Glossary /

    Customer churn

    Written by PEKO Team.Last updated: 2026. 05. 21..

    Customer churn is the percentage of customers who stop visiting your restaurant or buying from your café over a defined period (typically 30, 60, or 90 days).

    Published: 2026. 05. 01.

    In F&B, churn is harder to spot than in subscription businesses because there is no cancel button. A guest doesn't tell you they've left — they just stop showing up. By the time you notice, you've usually already lost them.

    Industry benchmarks suggest 60–75% of first-time F&B customers never return. Over a 90-day window, the average independent restaurant loses 30–40% of its 'active' customer base. The real cost is not the lost transaction — it's the cost of replacing each lost guest with a new one through paid acquisition.

    Modern customer retention platforms like PEKO use AI to predict churn before it happens, looking at recency, frequency, average ticket size, and visit cadence to flag guests who are drifting away. Once flagged, automated win-back campaigns (Zalo OA, email, SMS) re-engage them at the right moment.

    Worked example

    If you had 1,000 active customers at the start of the quarter and only 620 visited again within 90 days, your quarterly churn rate is (1,000 − 620) / 1,000 = 38%.

    FAQ

    What is a good churn rate for restaurants?

    There is no single 'good' number, but most independent restaurants and cafés sit in the 30–45% range over a 90-day window. Anything below 25% is excellent and usually requires an active retention program.

    How do I calculate customer churn for my café?

    Pick a window (30, 60, or 90 days). Count active customers at the start. Count how many of those visited again within the window. Churn = (start − returners) / start, expressed as a percentage.

    Can AI predict customer churn before it happens?

    Yes. AI churn models look at recency of last visit, average frequency, ticket size trend, and seasonality to score each customer's likelihood of churning in the next 30 days, often before the customer themselves has decided.

    Sources

    The definitions and figures on this page reference the sources below:

    Related terms

    People also read