MandM - Using Customer Lifetime Value Scores to Understand Inherent Future Value and Deliver Persona

Team Members:

• Ben Powis, Head of Data Science
• Joel Lenden, Data Scientist

Country: United Kingdom

Organization: MandM

MandM is an online discount fashion retailer. It is one of the largest online fashion retailers in the UK. They sell highly discounted products from many of the world's biggest brands, direct to customers throughout Europe.

Awards Categories:

• Best Acceleration Use Case
• Best ROI Story

Business Challenge:

MandM is a British fashion retailer, that exclusively sells products online at discounted prices. We have many satisfied customers throughout the UK and Europe, and we spend a lot of time understanding their individual preferences so that we can ensure our marketing is as relevant to them as possible.

As a business, we wanted to understand our customers' potential in terms of future value. This would help us target marketing activity more accurately at cohorts who show strong intent to engage with the brand.

Inversely, analyzing customers with low predicted future values helps us understand the traits that make up different customer profiles and understand which behaviors we would like to encourage more of in the future. This would give the business clear, data-driven engagement KPIs.

Business Solution:

We started our journey into LTV by developing our own custom Python model, which we set up to run in Dataiku. Shortly after we got this up and running, we noticed that the Business Solutions team at Dataiku had released their own starter project for Customer Lifetime Value. We jumped in and installed this in minutes.

Our Data Scientists who developed our initial model picked through the Business Solutions project step-by-step to understand how it worked, running some of the same datasets we had used in our initial model through the package and reviewing the results.

We were impressed with many elements of the Business Solutions project and, having access to all the code in Dataiku, decided to create a hybrid model, combining the best practices and Dataiku-specific implementations of the Business Solutions project with some of our own specific changes that fit MandM as a business.

The resulting project was far stronger, having been through this process, and was immediately ready to package up and deploy to our automation node, where it now scores millions of MandM customers every single day.

Business Areas Enhanced:

• Analytics
• Marketing/Sales/Customer Relationship Management
• Product & Service Development

Use Case Stage: In Production

Value Generated:

Once our model was deployed and scoring many customers each day, we were able to connect this information into teams across the business, driving actions and increased understanding:

• Insight and reporting: Our first steps were to analyze our customer lifetime value results and share these with key business teams, explaining how our model worked and what insights it identified. We were able to demonstrate how high and low CLTV customers expressed different behaviors and discuss how we could personalize experiences based on these.
• Customer segmentation: Once we had identified key groups of customers, we could segment them and share them with business teams so that we could personalize experiences for them. For example, a high lifetime value customer may receive more on-site content related to their favorite brands.
• Measurement: Having all the above in place also allowed us to use CLTV as a business-wide KPI for measuring success. With our model running every day, we were able to understand the inherent future value in our customer base and trend any changes in this as we personalized experiences over time. We were able to answer questions such as: “Does X brand increase CLTV if purchased?” or “If shoppers interact with X features of our website, how does it impact their onward value?”.

Customer lifetime value scores are now widely used across MandM, delivering added value daily.

Value Brought by Dataiku:

Firstly, the development of the solution inside Dataiku led us to the Business Solutions team and the built-in projects inside the platform. This enabled our resulting project to be even stronger as a result of additional input, while also reinforcing some of the decisions we had made as a team when building our first model.

Secondly, the deployment infrastructure within Dataiku allowed us to get our project into production at lightning speed. Within days of completing development, we had the project live and automating predictions over millions of customers, with any changes easy to manage and deploy.

Value Type:

• Improve customer/employee satisfaction
• Increase revenue
• Reduce cost
• Save time