About 9% of recurring payment attempts fail. That's the number that shows up consistently across Recurly, Stripe, and Baremetrics reports. If you haven't calculated what that means for your specific MRR, you're probably underestimating the damage.
Failed payments don't just cost you one month of revenue. They compound. A customer lost to a failed payment in January isn't just $50 lost in January. It's $50 lost every month for as long as that customer would have stayed. And the longer you wait to address it, the harder recovery becomes.
The basic math
Start with your MRR. Multiply by your failure rate. That's your monthly exposure.
Monthly revenue at risk = MRR x failure rate
| Your MRR | At 5% failure rate | At 9% failure rate | At 12% failure rate |
|---|---|---|---|
| $10,000 | $500 | $900 | $1,200 |
| $25,000 | $1,250 | $2,250 | $3,000 |
| $50,000 | $2,500 | $4,500 | $6,000 |
| $100,000 | $5,000 | $9,000 | $12,000 |
| $250,000 | $12,500 | $22,500 | $30,000 |
| $500,000 | $25,000 | $45,000 | $60,000 |
If your MRR is $50K and your failure rate is 9%, you have $4,500 at risk every single month. Not some months. Every month.
But "at risk" doesn't mean "lost." Some of those payments recover naturally through automatic retries. Some customers notice and update their card on their own. Without any dunning system, roughly 15-20% of failed payments resolve on their own.
That still leaves 80-85% unrecovered without intervention. At $50K MRR:
Monthly loss without dunning = $4,500 x 0.80 = $3,600
$3,600/month. $43,200/year. From a problem most teams haven't sized.
The compounding effect
Lost revenue from failed payments doesn't sit still. It compounds because of customer lifetime value.
A customer paying $100/month with an average lifespan of 18 months represents $1,800 in LTV. If that customer churns involuntarily in month 6, you didn't lose $100. You lost the remaining $1,200 they would have paid over the next 12 months.
Here's what compounding looks like for a SaaS at $50K MRR with a 9% failure rate and no dunning:
| Month | Monthly MRR Lost | Cumulative Revenue Lost | LTV Lost (18-mo avg lifespan) |
|---|---|---|---|
| 1 | $3,600 | $3,600 | $64,800 |
| 3 | $3,600 | $10,800 | $194,400 |
| 6 | $3,600 | $21,600 | $388,800 |
| 12 | $3,600 | $43,200 | $777,600 |
The LTV column is the one that should get your attention. $777,600 in lifetime value lost over a year, from a $50K MRR business. That's not a rounding error. That's a funding round's worth of revenue walking out the door.
Now, these are simplified calculations. Not every failed payment leads to permanent churn, and LTV assumptions vary. But even if you cut these numbers in half to be conservative, you're still looking at $380K in lost lifetime value annually.
What your failure rate actually looks like
The 9% average hides significant variation. Your specific failure rate depends on several factors.
Payment method mix. Credit cards fail more than debit cards in some regions, less in others. ACH/bank transfers have different failure profiles entirely. If a large portion of your customers pay with prepaid cards, expect higher failure rates.
Customer geography. International payments fail at higher rates due to cross-border processing, currency conversion issues, and varying bank fraud policies. US domestic transactions typically see 5-8% failure rates. Cross-border can hit 10-15%.
Average transaction size. Higher charges trigger more fraud checks and run into transaction limits more frequently. A $500 monthly charge fails more often than a $50 one.
Industry vertical. B2B SaaS sees 3-5% failure rates because customers tend to use corporate cards with higher limits and better card updater coverage. B2C subscriptions run higher at 7-12%.
Billing frequency. Annual plans have fewer failure opportunities (one charge per year) but each failure is more significant. Monthly plans accumulate small losses consistently.
To find your actual failure rate, pull your data from Stripe. Go to your Dashboard, look at payments over the last 90 days, and calculate: (failed charges / total charge attempts) x 100. If you're using Stripe's automatic retries, make sure you're counting unique invoice failures, not individual retry attempts.
The recovery curve
Failed payments don't all behave the same way over time. There's a recovery curve, and it drops fast.
| Time after failure | Cumulative recovery (with dunning) | Cumulative recovery (without dunning) |
|---|---|---|
| 1 hour | 5-10% | 5-10% |
| 24 hours | 25-35% | 10-15% |
| 3 days | 40-50% | 12-17% |
| 7 days | 55-65% | 14-18% |
| 14 days | 70-80% | 15-20% |
| 30 days | 75-82% | 15-20% |
Two things stand out here. First, the gap between dunning and no dunning is enormous. Active recovery triples or quadruples your recovery rate. Second, most recovery happens in the first week. After day 7, the incremental gains slow significantly.
