top of page

How we turned RFV into a churn-predictor

investdata

We recently completed a churn audit for a subscription-based platform and have some insights that could be relevant for all subscription-based models.


We developed a classic Recency-Frequency-Value (RFV) metric for each customer, which correlated with 61% of the churn.


By enhancing the RFV concept with our own analysis of customer "interactions"—what we call the RFV-I model (details on the methodology will be shared in an upcoming post)—we were able to predict 89% of the churners.


Key findings include:

  • 21% of churners contacted the team via forms in the month preceding their churn, with interactions fragmented across the platform.

  • 10% of churners experienced login issues.

  • 37% encountered payment errors during the renewal process.

  • Additional issues included unsubscribing from marketing emails, sending personal emails to the team, encountering page errors on the platform or excessive activity on specific account pages in the client zone.


Overall, 55% of unique churners faced 'administrative' issues before the renewal moment. These issues shifted their mindset in a way that traditional retention strategies couldn't address.


This realization was a major aha moment for us: all retention efforts are futile if a customer is facing personal hurdles. Sending "personal emails with personal offers and personal links" to customers who can't log in to your platform is ineffective. The metrics are clear: on average, these customers churned 26 days after experiencing (failed?) interactions. This common sense needed to be translated into dashboards and automation.


In addition to predicting churn, could we prevent it? Using our RFV-I scoring, we identified individuals who needed a personalized retention approach. Compared to the blue control panel, this resulted in a yellow engagement boost for those nearing the churn danger zone (see graph). We believe this demonstrates a positive effect (retention) from addressing a negative intention (churn). We welcome any suggestions!




Our main takeaway: RFV is a good start for churn prediction, but it becomes truly powerful when combined with tracking personal interactions. This makes your retention tactics more relevant and personal, creating a unique opportunity to turn potential churners into brand ambassadors.



Recent Posts

See All

Comments


bottom of page