Email is the default channel for dunning. It's cheap, scalable, and customers expect it. But email has a ceiling. Open rates top out around 40-45% for the first dunning email and drop from there. By email 3, you're reaching maybe a third of recipients. The rest either missed it, ignored it, or filed it as "I'll deal with this later" and never did.
SMS cuts through that noise. A 95-98% read rate, with most messages read within 3 minutes of delivery. For payment recovery, that directness translates to 15-25% more recovered payments when SMS is layered on top of email.
But SMS isn't just "email but on their phone." It has different rules, different costs, different customer expectations, and different compliance requirements. Getting it right matters.
The case for SMS in dunning
The performance gap between email and SMS is well-documented across multiple sources (Gartner, MobileSquared, and payment recovery platform data):
| Metric | SMS | |
|---|---|---|
| Open/read rate | 35-45% | 95-98% |
| Time to read | 2-6 hours | 3-5 minutes |
| Click-through rate | 12-20% | 20-35% |
| Response rate | 2-5% | 30-45% |
| Recovery rate per message | 8-15% | 15-25% |
These aren't marginal differences. SMS gets read by essentially everyone, almost immediately. For a payment recovery message where timing matters and the action is simple (click a link, update a card), that immediacy is the difference between recovering $4,500 and recovering $3,600 on a $50K MRR business.
Combined channel recovery data:
| Approach | Total Recovery Rate |
|---|---|
| Email only (4 emails) | 32-39% |
| SMS only (2 messages) | 28-35% |
| Email + SMS (3 emails + 1 SMS) | 40-50% |
| Email + SMS (4 emails + 2 SMS) | 45-55% |
Neither channel alone matches the combined approach. Email provides context and details. SMS provides urgency and immediacy. Together, they cover different engagement patterns and catch customers in different moments.
When to send SMS in the sequence
Placement matters more than most teams realize. The two most common mistakes: sending SMS first (feels intrusive) and sending it last (too late to recover most payments).
SMS as the first touch: not recommended. A text message about a failed payment before the customer has received an email feels aggressive. Most customers expect billing communications via email first. Starting with SMS also means you've used your highest-impact channel before the customer has had a chance to self-resolve.
SMS in the middle of the sequence: recommended. Day 4-7 is the optimal window. By this point, the first email or two has gone out. Customers who were going to fix it from email alone have already done so. The remaining pool is people who either missed the emails or procrastinated. SMS catches both groups.
SMS as the last touch: decent but suboptimal. A final "your account will be canceled tomorrow" text message does drive action. But by day 12-14, many recoverable customers have already churned mentally. You're getting the urgency of SMS but missing its speed advantage.
Recommended placement:
| Touch | Channel | Timing | Purpose |
|---|---|---|---|
| 1 | Day 0-1 | Notification and context | |
| 2 | Day 3 | Follow-up reminder | |
| 3 | SMS | Day 5-6 | Mid-sequence recovery push |
| 4 | Day 10-11 | Urgency, approaching cancellation | |
| 5 (optional) | SMS | Day 13 | Final text before cancellation |
The day 5-6 SMS consistently outperforms other placements in recovery tests. It catches the "email avoiders" while the payment failure is still fresh enough for customers to act.
If your grace period is only 7 days, move the SMS up to day 3-4 and skip the second SMS.
TCPA compliance
In the US, sending marketing or non-transactional SMS requires compliance with the Telephone Consumer Protection Act (TCPA). Dunning messages fall into a gray area, and the safest approach is to treat them as requiring explicit opt-in.
What TCPA requires for SMS:
- Prior express consent. The customer must have agreed to receive text messages from you. Ideally, this consent is collected at signup with clear language.
- Opt-out mechanism. Every SMS must include a way to stop receiving texts. The standard is "Reply STOP to unsubscribe."
- Clear identification. The message must identify who's sending it.
- Time restrictions. Don't send between 9 PM and 8 AM in the recipient's local timezone.
Are dunning messages transactional or marketing? Dunning messages about a customer's existing subscription and payment are generally considered transactional, which has a lower consent threshold than marketing. But "generally considered" isn't the same as "definitively protected." The FCC has been expanding its interpretation of what requires consent.
The safe approach:
- Collect SMS consent at signup. A simple checkbox: "I agree to receive account-related text messages."
- Keep records of consent (timestamp, IP, what they agreed to).
- Include opt-out language in every message.
- Honor opt-outs immediately and permanently.
- Restrict sends to 8 AM - 9 PM in the customer's timezone.
If a customer opts out of SMS, your dunning sequence should automatically switch to email-only for that customer. Never send another SMS after an opt-out, even if a future payment fails.
For international customers: GDPR (EU), CASL (Canada), and similar regulations in other countries have their own SMS rules. GDPR requires explicit opt-in for any electronic marketing, including SMS. If you have EU customers, you need consent before sending dunning texts.
SMS message templates
SMS has a 160-character limit for a single segment. Longer messages get split into multiple segments and cost more. Keeping dunning SMS concise isn't just a best practice for the customer experience, it's a cost optimization.
Template 1: Mid-sequence recovery (Day 5-6)
[CompanyName]: Your payment didn't go through.
Update here to keep your account active:
[link]
Reply STOP to opt out
Characters: ~130 (varies with company name and link length). Stays within a single segment. Clear, specific, actionable.
Template 2: Final notice (Day 13)
[CompanyName]: Your account will be canceled
on [date] due to a payment issue.
