Case: Attribution modelling in Retail

Solving the complexity of the customer journey

Subjects:

  • Attribution modelling
  • Offer personalization
  • Campaign analytics

Do you face challenges with attribution modelling? Read on to find out how we increased sales and ROI for marketing expenses for an international supplier of office products and solutions.

To have better insight into the particular marketing activities that are most effective, we created an attribution model template that enables the revenue for each order to be tracked back to its original online channel. This helped our customer to fine-tune customer contact and to increase the number of conversions relative to the volume of communication.

Complex puzzle

Our customer wanted to improve its marketing efforts, because the return on investment for its campaigns was below expectations. The interaction with its customers and the marketing offers needed to be more tailored and effective. However, without insight into customer behaviour, manual campaign selections, and customer data in different systems, this was a complex puzzle. Therefore, Crystalloids helped to implement the SAS Marketing Automation solution, starting in the UK. The first step was to automate the gathering and de-duplication of customer data, including a complete purchase history spanning all on- and offline channels.

Campaign performance

With SAS, Crystalloids configured advanced analytics templates to predict churn, cross-sell opportunities and acquisition possibilities. The output of these models is automatically imported into a campaign management environment, to optimise and refine campaign selections. With advanced multi-channel campaigns, our customer can now target more companies, with more personalised offers. This applies for the retail- and the direct sales channel. Crystalloids also advised on customer segmentation models and on how to calculate customer lifetime value, in order to serve large business customers across Europe. To be able to monitor the performance of all campaigns, we developed a fully automated performance template that is updated daily.

Multi-touch attribution modelling

To further understand the customer journey and to be able to justify marketing expenditures, we developed a multi-touch attribution model that can estimate the real effect of every online and offline ‘touch’ for every individual customer. This predictive model calculates the weighted contribution of each touch point in different phases of the journey. Each conversion is then linked back to the specific customer’s journey to determine the total contribution and ROI for each touch point. Attribution modelling helps our customer to make better decisions about campaigns and about the particular channels to use. It answers questions like:

  • Are my campaigns worth the effort, if most of my conversions come from SEO and direct traffic?
  • Should I keep investing in online advertising, if it creates a lot of traffic but no direct conversion?
  • What keyword searches lead to purchases?

Media mix modelling

On a higher level, we developed media mix models for every European country. These models calculate the optimal distribution of the marketing budget above and below the line for each country. How much do we have to spend on TV / radio / advertising, digital marketing and print to reach our sales targets, based on the ROI for each channel? The media mix models are calculated on a vast amount of data, ranging from market price levels and promotions to economical and conjectural data. Using the multi-touch attribution model, the budget for digital marketing is divided over the different online channels. With automatic ROI reports, each country can monitor campaigns and make adjustments where necessary.