What's the honest churn-risk estimate for a subscription-program migration, and what benchmark should we hold the migration agency to?
Subscriber churn from a subscription-platform migration runs 2 to 5% by the community-reported envelope (July 2026 research), and payment-method transfer is where that loss concentrates. Hold the agency to a card-retention rate, not a churn rate: the number to write into scope is the percentage of active payment methods that transfer without re-collection. Churn follows from it.
Where the 2–5% comes from, and what it hides
| Loss channel | Typical share of the loss | What moves it |
|---|---|---|
| Payment credentials that do not transfer | Largest | Gateway pairing, vault portability, whether cards were tokenized to the processor or the app |
| Subscribers who use the email as a cancel prompt | Second | Message framing, sender identity, timing |
| Billing-date drift and duplicate or missed charges | Third | Length and rigor of the parallel run |
| Portal confusion after cutover | Smallest, but noisiest in support | Whether the new portal keeps the same URL and login method |
The 2–5% band is the community-reported envelope for subscription-platform migrations (July 2026 research), not a Deploi measurement.
What moves the number most. Whether the card vault moves. If both platforms sit on the same processor and the tokens are processor-held, transfer is close to lossless. If the outgoing platform holds the vault, or the gateways differ, some subscribers must re-enter a card, and a meaningful share of them simply will not. Ask that question in week one, not week six.
The benchmark to write into the statement of work. Not "churn below 3%." Churn is downstream of things the agency does not control, including your own email. Write this instead:
- Payment-method transfer rate, stated as a target percentage of active contracts, measured seven days after cutover.
- Billing-accuracy rate across the first full cycle: charges that fired on the right date, at the right amount, once.
- Contract-field parity: cadence, next-charge date, discount, and line items matching source to destination, sampled and signed off.
Those three are measurable, attributable, and the agency can actually be held to them.
The honest boundary. Nobody can quote you a credible churn number in advance, because the dominant variable is your gateway and vault situation, and the second is your own comms. An agency that offers a guaranteed churn figure before auditing your payment stack is selling confidence, not diligence.
The Deploi point of view
Our own position, from building on Shopify. Separate from the facts above.
- Our take: Churn is the wrong contractual metric. Payment-method transfer rate is the right one, and it is the number that predicts the churn anyway.
- What we’ve seen: When a subscription migration goes badly, the post-mortem almost always lands on a card-vault assumption made verbally in a scoping call and never tested against the processor.
- Where we disagree: The vendor migration pitch leads with contract counts moved. Contracts are the easy part. The payment methods behind them are the project.
- What this page adds: which loss channel dominates, and the three measurable benchmarks that replace an unenforceable churn promise.
Reviewed by Martin Dejnicki, Director of SEO & AI Search. Facts verified 2026-09-13.