August 18, 2015

Home Shopping


 The Situation
A large home shopping fashion retailer introduces contact optimisation to pick the best catalogue mix for each individual customer

Executive Summary
This client with multiple million customers was able to obtain a double digit uplift in its returns from catalogue marketing by introducing a combination of insight tools, contact density tests, and contact optimisation software

The Client’s Business Goals
• Develop propensity models to predict both individual response and purchase value from each of the main brands and catalogue types
• Understand the impact of varying the overall contact density for individual customers throughout a season
• To move away from a near universal distribution of the expensive main catalogue at the start of a season, to the use of smaller more targeted catalogues

Our Solution
• We set up a joint analysis development team to build the propensity models, and transfer knowledge in model development to the client. The propensity models were built to replace the traditional reliance on RFM models as the main targeting tool.
• Tests on several smaller tranches of customers were set up to reduce contact density from business as usual levels to around 50 % of that amount. For each tranche of customers the impact of contact density levels in terms of overall demand in the season was measured. We then developed models to investigate how much contact density was a driver of changes in demand compared to other customer characteristics.
• Based on contact density test results, customer level budgets for catalogue contact density were introduced.
• Finally the propensity model scores and the contact density limits were introduced into our contact optimisation software and scenarios run, both for individual selection periods and an entire season.

Key benefits
• We were able to identify different groups of customers for whom ROI from catalogue despatch was maximised at different levels of contact density – broadly those customers with a high overall spend propensity had a much stronger appetite for greater levels of contact density.
• The gains from introducing the propensity models were equivalent to those from using contact optimisation software. In each case the uplift in terms of returns from the same level of catalogue budget was around 10%.