image

4 Data Scientists

image

Split between 2 criteria

image

>75% accuracy

Client

Rais

Purpose

Rais offer a CRM solution as a SaaS for small and medium sized retailers who want to understand their customers better to drive customer acquisition and retention. Small and medium sized retailers often don’t have the resources and expertise to conduct effective targeted marketing campaigns. The S2DS project objective was to create a visualisation tool to offer more actionable insights for marketers for a chosen customer segment.

Approach

A visualisation tool was developed to segment the user base according to two criteria: the churn probability of each customer, obtained using a Random Forests model, and the expected lifetime value of each customer, computed with a Buy'Til You Die model. The churned/non-churned status of each customer can be predicted with 75% accuracy using the available features and without using any information about the time since the last purchase. . The model thus gives the ability to assign a churn probability to each customer who has not churned yet. This predictive analysis gives the marketer information on which customer risks becoming dormant. To distinguish customers who might spend little in the future from large spenders, a Buy'Til You Die model (based on the beta/geometric-NBD model) was applied on sales data to give an estimate of the expected future expenditure of a customer.

We partnered with S2DS to build a new analytical model for our product. We were very impressed with our team; their practicality and commercial thinking. S2DS is a great programme, well run and beneficial not only for the data scientists but for the participating businesses as well

Agata Boczkowska, COO, Rais
The Outcome

The S2DS project team developed a visualisation tool which allows marketers to segment the buyers and identify customers of high value and those who risk becoming dormant. The tool classifies customers according to their probability of churning and their future value to the retailer. Marketers are better able to target customers as they have actionable information on their spending behaviour. The visualisation tool is written in JavaScript using the visualisation library D3 giving Rais the ability to incorporate the tool into their web based CRM platform in minimum time and cost. Rais’ competitors typically rely on descriptive analysis instead of predictive analysis, and this gives Rais an edge in the highly competitive customer analytics industry.

Connect with the Data Science Community

For Data Scientists & Businesses

TOP