Experimental Design to Unlock Promotional Potential

Listen to article

A multi-national brand was looking to quickly assess effects of new delivery promotions given changes in the business around COVID-19

To understand the true incrementality of different offers, they established an advanced test and control experiment within Analytic Partners GPS Enterprise Agile Learning solution. Insights into offer optimization led to a per-weekend incremental lift in sales of £625k.


Following the industry disruption instigated by the pandemic, the brand needed to adapt strategies and experiment around the growing sales channel of home delivery. Together with Analytic Partners, they sought to understand how to optimize offers and answer three key questions:

  • Are delivery promotions incremental or are we just rewarding loyal customers?
  • Which offers perform best, weekend or weekday promotions?
  • Which free products provide the biggest uplift to sales?


The client leveraged Analytic Partners Agile Learning, a solution within GPS Enterprise that utilizes proprietary algorithms integrated with intelligence from Analytic Partners’ ROI Genome and commercial mix models. The solution guided the design of a series of multi-cell experiments that went beyond traditional test/control comparisons to capture and quantify complexities, balance multiple attributes and better deal with bias and contamination through a proprietary Contamination Index. For example, Agile Learning was able to quantify and apply the correct adjustments for differences in regional lockdowns that occurred across the UK.


Implementing Agile Learning provided a single version of the truth regarding the incremental lift of offers to the client who was now able to understand the true value of promotions and optimize accordingly. The experiment confirmed that delivery promotions are incremental executions, providing an incremental uplift of 3.4% onto the participating stores delivery sales. The experiment also highlighted the importance of timing, proving that weekend offers achieved an incremental uplift 4.8 times higher versus the weekday offers. The client was also able to isolate the impacts from various products that were offered free for the promotion.

The experiment revealed that a single national campaign would have
an impact of £625k per weekend on incremental sales.