The starting point was understanding what data was held and then deciding what data we would use. This required us to determine the data variables that we needed extracting such as categories shopped, brands bought, order value and frequency. We then sliced and diced the data to analyse customer transactional behaviour to decide on the scope of the data profiling activity.
We eventually decided that there were 25 customer profiles that would gives us a clear picture of who All Beauty’s customers are and how they differ by spend patterns, product categories, location and brands.