Resources

Dunning by SMS: The Text-Message Recovery Playbook

Email is where dunning has always lived, and email has a problem: the customer who most needs to see your message is the one least likely to open it. A failed payment sits unnoticed in a promotions tab while the subscription quietly dies. A text message doesn't fix everything about dunning — sent badly, it's the fastest way to annoy a customer you were trying to keep — but as one layer inside a recovery sequence, SMS reaches people email structurally can't. This guide covers when a text recovers a payment an email won't, the consent and compliance rules that are not optional, the cadence that works, and what to actually say in one short message.

Why SMS works where email stalls

The mechanism is mundane, which is why it works. A text arrives on the device the customer is already holding, surfaces on the lock screen, and asks for ten seconds of attention. An email asks for a trip to an inbox the customer may check twice a day — alongside forty other senders having the same idea. For a certain class of failure, that difference is the whole game:

• The customer never knew. The most recoverable failed payments aren't refusals; they're accidents — an expired card, a replaced card, a bank that swapped a number. The customer intends to pay. They just need to find out, once, from a channel they'll actually see, with a link that works from the phone in their hand.
• The clock is short. Some failures resolve themselves if you wait (insufficient funds around payday). Others decay: every day an expired card goes unaddressed, the chance the customer bothers shrinks. SMS is the channel you spend on the failures where speed matters most.
• The ask is tiny. "Your card expired — tap to update it" is a complete, honest message. Dunning emails have to carry context, reassurance, and detail. A text carries one fact and one link, and for a large share of failures that's all the customer needed.

What SMS is not: a replacement for email, or a louder version of it. Email still carries the full explanation, the receipt trail, and the customers who prefer it. The text is the tap on the shoulder; the email is the letter. Our guide to dunning emails that work covers the letter — this guide covers the tap.

The rules are not optional

Business texting is regulated in a way business email is not, and "our dunning tool sent it" is not a defense. The load-bearing requirements, in the US framework most subscription brands operate under:

• Consent before the first text. Marketing and payment texts to a mobile number need the customer's prior consent — collected at checkout or signup, and kept on record. A phone number given for shipping updates is not automatically consent to payment texts; know what your opt-in actually said.
• Opt-outs are instant and permanent. STOP (and its equivalents) must end texting to that number immediately, and the opt-out has to be honored across campaigns, not per sequence. One confirmation message is allowed; then silence.
• Identify yourself. Every text names the business. A payment request from an unknown number reads as a scam, because it usually is one.
• Mind the clock. US rules restrict calling and texting to daytime hours in the recipient's local time (the commonly applied window is 8 a.m. to 9 p.m.). A 2 a.m. "your payment failed" text is both a compliance problem and a churn decision made on your behalf.
• Register your traffic. US carriers require application-to-person messaging to run through registered campaigns (A2P 10DLC for local numbers). Unregistered texts get filtered or blocked — quietly, so your dashboard says "sent" while nobody received anything. If your provider handles registration, confirm it actually happened.

None of this is a reason to avoid the channel. It is a reason to run it through a system that treats the rules as mechanics — consent on file, opt-outs honored globally, sending windows enforced — rather than as a template someone set up once.

The cadence: two texts, placed where they earn their keep

More texts do not recover more payments; they recover more opt-outs. In the sequence we run, SMS appears twice for a standard failure, positioned where its speed matters:

• Day 1, alongside the first email. The failure is fresh, the customer still expects the charge, and the fix — usually a card update — takes a minute. The text carries the merchant's name, the one fact, and the secure update link. The email sent the same day carries the detail for anyone who wants it.
• Day 7, as the sequence's midpoint warning. By now the retry schedule has had its chances and the emails have had theirs. The day-7 text says plainly what happens next: service pauses or the subscription ends on a specific date unless the card is updated. Specific dates outperform vague urgency, and honest stakes outperform manufactured ones.

Two cases bend the pattern. Expired cards get a dedicated cadence — texts in the day 1 / 3 / 7 rhythm — because the customer's action is the only cure, and every day of delay is pure loss. Hard declines (lost, stolen, closed accounts) get at most one clear text pointing at the update page; texting repeatedly about a card that can never work is harassment with extra steps. And the moment a payment lands — by retry, by update, by any path — every channel stops. A "your payment failed" text that arrives after the payment succeeded teaches the customer to ignore the next one, which might be real.

