Description

Generalized discounts for your customer base can lead to loss of margin without even increasing their brand engagement.


Being able to segment your customers in elastic and inelastic clusters gives corporates the possibility to optimize their revenues giving discounts only to sensitive groups of clients.

Low uplift
Group
High uplift
Group
Control
Group
Different discount offered in order to test the impact on the volume purchased and revenue
Determining the optimal discount to offer to each group
A/B testing of the impact of the optimal discounts in the orders

Solution


Segmentation Model

Creation of a segmentation model that first analyze customers reaction to promotion and incentives and then cluster retailers in 2 categories, sensitive or non-sensitive to discounts.

Recommendation Model

Development of a recommendation model that suggests the best discount for each client segment, along the consumer journey

A/B testing

Implementation of A/B testing campaigns for testing the model hypothesis and continuously improve both customer segmentation and discount optimization.

Results

Increased savings by 40% avoiding promotion for inelastic customers

Increased customer engagement

Promotion campaigns sent only to discount sensitive customers, optimizing campaigns delivery workload


Solution


Segmentation Model

Creation of a segmentation model that first analyze customers reaction to promotion and incentives and then cluster retailers in 2 categories, sensitive or non-sensitive to discounts.


Recommendation Model

Development of a recommendation model that suggests the best discount for each client segment, along the consumer journey


A/B testing

Implementation of A/B testing campaigns for testing the model hypothesis and continuously improve both customer segmentation and discount optimization.


Results


Increased savings by 40% avoiding promotion for inelastic customers


Increased customer
engagement


Promotion campaigns sent only to discount sensitive customers, optimizing campaigns delivery workload