Separate your churn
Involuntary churn tracked apart from voluntary cancellation, because they have different fixes.
SOFTWARE AND SAAS
Your churn number has two components and only one is about your product
The other is cards failing, and most teams underestimate how much of their attrition it accounts for. We advise on that side of it.
Recover mechanical failures
Retry timing, automatic card updates, and escalation to a person when the automation runs out.
Know your acceptance
The share of recurring charges that go through is a number that moves, and most teams have never looked at it.
Defend disputes you cannot see
Every transaction is card not present, so your defence is built from what your own system logged.
Two kinds of churn, and only one is your product
When a customer leaves because your product did not do what they needed, that is a product problem and your team probably knows about it. When a customer leaves because their card expired and the retry logic gave up after three attempts, that is a payments problem and nobody in the building owns it.
The second kind is worth more attention than it gets, because those customers were not trying to leave. They wanted to keep paying. Recovering them costs a fraction of what acquiring them cost, and unlike most retention work, the outcome is measurable within a month.
Most teams we talk to have never separated the two numbers. That is the first thing worth doing, because you cannot fix a problem you are currently reporting as customer dissatisfaction.
Recovering payments that failed for mechanical reasons
Cards expire on a schedule you can see coming. They get reissued after a breach at some retailer with nothing to do with you. Balances run short on the wrong day of the month. Issuing banks decline recurring charges they do not recognize.
None of that reflects a decision by the customer, which is why the recovery work is mechanical rather than persuasive. Retry timing matters more than retry volume, because hammering the same card daily gets you nothing except a worse relationship with the issuer. Card account updater services, where the networks support them, refresh details automatically before you ever see a failure.
Then there is the part software does not do for you. Somebody has to look at the accounts that failed everything and reach out to them like a human being, before the subscription cancels itself.
- Involuntary churn tracked separately from voluntary cancellation
- Retry schedules spaced deliberately across days
- Automatic card updates where the networks support them
- Customers warned before expiry rather than after failure
- Escalation to a person when automated retries are exhausted
Authorization rates are a number you should know
Every recurring charge either goes through or does not, and the share that goes through is a number most teams have never looked at. It is not fixed. It moves based on how transactions are submitted, how they are described, whether the right data accompanies them, and how your acquirer routes them.
A few points of difference here is real money at any meaningful volume, and it does not show up anywhere in your product metrics. It looks like churn.
This is a genuinely technical area, and the honest thing to say is that improving it is an ongoing exercise rather than a setting you switch on. What we can do is tell you where your current numbers sit relative to what is reasonable, and whether the gap is worth working on.
Selling across borders
Software crosses borders by default. A company can be billing customers in twelve countries before anyone has thought about what that means for acceptance.
Two things happen. Authorization rates drop when a transaction is routed a long way from the cardholder's bank, because issuers are more cautious about charges that look foreign. And customers get charged conversion costs they did not expect, which produces support tickets and occasionally disputes.
Presenting prices in the customer's own currency and settling appropriately addresses part of this. Whether it is worth the complexity depends on how much of your revenue actually sits outside Canada and the US. For a business with a handful of international customers, it is not. For one with a third of its revenue in Europe, it usually is.
Disputes when there is nothing physical to point to
You will never have a signature, a chip read, or a delivery record. Every transaction you take is card not present, which is the highest exposure category there is.
Most subscription disputes are one of three situations. The customer forgot they signed up. The customer tried to cancel and could not, or thought they had. Or the charge appeared under a name they did not recognize.
All three are addressable before they happen. Your descriptor should match the product name your customer knows, not your holding company. Renewal notices before annual charges cost you nothing and prevent a category of dispute entirely. Cancellation should be self serve, because a customer who cannot cancel calls their bank instead, and that outcome costs you the revenue plus a fee plus a mark against your account.
- A descriptor matching the product name customers recognize
- Renewal notice sent before annual and higher value charges
- Self serve cancellation that produces a record
- Signup confirmation retained with timestamp and IP
- Usage logs available as evidence when a dispute arrives
Getting classified as high risk without expecting it
Plenty of software businesses end up in the high risk category and are surprised by it. The characteristics that cause it are ordinary in your world. Recurring billing. Card not present volume. Annual prepayment, where you hold money for service you have not yet delivered. Fast growth, which underwriters read as unpredictable.
You may not encounter this at all. But if your account gets reviewed, a reserve appears, or an application gets declined for reasons nobody explains clearly, that is what happened.
Being prepared for it means keeping your dispute ratio low deliberately rather than incidentally, and having your documentation in order before anyone asks. We advise clients on this, and we would rather have that conversation early than after a reserve shows up.
Growth costs money before it makes money
Subscription economics run backwards at the start. You spend to acquire a customer now and recover it over a year or more, which means growing faster makes your cash position worse before it makes it better.
That is a well understood problem and it is also why so many otherwise healthy software businesses are constrained. The unit economics work. The timing does not.
We work with software clients on business capital against recurring revenue, which is a different conversation from lending against inventory or equipment. Whether it fits depends on your retention, your payback period, and how predictable your revenue actually is once involuntary churn is accounted for. That last part is why the earlier sections matter here.
Built for how you sell
Different models, the same recurring charge underneath.
Self serve SaaS
Card on file, monthly renewals, and cancellation that has to work.
Enterprise
Annual contracts, invoicing, and procurement that does not use a card.
Usage based
Charges that vary each cycle, which customers dispute more often.
Platforms
Payments inside your own product, on behalf of your users.
Marketplaces
Money moving between parties, with you in the middle.
Hybrid billing
Self serve and enterprise running side by side on one stack.
Book a call about your billing
Tell us how you bill, what your failed payments cost you, and where customers drop off. We will tell you what we would change and what we would leave alone.