What a dunning text should say

One text, dissected. The anatomy is: brand, fact, link, exit.

• Brand first. "Coffee Club:" before anything else. The customer should know who's texting before they know why.
• One fact. "Your payment didn't go through" or "Your card on file has expired." Not both, not a paragraph. If the reason is known and safe to state (expiry), state it — specifics are what separate a helpful text from a phishing lookalike.
• One link. To a secure, hosted update page — never a bare request for card numbers by text or reply. The page should open on the phone, show the brand, and take under a minute.
• An exit. "Reply STOP to end" — required, and also a trust signal. A business confident enough to offer the exit reads differently from one that hides it.

Example, at full length: "Coffee Club: your card on file expired, so today's box is on hold. Update it here to keep your subscription: [link] Reply STOP to end." That's the whole genre. Anything longer is an email wearing a costume.

Pairing SMS with the rest of the system

SMS does its best work as the middle layer of a sequence that starts before the failure and ends after the retry schedule does:

• Prevention first. Expiring-card alerts before renewal and card-updater services quietly eliminate the failures SMS would otherwise chase. The cheapest text is the one a prevented failure makes unnecessary.
• Retries run silently. A soft decline that clears on a well-timed retry never needs a message at all. Texting about failures the customer never had to know about spends trust for nothing — the sequence messages when customer action is the cure, not on every attempt.
• Voice is the escalation, not the default. For a high-value account still unresolved late in the sequence, a human call converts in a way no message can — and costs like one. Reserve it for accounts where the subscription's value justifies a conversation.
• Attribution closes the loop. Every recovered dollar should record which lever won it — retry, email, text, call — so the sequence is tuned on evidence instead of instinct. Our counting rules, including how we separate total recovered from directly recovered, are public: how Swipe Again counts recovered revenue.

Swipe Again runs exactly this layering: decline-code retries on a schedule the decline dictates, branded email under your control each day of the sequence, SMS on day 1 and day 7 (with the expired-card cadence above), and voice reserved for high-value accounts — one flat $179/month, no percentage of what it recovers. And the guarantee that stands behind it: recover $179 in your first 30 days, guaranteed — or Swipe Again is free the next 30 days. Recovered revenue means every failed payment collected while Swipe Again is active, from any recovery source, as defined in Section 3 of our Terms. The full system this channel belongs to is laid out in our guide to recovering failed payments.

Frequently asked questions

Will texting customers about failed payments annoy them into cancelling?

Bad texting will. Two restrained, useful texts — one when the problem is fresh, one before a real consequence — read as service, not harassment, because for most customers the failure is news and the fix is easy. The annoyance risk lives in volume and irrelevance: daily texts, texts after payment, texts about cards that can never work. Cap the channel, aim it, and stop it the moment it succeeds.

Can I just add SMS to my existing email dunning?

The channel, yes — bolt-on texting exists. The question is coordination: if the texts run on a separate clock from your retries and emails, customers get a "payment failed" text the same morning a retry quietly succeeds, or three messages in a day from two systems. SMS earns its place inside one sequence with one schedule, not as a second, parallel nag.

What about customers who never gave a phone number?

Then the channel doesn't exist for them — which is a checkout decision, not a dunning one. Collecting a mobile number at signup, with honest consent language, is what makes the day-1 text possible later. For everyone else, the sequence leans on email and timing, which is why the email frameworks matter as much as they do.

The short version

SMS recovers the failures email can't reach: the unnoticed, the time-sensitive, the one-tap fixes. It works when it's consented, identified, restrained — two texts placed where speed matters, an expired-card cadence for the case that needs one — and silent the moment the payment lands. Run it as one layer inside a real recovery sequence, measured like everything else, and it's the highest-leverage message in dunning. Run it as a megaphone, and it's an opt-out generator. The wider system — retries, email, SMS, voice, and the cancel moment — is mapped in our involuntary churn guide.

Related reading: Dunning channels compared · In-app payment alerts · Winback SMS templates