What is churn rate, and how do you measure it in Metabase?

Churn rate is the share of recurring revenue — or customers — you lose in a period. It's the counterweight to growth: even fast acquisition can't outrun high churn. Measure it in Metabase from billing data synced into a database (Stripe, Chargebee, Paddle, or Recurly).

How is churn rate calculated?

Pick a period (usually a month) and a fixed starting cohort:

• Revenue churn rate = churned MRR in the period ÷ MRR at the start of the period.
• Customer churn rate = customers lost ÷ customers at the start.

Measure loss against the cohort that existed at the start of the period, not the end — otherwise new signups dilute the rate and hide the problem.

Revenue vs. customer churn — and voluntary vs. involuntary

• Revenue churn weights each loss by its MRR. Losing one large account can dwarf many small ones.
• Customer (logo) churn counts accounts equally. Good for product-market-fit signals.
• Voluntary churn — the customer chose to cancel. Fix with product and value.
• Involuntary churn — a payment failed. Fix with dunning and card-update flows; see failed-payment rate.

When expansion outweighs churn, net revenue retention exceeds 100% — track it alongside churn.

What data does churn rate need?

• Subscription start and cancellation/expiry timestamps.
• A cancellation reason or type to separate voluntary from involuntary.
• A monthly MRR model for revenue churn.

SQL patterns

-- Requires a monthly MRR model: one row per subscription per month
WITH by_month AS (
  SELECT
    month,
    SUM(mrr)                                          AS mrr,
    LAG(SUM(mrr)) OVER (ORDER BY month)               AS mrr_start,
    SUM(CASE WHEN is_churned THEN mrr_lost ELSE 0 END) AS churned_mrr
  FROM modeled_mrr
  GROUP BY month
)
SELECT
  month,
  ROUND(100.0 * churned_mrr / NULLIF(mrr_start, 0), 2) AS revenue_churn_pct
FROM by_month
ORDER BY month;

WITH months AS (
  SELECT generate_series(
    date_trunc('month', CURRENT_DATE - INTERVAL '11 months'),
    date_trunc('month', CURRENT_DATE),
    INTERVAL '1 month'
  ) AS month
)
SELECT
  m.month,
  COUNT(*) FILTER (
    WHERE s.created <= m.month
      AND date_trunc('month', s.canceled_at) = m.month
  ) AS churned,
  COUNT(*) FILTER (
    WHERE s.created <= m.month
      AND (s.canceled_at IS NULL OR s.canceled_at > m.month)
  ) AS active_at_start,
  ROUND(100.0 * COUNT(*) FILTER (
    WHERE s.created <= m.month
      AND date_trunc('month', s.canceled_at) = m.month
  ) / NULLIF(COUNT(*) FILTER (
    WHERE s.created <= m.month
      AND (s.canceled_at IS NULL OR s.canceled_at > m.month)
  ), 0), 2) AS customer_churn_pct
FROM months m
CROSS JOIN subscriptions s
GROUP BY m.month
ORDER BY m.month;

Pitfalls

Where this metric applies

• Stripe + Metabase — subscription canceled_at and status
• Chargebee + Metabase — subscription cancelled_at and cancellation reason
• Paddle + Metabase — subscription canceled_at and scheduled changes
• Recurly + Metabase — subscription state and expires_at

Related

Analytics

Metrics

FAQ