Predict and prevent customer churn.

Finsi uses machine learning models trained on your store's historical data to predict churn. The models analyze purchase recency, order frequency, subscription behavior (skips, pauses, downgrades), engagement signals, and product-level patterns to assign each customer a churn probability score. These scores update regularly as new data comes in, so you always know who is most at risk.

Cohort analysis groups customers by the time period they were acquired (e.g., all customers from January) and tracks their behavior over subsequent months. You can see retention rates, repeat purchase rates, revenue per cohort, and subscription survival curves. This helps you identify whether retention is improving or declining over time and which acquisition channels bring the most loyal customers.

Customer health scores in Finsi are calculated using recency of last purchase, purchase frequency, monetary value (total and average spend), subscription status and tenure, engagement signals (email opens, site visits), and recent behavioral changes like skipping deliveries or reducing order size. Each factor is weighted based on its predictive power for your specific business.

Finsi connects to subscription platforms like Recharge, Skio, Smartrr, and Loop to track every event in the subscription journey—activation, renewals, skips, pauses, product swaps, upgrades, downgrades, and cancellations. This timeline view helps you identify critical moments where customers are most likely to churn and the optimal timing for retention interventions.

Finsi generates specific, data-driven retention recommendations such as targeted discount offers for at-risk segments, optimal timing for re-engagement emails, product swap suggestions for customers showing fatigue, personalized win-back campaigns for recently churned customers, and subscription upgrade offers for high-engagement users. Each recommendation includes the target segment, suggested timing, and expected impact on retention.