Skip to content
ADVISORY RECONCILIATION · STRIPE ↔ ENTITLEMENTSRun the browser-only CSV audit
The journal
DetectionNovember 14, 2026 · 8 min read

How to detect premium access without an active subscription

Stale premium access is one of the clearest revenue leaks — and one of the easiest findings to get wrong. The difference is policy and evidence, not a clever query.

The basic test sounds simple: join the subscription table to the entitlement table and flag users who have premium access but no active subscription. In practice, each side of that sentence needs a definition. What counts as active? What counts as premium? How long is the billing system allowed to be out of date before the comparison is meaningless?

1. Define “active” as a policy, not a status string

Stripe subscriptions can be active, trialing, past_due, unpaid, paused, canceled, incomplete, or incomplete_expired. A product may reasonably treat trialing users as entitled and past_due users as entitled during a grace window. Canceled subscriptions may still be paid through the end of the current period.

Write these decisions down before generating findings. A stale-access report with no stated policy will be argued about rather than fixed.

  • Active: normally entitled.
  • Trialing: entitled only if the product grants trial access.
  • Past due: often entitled during a dunning window, but not indefinitely.
  • Canceled: usually entitled only through current_period_end, if the subscription was canceled at period end.
  • Unpaid, incomplete, incomplete_expired: normally not entitled, unless a documented manual exception applies.

2. Resolve identity before comparing state

The most common reason an audit fails is not a billing rule. It is identity. An application user may have no Stripe customer ID, several historical customer IDs, or a shared organization account with multiple seats. Comparing email addresses is risky: people change email, share inboxes, and use aliases.

Prefer an explicit mapping table written at signup or checkout time. Fall back to email only as a review signal, never as automatic proof.

-- Illustrative shape, not a drop-in schema.
select
  e.user_id,
  e.stripe_customer_id,
  e.is_premium,
  s.status,
  s.current_period_end
from app_entitlements e
left join latest_stripe_subscriptions s
  on s.customer_id = e.stripe_customer_id
where e.is_premium = true
  and s.customer_id is null;

3. Compare only complete, recent snapshots

A missing subscription in a partial reconciliation is indistinguishable from a canceled subscription. Require an explicit complete marker from both sources, and record the snapshot time. If the Stripe snapshot is six days old, a subscription canceled yesterday will look active. If the entitlement snapshot is partial, a user may appear to have access when they do not.

Staleness is not a cosmetic warning. It changes the meaning of the finding, so it belongs in the finding itself.

4. Separate access from intention

Not every stale entitlement is a mistake. Support may have granted access while a payment dispute is resolved. A founder may have a lifetime account. A partner may be on a manual contract. Record these as expected exceptions with an expiry date, and make the reconciliation engine cite the exception rather than silently suppressing the case.

The output should tell a reviewer what changed, when it changed, and what evidence is missing. “Stale access” alone is not enough to act on.

5. Measure the value only when the price is known

If the subscription record contains a verified price and currency, you can show a monthly value for the plan. If it does not, show “unknown.” Do not infer a price from the customer’s plan name or from a marketing page. An estimate that cannot be defended is worse than a blank field.

FROM READING TO EVIDENCE

Your exports are more interesting than our examples.

Run the local reconciliation and see which categories actually appear in your data. The audit needs no account.