Codrva Digital

A Quarter of Your Churn Is Expired Bank Cards

Most retention work targets customers who decided to leave. A meaningful share never decided anything, and winning them back needs no product change at all.

C
Codrva Team
Published Nov 16, 2026
7 min read
Light cream cover with a SaaS icon, the headline in bold condensed type and a panel listing four findings: These customers did not leave; Most retries are configured badly; Tell the customer, and tell them usefully; Do not cancel on the last retry

Light cream cover with a SaaS icon, the headline in bold condensed type and a panel listing four findings: These customers did not leave; Most retries are configured badly; Tell the customer, and tell them usefully; Do not cancel on the last retry

Median monthly churn for B2B SaaS sits around 3.5 percent. That figure splits into roughly 2.6 percent voluntary and 0.8 to 0.9 percent involuntary, the latter caused by failed payments, expired cards and declined transactions rather than by anybody choosing to leave.

So somewhere around a quarter of your churn is not a product problem, a pricing problem or a competitor problem. It is a payment problem, and it is the only category of churn you can reduce without changing anything a customer experiences.

These customers did not leave

The distinction matters because the two require completely different work.

A voluntary churn is a decision. The customer evaluated the product, decided it was not worth the money, and acted. Reducing that means changing the product, the price or the onboarding, all of which take quarters.

An involuntary churn is an accident. A card expired, a bank declined a transaction as unusual, a corporate card was reissued after a fraud alert, or the person whose card was on file left the company. The customer still wants the product. Nobody made a decision at all.

Those customers are, in retention terms, the easiest population you will ever work with. They have already decided in your favour and the system cancelled them anyway.

Most retries are configured badly

Almost every billing system retries a failed payment, and almost every default schedule is wrong in the same way.

Retrying the next day and then giving up after three attempts fails the most common real cases. A reissued card takes days to arrive. A bank hold clears when somebody notices the alert. A finance team updates payment details when they return from leave.

A schedule spread across two to three weeks recovers substantially more than one compressed into three days, and it costs nothing but configuration. Vary the time of day too, because a decline caused by a daily limit will decline again at the same hour.

We look at retry configuration before anything else in retention work, and it is almost always at whatever the platform shipped with. Nobody chose it; it simply was never revisited.

Tell the customer, and tell them usefully

The second lever is the email, and most are written as though the customer did something wrong.

Say what happened, say what will happen next, and give one link that goes straight to updating the card. Not a link to a login page, then a settings menu, then a billing tab. The number of steps between the email and the corrected card is the whole conversion rate of that message.

Send it to more than one address where you can. The person whose card is on file is frequently not the person who reads billing email, and in larger accounts neither is the person using the product daily.

And tell the account owner inside the product too, not only by email. An in-app notice reaches somebody who never opens billing messages, and it arrives when they are already logged in and able to act.

Do not cancel on the last retry

The default behaviour at the end of a failed retry sequence is usually immediate cancellation, and that is the most expensive line of configuration in the system.

A grace period during which the account still works, clearly flagged, recovers customers that hard cancellation loses permanently. The customer who returns from two weeks of leave to a working account with a warning fixes their card. The one who returns to a cancelled account and lost data frequently does not come back.

Set the grace period against your own numbers rather than a default. If most recoveries happen within ten days, a fourteen day window costs you very little and catches nearly all of them.

What you must not do is let it run silently. An account in grace should say so every time somebody opens it, or you have simply delayed the same loss.

We have seen a client recover a meaningful share of a quarter's lost accounts by changing two settings: the retry window and what happens at the end of it. Neither touched the product, and the work took an afternoon.

Prevent the failure before it happens

The cheapest recovery is the one you never need.

Card expiry dates are known in advance, so prompt before they arrive rather than after they bite. Most payment providers also offer an updater service that refreshes reissued card details automatically, and it is frequently available and switched off.

For accounts above a certain value, offering invoicing rather than card payment removes the failure mode entirely. That is more administration and it is worth it on the accounts where a declined card costs you a meaningful sum.

Measure the two separately or you will chase the wrong one

The reason this goes unaddressed is that most reporting shows one churn number.

Split it. Report voluntary and involuntary as separate lines, and watch them separately, because they respond to entirely different work and averaging them hides which one is moving.

Then judge the effort proportionally. Reducing involuntary churn by half is a fortnight of billing configuration. Reducing voluntary churn by half is a year of product work and may not be achievable. Both are worth doing and only one is available this month.

It also feeds the retention figure everyone reports, and expansion revenue now accounts for 40 to 50 percent of new ARR, which makes keeping existing accounts arithmetically more valuable than it used to be. The pricing side of that is in why your SaaS pricing model is breaking, and the cost side in setting a price when your own cost moves.

Our SaaS development team treats billing recovery as retention work rather than as finance plumbing, because that is where it actually sits.

Frequently Asked Questions

How much churn is involuntary?

Roughly 0.8 to 0.9 percentage points of a 3.5 percent median monthly rate, so about a quarter. Those customers never decided to leave; a payment failed and the system cancelled them.

What retry schedule works?

One spread over two to three weeks rather than compressed into three days, with attempts at varying times of day. Reissued cards and cleared bank holds both take longer than most default schedules allow.

What should the failed payment email contain?

What happened, what happens next, and one link straight to updating the card. Every step between the email and the corrected card costs you recovery.

Should we cancel at the end of the retries?

Not immediately. A clearly flagged grace period recovers customers that hard cancellation loses for good. Size it against when your own recoveries actually happen.

What prevents the failure entirely?

Prompting before known expiry dates, enabling your provider's card updater service if it is available, and offering invoicing on higher value accounts where a decline costs real money.

Why is this not already handled?

Because it sits between finance and product and belongs to neither, and because most reporting shows a single churn number that hides it. Split the metric and it becomes visible.

Which should we work on first?

Involuntary, because halving it is a fortnight of configuration while halving voluntary churn is a year of product work. Do the available one first, then the hard one.

Does this matter at small scale?

Proportionally, yes, and the fix costs the same regardless of size. Retry schedules and a good recovery email are a day of work whether you have eighty customers or eight thousand.

Talk to our team if your churn number is one line and nobody knows how much of it is failed cards.

Share this post:
← Back to Blog

Comments (0)

Leave a Comment

Minimum 10 characters

No comments yet. Be the first to comment!

Chat with us