This is why timing matters so much. The first email sent within hours of a failure captures the customer while the failed charge is still fresh. By day 14, many customers have mentally moved on.
Building your own cost model
Here's a formula you can apply to your business:
Annual cost of failed payments =
MRR
x failure rate
x (1 - natural recovery rate)
x 12
x average remaining customer lifespan (in months)
Plug in your numbers:
- MRR: Your current monthly recurring revenue
- Failure rate: Your actual rate from Stripe, or use 7% as a starting estimate
- Natural recovery rate: 15-20% if you have no dunning system
- Average remaining lifespan: How many more months the average customer stays. For most SaaS, 12-24 months is reasonable.
Example with $75K MRR:
$75,000 x 0.07 x 0.82 x 12 x 14 = $724,248
That's $724K in lifetime value at risk annually. Even recovering half of it with a basic dunning sequence would save $362K.
What dunning actually recovers
Let's be specific about what a dunning system changes in these calculations.
Without dunning, your recovery rate sits at 15-20%. With a well-configured email and SMS dunning sequence, recovery rates jump to 60-80%. The exact number depends on your audience, pricing, and sequence quality, but these ranges are well-established across Recurly and Baremetrics benchmark data.
Here's the same $50K MRR example with dunning:
| Metric | No dunning | Email-only dunning | Email + SMS dunning |
|---|---|---|---|
| Monthly failures | $4,500 | $4,500 | $4,500 |
| Recovery rate | 18% | 55% | 72% |
| Monthly recovered | $810 | $2,475 | $3,240 |
| Monthly lost | $3,690 | $2,025 | $1,260 |
| Annual lost | $44,280 | $24,300 | $15,120 |
| Annual savings vs no dunning | — | $19,980 | $29,160 |
Email-only dunning saves roughly $20K/year. Adding SMS on top of email saves another $9K. For a $50K MRR business, that's meaningful revenue recovered with minimal effort.
The costs people forget
Revenue loss from failed payments goes beyond the direct MRR hit. There are secondary costs that compound the problem.
Acquisition cost wasted. You paid to acquire that customer through marketing, sales, or both. If your CAC is $200 and you lose 30 customers per month to failed payments, that's $6,000 in acquisition spend thrown away monthly.
Support burden. Customers whose payments fail generate support tickets. "Why was my account suspended?" "I didn't cancel, what happened?" Each ticket costs $5-15 to handle depending on your support structure. More importantly, it creates a negative experience that increases voluntary churn risk even after the payment is resolved.
Brand damage. When a customer's service gets interrupted because their payment failed silently and nobody told them, they don't blame their bank. They blame you. That frustration shows up in reviews, NPS scores, and word-of-mouth.
Cohort data pollution. Failed payments create messy data. Customers who churned involuntarily look identical to customers who left because they hated your product. This distorts your retention metrics and can lead you to solve the wrong problems.
Revenue recovery ROI
Most dunning tools cost between $50-300/month depending on your volume. Even the simplest ROI calculation makes the case.
For a $25K MRR business:
- Monthly failures at 7%: $1,750
- Recovery with dunning (65%): $1,138 recovered
- Recovery without (18%): $315 recovered
- Net monthly gain from dunning: $823
- Annual gain: $9,876
- Cost of dunning tool: $600-1,200/year
- ROI: 800-1,600%
For a $100K MRR business:
- Monthly failures at 7%: $7,000
- Recovery with dunning (65%): $4,550 recovered
- Recovery without (18%): $1,260 recovered
- Net monthly gain from dunning: $3,290
- Annual gain: $39,480
- Cost of dunning tool: $1,200-3,600/year
- ROI: 1,000-3,200%
There aren't many investments in a SaaS business that return 10-30x. Payment recovery is one of them.
What to do with this information
Three steps, in order of effort:
1. Measure your actual exposure. Pull your failure rate from Stripe. Multiply by MRR. See the number. This takes 10 minutes and often surprises people.
2. Calculate what dunning would recover. Use the 60-70% recovery benchmark for email + SMS dunning. Subtract your current natural recovery rate. Multiply by your monthly failure volume. That's your recoverable revenue.
3. Compare the cost of doing nothing. Add up 12 months of lost revenue. Factor in LTV. Compare that to the cost of implementing a dunning sequence. The math is rarely close.
Failed payments are a solved problem. The data on recovery rates is well-established, the tools exist, and the ROI is clear. The only real cost is the cost of not acting on it.