Promotion Effectiveness

Key Concept

Quantify sales uplift compared to a predicted baseline without promotion

Subtract discount to get net impact

Reduce discounts on promotions with negative net impact (discount > lift)

Use model to predict uplift of promotions with different mechanics

For example, BOGO vs 50% discount on Segment ABC

Use the prediction model to explore potential mechanics for new promotions

Narrow down promotion incentives to a small range for A/B Testing

Recommend optimized incentive range to maximize sales or profits

Depending on data, mechanics can be optimized for every level of business

Reduce overall budget with improved promotion outcomes

Allow more members to receive the best promotions

Stronger decision-making supported by statistics

Apply inferential statistics to improve test reliability and minimize experiment size

Calculate statistical significance levels and adjust control groups to sample size to a minimum

In this complex world,

we need to collaborate to deliver outstanding work.