a consumers repertoire and determines the purchase interaction the consumer has with each brand. The questions relate to preferred brand, brands recently bought and those they would consider in the future

•The probability of any one brand in the set being purchased next is then calculated. The total probability of all the brands is 100%

•The question set can be added on to any survey that you are running

•The analysis is derived at the respondent level, which allows us to work with any sample size

•Respondent repertoire and the consequent probability of purchase is not static and can change rapidly. The change will effect in the altered purchasing behaviour. It is thus beneficial to track repertoire sets

For more information please contact Duncan Brett: or 082 3333 444/+ 27 21 671 8653

Leave a Reply