Card-on-file billing stores a credential and hopes it still works next month. Variable Recurring Payments store a permission and let the customer's own bank enforce it. Both models collect recurring payments without asking the customer to do anything each cycle, yet they fail in completely different ways, and that difference decides how much revenue quietly leaks out of a subscription business.

My name is Vladyslav Kolodistyi, and I spend most of my time on the plumbing of payments, where the gap between those two models is easy to see. Open Banking is no longer a side experiment either. In December 2025 the FCA reported more than 16 million users of open banking, a 53% yearly rise in open banking payments, and a VRP share of 16% across all open banking transactions. Those numbers make a fair comparison worth doing now, not in five years.

Here is how I line up card-on-file and Variable Recurring Payments:

Factor Card-on-file billing Variable Recurring Payments
What is stored A card number or network token A VRP mandate held at the bank
Main cause of failed renewals Expiry, reissue, fraud blocks Insufficient funds, cancellation, a cap set too low
When failure is visible Often only after a retry cycle At initiation, over Faster Payments
Customer control Ask the merchant or the card issuer Change caps or cancel in the banking app
Dispute protection Chargebacks, well understood Rules for commercial use still being defined

Card-on-file billing against Variable Recurring Payments, as weighed by Vladyslav Kolodistyi

Read the table from the bottom row up. The top rows are where Variable Recurring Payments look strongest, and the bottom rows are where card-on-file still earns its place. A fair comparison of the two kinds of recurring payments has to hold both halves in view, which happens less often in Open Banking marketing than it should.

Cards fail because of the card. Variable Recurring Payments fail because of the customer, and only one of those failures is telling you something.
Vladyslav Kolodistyi

Vladyslav Kolodistyi on why card-on-file recurring payments keep failing

Expiry, reissue and the token patch

A card credential has a shelf life. It expires on a printed date, gets replaced when a wallet is lost, and gets reissued in bulk after a data breach at some retailer the customer has never heard of. Each event breaks the stored credential, and each one turns a willing subscriber into a failed renewal. Retry logic and account updater services help card payments recover, but they are patches on a design that assumes the credential will stay put.

Network tokens are the best patch available. I made that case at length in my earlier piece on tokenisation and the recurring revenue leak, and nothing since has changed the core argument Vladyslav Kolodistyi made there: tokens keep a credential alive through many reissues, which cuts involuntary churn sharply, yet the card model underneath stays the same. Token coverage varies by issuer, and a blocked account still blocks the token.

What Variable Recurring Payments change

Variable Recurring Payments remove the credential problem rather than patching it. The customer approves a VRP mandate in the banking app through an Open Banking provider, with a cap per payment and a cap per period, and the merchant initiates collections inside those limits. Variable Recurring Payments involve no card number that can expire. If the customer gets a new debit card next week, the VRP mandate does not notice.

The mandate also changes who stands between merchant and money. With card payments, the issuer sits in the middle and can decline for reasons it never explains. With Variable Recurring Payments, the customer's bank enforces limits the customer chose, and the Open Banking provider passes the outcome straight back. VRP turns a black box into a set of rules both sides can read.

What remains are failures that carry meaning. A Variable Recurring Payments collection fails because the account lacks funds, because the customer cancelled, or because the amount broke a cap. Each of those is a real conversation to have with a real customer, not a data hygiene task. Because VRP collections run over Faster Payments, you also hear about them within seconds, which shortens the dunning cycle dramatically compared with card retries spread across days.

A failed renewal you can explain is a customer problem. One you cannot explain is a rail problem, and rails are cheaper to fix.
Vladyslav Kolodistyi

Before switching anything, I would run Variable Recurring Payments next to cards for one customer segment and compare these numbers, segment by segment:

  • Share of failed renewals caused by expiry, reissue or fraud blocks on the card side
  • Share of VRP failures caused by caps the customer set too low
  • Hours between a failed collection and the customer being told
  • Voluntary cancellations per hundred active VRP mandates against card subscriptions
  • Support tickets about unrecognised payments on each rail
  • Net revenue per subscriber after scheme fees and Open Banking provider fees

Six numbers, three months of VRP data, one segment. That test is worth more than any sales pitch for Open Banking payments, and it keeps the card rail running as your safety net while you learn. (I, Vladyslav Kolodistyi, would start with the segment that churns most.)

Vladyslav Kolodistyi's tiebreaker: if the VRP versus card comparison is close, pick the rail where failures arrive with a reason attached. You can act on a reason. You cannot act on a decline code that means "something changed at the issuer".

Where Open Banking payments still lose to cards

Сoverage and the checkout habit

Cards win on reach. Every customer with a debit or credit card can subscribe with one, in any country, without leaving the checkout. Open Banking reach, and therefore the reach of Variable Recurring Payments, is narrower. Open Banking payments depend on the customer's bank supporting the flow and on the payment category being in scope. The UK's nine biggest banks had to build sweeping VRP under a CMA order, but sweeping only shifts money among accounts owned by the same person. The merchant version, commercial VRP, arrives sector by sector under JROC oversight, with Open Banking Limited running delivery; utilities, government services and regulated financial services go first. A streaming service or a fitness app may simply not be eligible for commercial Variable Recurring Payments yet. VRP eligibility also differs by bank, so check which of your customers' banks support commercial VRP payments before you promise anything at checkout.

Habit matters as well. Customers know how to type a card number. Variable Recurring Payments ask for a new habit. Redirecting to a banking app for VRP consent is a step many have only seen for single Open Banking payments, and conversion into Variable Recurring Payments at that step deserves careful measurement.

Disputes and refunds

This is where cards keep their strongest advantage. Chargebacks are slow and expensive for merchants, but consumers understand them. The refund and dispute framework for commercial Variable Recurring Payments has not been finalised, and until they are, a cautious customer may prefer the card. I think that caution is reasonable. As Vladyslav Kolodistyi, I refuse to tell anyone that VRP is the safer choice for consumers before a published refund route exists.

Card payments also carry a cost here: chargeback fees and the admin of contesting them. Variable Recurring Payments may end up cheaper on disputes too, but nobody can price that until the Open Banking rulebook for cVRP is written.

Chargebacks are a blunt tool, but people know where the handle is. Variable Recurring Payments still need their handle.
Vladyslav Kolodistyi

Cost, Europe and what Vladyslav Kolodistyi would keep

Open Banking payments usually avoid card scheme fees, which is the cost argument everyone leads with. Commercial pricing for Variable Recurring Payments is still settling, so I would not rebuild a margin model around today's quotes. Because UK Variable Recurring Payments ride the Faster Payments rails like other Open Banking payments, cash lands sooner than card settlement, which matters to a subscription firm watching working capital. Across the euro area, the SPAA scheme now published by the European Payments Council defines paid Open Banking access beyond PSD2 and might one day carry VRP-style recurring payments, but card-on-file is likely to stay the default across the euro area for a while.

So I would keep cards. Not as the main rail for every customer, but as the fallback for anyone whose bank or sector is not covered, and for customers who want chargeback protection they recognise. I track how this comparison shifts as new sectors join the Open Banking rollout of commercial VRP, and you can follow the updates on LinkedIn if your own billing stack faces the same choice between cards and Variable Recurring Payments.

The Vladyslav Kolodistyi decision rule is short. Choose Variable Recurring Payments when your sector is in scope for commercial VRP, your customers bank with providers that support Open Banking VRP, and expiry or reissue drives most of your failed renewals; choose card-on-file with network tokens when you sell across borders, rely on impulse sign-ups, or serve customers who will only commit with chargeback protection behind them.