Back to all case studies
Entertainment

Customer Churn Prediction for Online Sports Betting Company

Customer Churn Prediction
Key Details
  • Challenge
    Reduce customer churn on a betting platform
  • Solution
    Customer churn prediction model
  • Technologies and tools
    Python

Client

Sports betting company that stands out among other online bookmakers as a reliable company that strives to create a long-term association with each customer. Although the company is relatively young, it is already trusted by almost one million active fans.

Challenge: reduce customer churn on a betting platform

Hundreds of betting websites offer their services to millions of fans who like to bet on sports online. The client spends large amounts of money to acquire new customers, that’s why it is more important than ever to prevent customer churn and turn new customers into loyal ones.

The rate of customer churn directly affects the growth of the company. To keep that rate low marketing and IT department leaders are looking for a solution that can predict customer churn.

Solution: customer churn prediction models

InData Labs was chosen for 3 main reasons:

  1. proved experience in building predictive analytics solutions
  2. proved experience in the betting and gaming industries
  3. readiness of InData Labs to customize its solution for specific needs of the client

InData Labs starts to work closely with client’s business analysts and marketing team to get familiar with the business, processes and industry specifics.

Customers in the betting industry appeared to be very dynamic, new players come and go very fast. Churn rates in this industry are much higher than in telecom or gaming.

The betting company saw 40% of their customers churning just after submitting a registration form, even before placing the first bet. Marketing team believes that retaining these new customers is essential for the future growth of the company. For that reason, our data science team builds a special model that can predict churn for that segment of new customers with very limited information. As a result, they were able to see that the model can successfully identify active and passive users based on the information they’ve provided with a registration form. InData Labs also helped to identify motivational bonuses for different customer groups. Within a couple of months, the client runs numerous tests to define which bonuses work better for different customer segments.

Other models in InData Labs solution aim to make predictions for players ‘with a history’. The inputs for these churn prediction models are:

  • customer profile data: age, gender, location, time and date of registration
  • historical data about clients activity at the betting platform: lifetime, current account balance
  • customer recent activity: quantity of bets, sum of bets, number of gains and losses, amounts of gains and losses, frequency of deposits and withdrawals from the account, amounts of deposits and withdrawals from the account
  • customer preferences: mobile app/ website, popular sports events/ specific sports events, sports preferences.

InData Labs’ solution transforms the raw customer data into features for churn prediction models. The outputs of the models are probabilities of churn in the course of 3 weeks. It’s easy for the client’s marketing team to interpret outputs of the machine learning system and to operationalize the insights.

The final solution for the betting company includes three predictive models, allowing the company stay data-driven at different stages of customer lifecycle:

  • the 1st model works ‘on the fly’ predicting the propensity of just registered user to bet at least once at the website
  • the 2nd model predicts churn of new customers in the course of 3 weeks
  • the 3d model predicts churn of a loyal customer in the course of 3 weeks

Result: better understanding of customer behaviour and higher customer retention rate

InData Labs solution allows its customer to predict customer churn and take up necessary measures to prevent it. The company saw a 20% higher customer retention rate within the first three months after deployment of InData Labs solution.

Tags:
  • Entertainment
  • Machine Learning
  • Python
Share this post

Contact Us

Drop us a line about your project or describe a challenge your company needs help solving. We’d love to discuss how InData Labs can work with you.

By clicking Send Message, you agree to our Terms of Use and Privacy Policy.