Update now: [link]
Reply STOP to opt out
Characters: ~135. Adds urgency with a specific date. Still concise.
Template 3: With amount (higher ARPU)
[CompanyName]: We couldn't process your $[amount]
payment. Update your card to avoid cancellation:
[link]
Reply STOP to opt out
Characters: ~140. Including the amount works well for higher-price subscriptions where the customer may not remember what they're paying.
What to avoid in SMS:
- Don't use ALL CAPS. It feels like shouting and can trigger spam filters.
- Don't include multiple links. One link, one action.
- Don't use emoji. Keep the tone professional. This is about their money.
- Don't use URL shorteners from generic services (bit.ly, tinyurl). They look suspicious and some carriers filter them. Use your own branded short domain.
- Don't say "Dear customer" or include formal greetings. Get straight to the point.
Cost analysis
SMS costs more than email, and the cost structure is different. Understanding the economics helps you decide how much SMS to include.
Per-message costs (US):
| Channel | Cost per Message | Cost per 1,000 Messages |
|---|---|---|
| Email (SES) | $0.001-0.005 | $1-5 |
| SMS (domestic) | $0.01-0.05 | $10-50 |
| SMS (international) | $0.05-0.15 | $50-150 |
SMS is roughly 10-50x more expensive than email per message. But cost per message isn't the right metric. Cost per recovered payment is.
Cost per recovery comparison:
| Approach | Cost per 1,000 Sends | Recoveries per 1,000 | Cost per Recovery |
|---|---|---|---|
| $3 | 100-150 | $0.02-0.03 | |
| SMS | $30 | 150-250 | $0.12-0.20 |
| Email + SMS | $20 (blended) | 180-280 | $0.07-0.11 |
Even at 10x the per-message cost, SMS delivers recoveries at a cost per recovery that's trivial compared to the revenue saved. If your average subscription is $50/month and SMS costs $0.15 per message, you need to recover just 1 in 333 sends to break even. The actual recovery rate is 150-250 per 1,000, which is roughly 500x the break-even rate.
Budget framework:
For a SaaS with 500 failed payments per month and a 4-email + 1-SMS sequence:
| Item | Monthly Cost |
|---|---|
| 2,000 emails (4 per failure) | $2-10 |
| 500 SMS (1 per failure) | $5-25 |
| Total | $7-35 |
At $7-35/month for messaging costs, the ROI question isn't whether SMS is worth it. It's why you haven't added it already. Even the most expensive SMS plan is a rounding error compared to the revenue it recovers.
Combining email and SMS: the lift
The lift from adding SMS isn't additive in a simple way. It's not "email recovery + SMS recovery = combined recovery." Some customers respond to both channels. But SMS captures a distinct segment that email alone misses.
Who SMS recovers that email doesn't:
- Email non-openers. About 55-65% of dunning emails go unopened. Many of those recipients will read a text.
- Inbox-zero abandoners. Some people have inboxes with thousands of unread messages. They live on their phone.
- Procrastinators. SMS creates immediate response. "I'll do it now" replaces "I'll do it later."
- Gmail Promotions tab residents. If your dunning emails land in the Promotions tab (despite being transactional), SMS bypasses this entirely.
Measured lift from adding SMS to an email-only sequence:
| Email-only Recovery | Email + SMS Recovery | Lift |
|---|---|---|
| 32% | 42% | +31% relative lift |
| 38% | 48% | +26% relative lift |
| 44% | 54% | +23% relative lift |
The absolute lift is 8-12 percentage points. The relative lift is 23-31%. Both of those numbers are significant enough to justify the minimal cost and implementation effort.
Handling opt-outs
SMS opt-outs in dunning require careful handling because the customer might want to stop getting texts but still wants to keep their subscription.
When someone replies STOP:
- Immediately stop all SMS to that number. No exceptions.
- Continue their dunning sequence via email only.
- Don't add them back to SMS for future failures (unless they explicitly re-consent).
- Log the opt-out with a timestamp for compliance records.
Don't confuse SMS opt-out with service cancellation. A customer who texts STOP doesn't want texts. They may still want your product. Your dunning system needs to distinguish between "unsubscribe from SMS" and "cancel my account."
Opt-out rates for dunning SMS are typically 0.3-0.8% per send, which is low but not zero. The rate increases if you send more than 2 SMS in a sequence, which is another reason to limit SMS to 1-2 messages.
Getting started with SMS dunning
If you're currently running email-only dunning, here's a minimal path to adding SMS.
Step 1: Collect phone numbers. If you don't have customer phone numbers, start collecting them at signup or in your account settings. A simple "Add a phone number for account notifications" field works. You need the phone number and SMS consent before you can send anything.
Step 2: Start with one SMS. Add a single text message at day 5-6 of your sequence. This is the lowest-effort, highest-impact change. Don't try to build a complex multi-SMS flow from day one.
Step 3: Measure the lift. Compare your recovery rate before and after adding SMS. Give it at least 30 days of data. You should see a 15-25% relative increase in total recovery.
Step 4: Optimize placement and add a second SMS if warranted. If the first SMS performs well, test a second one at day 12-13 as a "final notice" before cancellation. Track whether the second SMS adds meaningful recovery or just incremental cost.
SMS in dunning isn't complicated. It's a simple message with a payment link, sent at the right time, to a channel that almost everyone reads. The compliance requirements need attention, but the implementation itself is straightforward. For most SaaS businesses, one well-timed text message adds more recovery than two additional emails would.