.png)
.png)
Stripe Smart Retries is Stripe's built-in feature that automatically re-attempts failed subscription charges using a machine-learning model trained on billions of historical transactions. Instead of retrying on a fixed schedule, it picks the time window most likely to succeed for each individual card.
Stripe publishes that businesses using Stripe Billing recover 55% of failed payments on average. That's the headline figure on stripe.com/billing. But after auditing 200+ Stripe Billing accounts at Redux Payments (all B2C subscriptions, representing over $500M in failed-payment volume), the recovery rate we consistently see lands in the 25–35% range. That's a 20+ point gap from Stripe's headline number, and the gap isn't a mistake. It's structural, and it matters if you're running a B2C subscription on Stripe and benchmarking against the wrong number.
This guide explains how Smart Retries actually works, why the recovery rate B2C operators experience on involuntary failures is materially lower than Stripe's published aggregate, and where Stripe's recovery suite leaves money on the table for consumer subscriptions.
When a subscription charge fails, Smart Retries doesn't immediately try again on a fixed cadence. Instead, the model scores the probability of success for each potential retry window over the next several days, then schedules the retry attempts at the highest-probability times.
Per Stripe's engineering team, the model is trained on "billions of data points" with "more than 500 attributes," using an Auto-ML–based ensemble architecture that replaced their original XGBoost setup in 2023. The signals it incorporates include the card brand, issuing bank, original decline code, the customer's historical payment behavior, and time-of-day patterns from across Stripe's network. The recommended default configuration in Stripe's docs is up to 8 retries within 14 days. When a charge returns a hard-decline code (e.g. incorrect number, lost card, stolen card), retries are still scheduled and the attempt count still increments, but they only execute once Stripe detects a new payment method on the customer.
It's worth being precise about scope: Smart Retries itself is just the ML retry timer. Stripe Billing pairs it with Customer Emails, a sibling feature inside the same revenue-recovery suite that sends failed-payment notifications with a hosted payment-update page where the customer can update their card. When this article cites a 25–35% B2C recovery rate, we mean Stripe's full recovery suite (Smart Retries plus Customer Emails) working together, not Smart Retries in isolation.
Stripe's official line is 55% of failed payments recovered on average, published on stripe.com/billing alongside claims of $8.2B recovered across the platform in 2025 and "$14 recovered for every $1 spent." Stripe also publishes that Smart Retries specifically helps businesses "recover $9 in revenue for every $1 spent on Billing."
Those numbers are real. They're also aggregate. Stripe doesn't segment recovery rates by business type, which means the headline figure blends every Stripe Billing customer, from B2B SaaS with $50K annual contracts and dedicated AP teams to B2C subscriptions charging $14.99 to consumer debit cards. The B2B half of that population pulls the average up substantially.
Independent data from Recurly confirms the gap. In their published benchmarks:
Recurly states it directly: "B2B businesses had higher recovery rates than their B2C counterparts across all decline reasons."
Our internal data, drawn from auditing 200+ Stripe Billing accounts that are exclusively B2C subscriptions and representing over $500M in failed-payment volume, shows the same pattern. The recovery rate we observe across these accounts using Stripe's recovery suite (Smart Retries combined with Customer Emails and the hosted payment-update page) sits at 25–35% of involuntary failures. That sits right alongside Recurly's 34.6% B2C number from a completely independent dataset, which is meaningful corroboration.
The takeaway: if you're running a B2C subscription and you're benchmarking yourself against Stripe's 55%, you're benchmarking against a number that includes the B2B half of Stripe's book. The B2C-relevant baseline is closer to 30%.
The gap isn't random. It's driven by structural differences in how the two populations behave when a card fails:
None of these favor B2C. They're the reason the same Stripe recovery suite produces a 53.5% recovery rate for one cohort and a 34.6% rate for the other.
Stripe's recovery suite is a great baseline, but it's a generalist solution. It treats a $2,000 B2B invoice the same way it treats a $79 B2C subscription. The Smart Retries model is trained on Stripe's entire transaction graph, which means it's optimizing for the average merchant across the platform, not for B2C consumer subscriptions specifically.
The places it consistently underperforms for B2C subscriptions:
By relying solely on "good enough" native recovery, B2C subscription companies unknowingly accept a recovery rate that leaves six or seven figures of revenue on the table every year.
"It's just a $79 fail," he laughed over the Teams call.
Ten seconds later, he wasn't laughing.
The math was simple, but the realization was heavy. At a $79 monthly price point and an average subscriber lifespan of 6.1 months, that single failed payment represented $481.90 in lost LTV. When we looked at the data, they had 800 of those failures last quarter.
He went silent. He scrolled through the audit, then scrolled again. Finally, the question came: "Wait... are we losing over $1M a year?"
The answer was a definitive yes.
That "small" failure is a signal. It's the difference between a healthy, compounding LTV and a leaky bucket that requires you to spend more on acquisition just to stay in the same place.
We often think of retention as something that happens at the end of a customer's journey, on the "cancellation page." In reality, retention starts at the failed payment. If your growth feels stagnant and your revenue feels like it's leaking, it's often because your "plumbing" is broken. You're losing customers who actually wanted to stay, but were forced out by a technical glitch or a poorly timed retry.
The conventional advice from third-party recovery vendors is to keep Smart Retries on and bolt their tool on top. We don't recommend that, and as the company building one, we have a specific reason for the opposite recommendation.
Redux is not a stack-on-top recovery layer. It's a full recovery suite purpose-built for B2C subscriptions, and we recommend turning Smart Retries off when you switch over. Stripe Smart Retries is a generalist ML model trained on every transaction type that flows through Stripe: B2B invoices, B2C subscriptions, e-commerce, marketplaces, all averaged together. Redux is built exclusively for B2C consumer subscriptions: no B2B invoices pulling the model toward enterprise patterns, no aggregate averaging, no generalist assumptions. It passively outperforms Stripe's recovery suite on consumer subscriptions because it doesn't have to optimize for anything else.
The way this typically works in practice:
That last point matters: because Redux replaces Stripe's recovery rather than supplementing it, we have to actually outperform Smart Retries to make money. If we don't beat your benchmarked baseline, you don't pay. (For how the lift number is computed, see our breakdown of incremental lift. For more on why purpose-built B2C recovery beats traditional dunning, see silent recovery vs. traditional dunning.)
The B2C-only design produces concrete differences from Stripe's generalist approach:
Stop benchmarking yourself against the wrong number. If you want to see what your true B2C baseline looks like, and what's recoverable above it, we'll run the audit before you commit to anything.
Start your zero-risk pilot with Redux Payments today
Methodology: The 25–35% recovery rate cited throughout this article is drawn from Redux Payments' internal data across 200+ Stripe Billing accounts, all of which are B2C subscription businesses, representing over $500M in cumulative failed-payment volume. The rate represents the percentage of involuntary payment failures recovered through Stripe's native recovery suite (Smart Retries combined with Customer Emails and the hosted payment-update page) measured as a baseline before customers switch to Redux's B2C-only recovery suite. Recovery-rate calculation window: rolling 12 months ending Q1 2026.
Primary sources cited:
.png)
.svg)
.svg)
