Analysis Using Transaction (Customer-level) Data

For disaggregate (customer-level) analysis, the tools and techniques used to study consumer panels are also applicable for retail analytics. The focus of analysis however shifts from brands to outlets.

The range of analysis includes:

  • Customer profile analysis
  • Loyalty and propensity
  • Assortment analysis
  • Overlap analysis
  • Outlet group analysis
  • Outlet repertoire analysis
  • Gain–loss analysis
  • Trial and repeat visit analysis for new outlets
  • Penetration and repeat rate
  • Sales forecasting (relevant for new outlets)

The case example, “Evaluation of opening of new petrol station”, which comes later in this chapter, illustrates the application of some of these analysis.

Note that data confined to the retailer’s own transactions, as is the case with loyalty panels, curtails the scope of the analysis. The analysis is then restricted in context to the retailer’s customers and their transactions at the retailer’s outlets.

For example, penetration would be redefined as proportion of loyalty card holders. The analysis of gain–loss between stores will not be able to capture switching between competing chains. Forecasts would be confined within the boundaries of the retail chain. And metrics such as loyalty and propensity cannot be computed. Both these measures require an assessment of customers’ transaction across the entire market.

In the FMCG sector, where they are prevalent, consumer panels provide a holistic view of the market and work well for most of the above analysis. They are however expensive to set up and maintain, and sample sizes would tend to be relatively small compared to sales transaction data or loyalty panel data.

Previous     Next