The Role of Judgement in Statistical Analysis
The wisest statisticians are those who readily accept that in business applications, managerial judgement provides a complementary role to some statistical techniques.
For example, there are some multivariate techniques that produce relatively definitive outcomes such as regression analysis or structural equation modelling where models are produced with identified drivers. Sure, these results need to be interpreted and explained to management, but the role that judgement plays is minor in comparison to its importance in techniques such as segmentation or cluster analysis.
These techniques attempt to identify groups within a market whose members have more in common with each other than with the rest of the market as a whole. If correctly identified, these sub-groups or segments within the market allow management to customise their product/service offerings to best match the needs of that segment.
Often the most statistically eloquent solution is not the most managerially relevant. In addition to the generally accepted rules of segmentation (measurable, accessible, substantial and actionable), the communal judgement of the analyst, the consultant and the client becomes the final arbiter of what is a suitable basis for segmentation – rather than what statistics might mislead us to believe is the 'best solution'.
Environmental factors, such as multiculturalism, place a heavier burden on communal judgement and often result in the need for culturally specific insights to be better able to hypothesise the most operational discriminating variable/s. Where this is achieved, the resulting segments are more likely to be both actionable and profitable (or sustainable).
Further, where the researcher is attuned to the multicultural aspects of the market and so better able to apply judgement, it is less likely that minority groups, with profit potential, will be overlooked.
Perhaps all too many analysts take too much comfort in believing the absolute truth lies in statistics. The reality is that judgement plays not only a crucial role but sometimes even a contradictory one.
For example, there are some multivariate techniques that produce relatively definitive outcomes such as regression analysis or structural equation modelling where models are produced with identified drivers. Sure, these results need to be interpreted and explained to management, but the role that judgement plays is minor in comparison to its importance in techniques such as segmentation or cluster analysis.
These techniques attempt to identify groups within a market whose members have more in common with each other than with the rest of the market as a whole. If correctly identified, these sub-groups or segments within the market allow management to customise their product/service offerings to best match the needs of that segment.
Often the most statistically eloquent solution is not the most managerially relevant. In addition to the generally accepted rules of segmentation (measurable, accessible, substantial and actionable), the communal judgement of the analyst, the consultant and the client becomes the final arbiter of what is a suitable basis for segmentation – rather than what statistics might mislead us to believe is the 'best solution'.
Environmental factors, such as multiculturalism, place a heavier burden on communal judgement and often result in the need for culturally specific insights to be better able to hypothesise the most operational discriminating variable/s. Where this is achieved, the resulting segments are more likely to be both actionable and profitable (or sustainable).
Further, where the researcher is attuned to the multicultural aspects of the market and so better able to apply judgement, it is less likely that minority groups, with profit potential, will be overlooked.
Perhaps all too many analysts take too much comfort in believing the absolute truth lies in statistics. The reality is that judgement plays not only a crucial role but sometimes even a contradictory one.
