What are failed payments costing you?

How the arithmetic works

Three inputs, one constant, four lines of maths. Nothing is hidden and nothing is rounded in our favour.

The constant is 9%. Subscription businesses lose an average of around 9% of recurring revenue to failed payments — the figure reported by Baremetrics and by Paddle / ProfitWell. We treat that 9% as your monthly exposure: the slice of recurring revenue that hits a declined card somewhere between the charge attempt and the retry ladder giving up. What you actually lose is the part of that exposure your recovery system never gets back.

1. Monthly exposure = MRR × 9%
2. Recovered today = monthly exposure × your recovery rate
3. Lost each month = monthly exposure − recovered today
4. Lost each year = lost each month × 12

Subscriber count never touches the revenue figure. It drives two context numbers only — average revenue per subscriber, and roughly how many charges decline in a month. That second one assumes declines fall evenly across order values, which in reality they don’t: a decline is marginally more likely on a larger basket. Read it as an order of magnitude, not a count.

If you read the published 9% the other way — as revenue already lost, net of whatever you recover today — set the recovery-rate field to zero. The headline becomes MRR × 9% × 12, which is exactly that reading. The calculator supports both because the published figure is genuinely ambiguous, and we would rather show you the seam than paper over it.

One number we deliberately don’t supply is a benchmark recovery rate. We haven’t found a source for it we’d be willing to put a client’s budget behind, and we aren’t going to invent one to make the second figure bigger. The ladder prices every rate instead. When we have enough of our own client data to publish a real one, it will appear here with the sample size next to it.

How to read your result

It is exposure, not a cheque. The figure is what is at stake if your recovery rate stays where it is for twelve months. It is not incremental revenue you can book next quarter, and no build recovers all of it — hard declines on closed accounts are gone whatever you do.

It annualises a single month. If your subscriber base is seasonal — and coffee, skincare and anything that sells hard in Q4 usually is — take the number as a run-rate at today’s MRR rather than a forecast of the next twelve.

The recovery rate is the input most people guess. If you typed a figure because it felt about right, that is itself the finding. Almost every subscription platform will tell you how many charges failed and how many eventually settled; almost nobody looks. The gap between the rate you assume and the rate you actually have is often larger than anything else on this page.

Cross-check it against your churn. Involuntary churn — cards, not customers — is commonly 20–40% of all churn (Baremetrics; Paddle / ProfitWell), and typical monthly subscription churn in consumer goods runs 6–7% (Recurly Consumer Goods Churn Benchmark; Recharge DTC consumer panel). Multiply the two and roughly 1.2 to 2.8 points of monthly churn is billing rather than dissatisfaction. If your own involuntary split sits far below that band, either your recovery is genuinely good or your reporting is filing failed payments as voluntary cancellations. In our experience it is the second.

What moves the number

Recovery rate is not a setting. It is the sum of seven pieces of infrastructure, and most brands have three of them:

• A retry ladder built on decline codes. A flat schedule retries a stolen-card decline the same way it retries an insufficient-funds one. Timing, attempt count and routing should all read the decline reason first.
• Pre-dunning. A card-expiry warning sent before the charge fails costs nothing and removes the failure entirely.
• A four-to-six touch dunning sequence across email and SMS, written to be helpful rather than threatening. One email is not a sequence.
• Account updater. Reissued cards get caught silently and the customer never learns there was a problem.
• A payment-update page that works on a phone in one tap. Most recovery links land on a desktop-shaped login form. That is where the recovery rate goes to die.
• Decline-reason reporting — soft versus hard, by processor, by cohort — so you can tell an infrastructure problem from a customer one.
• A recovery dashboard you actually read every week, so the number stops being an annual surprise.

That list is the payment recovery build, in order of how cheap each piece is to add.

What to do next

1. Open your subscription platform and find the real recovery rate — failed charges in a month against those that eventually settled. Run this again with it.
2. Split last month’s declines into soft and hard. Soft declines are recoverable revenue sitting in a report; hard ones tell you how much of the exposure is genuinely gone.
3. Fix the cheapest lever first. Pre-dunning and a one-tap update page are usually a week of work and move the rate more than a new dunning sequence will.
4. If you would rather have the number measured than modelled, that is what the Revenue Recovery Audit produces — with the retry logic, the cancel data and the flows read together.