Predictive churn model drives £1m growth in EBITDA

Profile

A PE-backed accounting and HR SaaS provider.

Situation

A PE-backed accounting and HR SaaS provider was aiming to reduce customer churn and drive value from its existing customer base. They wanted to be able to identify high-risk customers so that they could be prioritised for retention callouts.

Action

Over a 2-week period, we analysed and parametrised 2.5 years of revenue data (e.g. variance, gradient, averages, and other available customer data). This was then used to tune a Machine Learning model with using the most influential parameters.

The model was four times more likely to accurately identify churners compared to choosing at random. We also developed a bespoke and secure web-portal to disseminate the model outputs across the business and ensure ease of access for all stakeholders.

Impact

The business regularly used predicted churn scores from the model to prioritise the retention callouts, with an estimated >£1 million in EBITDA savings. Insights from the model drivers also helped optimise sales strategy (e.g. enhancing its customer training programme, an approach which made customers ‘stickier’).

So what?

We specialise in developing pragmatic solutions that are informed by advanced analytics and easily deployed within the operational context. Sometimes what’s needed is an advanced model with outputs in Excel. Other times it might be best to fully integrate with your CRM or sales management platform. Whatever the solution we understand that your team needs to trust the results and fully understand the rationale for the recommendations.