Churn Prediction (AI-Powered)

Automatically engage with players based on their likelihood of churning. With this lifecycle, you can trigger personalised offers at the right moment and keep players active for longer.

🚀 What Is It and How To Use It?

The New Prediction Model (AI-Powered) is an intelligent model available in FT CRM. It uses machine learning to evaluate player behaviour and predict future outcomes.
Every player is assigned a churn probability score (0–100%), which updates automatically every 24 hours. Based on these scores, players move dynamically between segments, which determine what offers they receive.
This lifecycle template uses this segmentation to make it easier than ever to launch AI-driven engagement campaigns with automated incentivisation strategies.
Use this template to:
  1. Anticipate player churn with the help of machine learning.
  2. Target players when their churn risk increases.
  3. Deliver offers tailored to each player’s probability of churning.
  4. Stop communications automatically once a player is active again.
🔗 Integration Requirements
The integration required for this lifecycle is explained in detail in Churn Prediction Model (AI-Powered). Please review that article before setting up this template.

🧮 Entry Conditions

Players can enter this lifecycle in two ways:
  1. Negative Movement from Active
    1. When a player’s churn probability increases, and their segment shifts downward.
    2. Example: moving from Active → At Risk.
  2. Custom Entry Conditions
    1. Automated
    2. Specific date and time
Once inside the lifecycle, players immediately receive an offer aligned with their churn probability.
Entry Conditions
Entry Conditions

🎁 Offer Strategy

In this lifecycle, offer generosity is directly tied to the player’s probability of churning. The higher the risk, the stronger the incentive should be to keep them engaged.
The default lifecycle is divided into four churn-risk buckets:
  1. 🟢 Low Risk (0–30%)
    1. Players show little risk of leaving.
    2. Light or no incentivisation is recommended (e.g., reminder emails, small-value perks).
  2. 🟡 Medium Risk (30–50%)
    1. Players are showing early signs of potential churn.
    2. Moderate incentives are effective here (e.g., small bonuses or personalised offers).
  3. 🟠 High Risk (50–70%)
    1. Players are at a strong risk of leaving soon.
    2. More attractive offers should be introduced to retain them (e.g., larger bonuses, targeted incentives).
  4. 🔴 Critical Risk (70–100%)
    1. Players are highly likely to churn.
    2. The most generous and compelling offers should be delivered at this stage to maximise chances of reactivation.
By scaling offer value according to churn probability, you ensure that resources are allocated efficiently—saving strong incentives for the players who need them most while maintaining healthy engagement across your broader player base.
Lifecycle Events
Lifecycle Events

🗣️ When To Communicate

  1. The model calculates scores every 24 hours (default update: 05:00:00 UTC).
  2. Segmentations update at the same time.
  3. The Lifecycle event trigger by default at 16:00 UTC, but this can be changed
Then, players remain in the lifecycle until:
  1. They become active again (on bet or deposit), or
  2. They become inactive (30 days with no real bets or deposit).

🔄 Next Actions

To ensure continuous engagement:
  1. Set up a Reactivation Lifecycle for players who exit this template as Inactive.
  2. Align the exit conditions of the AI-Powered Prediction Lifecycle with the entry conditions of the Reactivation Lifecycle.
  3. This guarantees a seamless automated journey for every player.