Image by Sergey Zolkin

Understanding the impact of online media using a regression model approach

Challenge

Robustly evaluate test and learn e-commerce investment focused over a three month period.  

Project highlight

The output of the model allows our client to lean on accurate and independent online performance data, allowing the team to make justified and data driven online media decisions within future commerce plans.

Approach

Using our bespoke regression model approach, we are able to understand the relationship of variables (e.g. seasonality, promotions etc) to sales – which means we can robustly isolate the impact of sales caused by media investment. We used this model to analyse the impact of test and learn media run on Tesco online across 4 bursts over a 3 month period.
 
The final output gave a thorough look into modelled vs. actual performance reporting on three key metrics, % uplift, £ uplift and ROI. Each burst was broken down by a FSKU and brand level allowing us key insight into the performance trends, leading onto learnings and recommendations for the team to take to plans in the future.

Project Statistics

4x
Bursts of online media robustly evaluated


£1.12
Total brand ROI across all 4 bursts of media


+£90,600
Total brand £ uplift across all four bursts of media