The Research Advisory Service had the great opportunity to partner with the MRS in designing and delivering an advanced segmentation course to help data analysts in a leading broadcaster to get the most out of their data both in terms of how well it served their business objectives and in terms of how assiduously they interrogated it using standard statistical software packages.
What struck me when I reviewed contemporary writing on segmentation was how compelling it is for organisations to focus on data that they can easily collect – shopping behaviour, mobile phone usage, TV viewing behaviour, on-line purchase patterns etc. It’s compelling stuff because it is in real time, it is volumous, it mostly allows for some basic demographic intelligence and there is nothing like seeing old patterns to predict new patterns. So shaping products and services using this type of data feels good and probably (if it yields a profit) works.
There are holes in the data and inferring from behaviour is not without its challenges – who is actually watching channel 87, or did they or their spouse just leave it on by mistake; is the mother using the loyalty card to buy gin or just the 16 year old daughter’s boyfriend. But, again patterns and outliers of data can lead to educated guesses that may be good enough for predicting outcomes with a desired level of confidence. The danger of this type of data could be that it often focuses on the customer and that a certain myosis sets in on how far behaviours (often induced by those organisations in the first place) will predict desired step changes or simply provide a mechanism for product refinement.
This ‘behavioural’ data often requires nothing other than our actions as recorded in third party technology and this provides a real challenge for the sample survey. Are our life conditions, attitudes and stated world views enough of a tell to warrant sample survey research that seeks to uncover a more felt self to shape business objectives regarding product development and marketing in that it is here we are able to exploit prevailing consumer sentiments and go on to build competitive advantage? Probably yes. But it is quite hard to find the evidence – I looked.
My own feeling is that appropriately sampled qualitative and quantitative research should always have a place in shaping and validating segmentations that are based or need to be determined from real time behavioural data. And, I suspect that only in work that looks beyond what we do and into how and why we do it will inducing real consumer step change be achieved.
I leave you with a thought – not buying ‘fair trade’ coffee doesn’t mean you have no interest in products that support the disadvantaged grower in the developing world it may simply mean you don’t believe that the term ‘fair trade’ is anything but a ruse to get you to spend more money.