Modelling the drivers of sales uplift for a large alcoholic drinks client
A key ingredient in delivering business value from behavioural data is joining datasets that are regularly used, but in isolation from each other. The reason for this disconnect is normally by accident – rather than design – and connecting them is often cheaper than acquiring new data.
The challenge
On a recent project engagement, our client wanted us to connect two massive databases in a fast growing market – one on daily sales and another on daily shelf placement data – to objectively:
- Analyse whether outlets which complied with product assortment targets performed better than outlets which followed another strategy
- Identify where the targets were in the ‘goldilocks zone’, too ambitious or too easy to meet – cut by different outlet groupings
- Estimate the ROI of simplifying the product assortment targets
The solution
Before starting any data crunching, we organised dedicated workshop sessions with users of the analysis to understand what was ‘must have’ from the analysis vs ‘nice to have’.
We then carefully extracted and combined daily data for over 25,000 outlets and 3 million individual sales invoice/ shelf placement records, working closely with multiple product owners and local market experts.
The result
Using data analysis and sophisticated econometric modelling, we were able to give clear guidance on where targets were working and where they needed to be reworked for the next financial year.
A key visual output was mapping all 50+ outlet segments into a single ‘current vs expected’ performance framework, with key actions to take depending on the quadrant the segment was located.
If you have any data analytics questions or challenges please get in touch with:
