Failed Payment Recovery Benchmarks

The published numbers on failed payments, involuntary churn, and recovery — every figure cited.

This page assembles the benchmark data that actually exists in public: research published by Recurly, and Churnkey's analysis of Stripe's subscription data. Swipe Again is new and has no first-party benchmark data yet — when we do, it will be published here alongside these, labeled as ours. Until then, we'd rather hand you the industry's numbers honestly than invent our own.

How big the leak is.

Recurly estimates the subscription industry left $129 billion on the table from involuntary churn in 2025, against a projected $1.5 trillion market — an estimate Recurly derived from billions of data points across its billing network (Recurly press release, January 2025). Recurly publishes the working range most operators plan against: 20 to 40% of all customer churn is involuntary — subscribers lost to payment failures rather than decisions (Recurly, churn management resource).

The recovery side is measurable too. In its 2026 State of Subscriptions report (2,200 subscription businesses, 76 million subscribers), Recurly reports the software industry reclaimed over $155 million through recovery tools in 2025, and digital media nearly $100 million (Recurly, State of Subscriptions 2026).

Involuntary churn as a share of all churn.

Churnkey partnered with Stripe to analyze subscription transaction data — Stripe processed $1.4 trillion across 200 million subscriptions in 2024 — and published the split of voluntary versus involuntary churn by industry (Churnkey, Involuntary Churn Benchmarks; underlying annual churn rates from Stripe):

The same analysis splits B2C from B2B: B2C businesses lose 24% of their churn to payment failures, B2B businesses 16%. And the lower the price point, the more of the churn is involuntary — at an average order value under $10, 35% of all churn is payment failure; between $100 and $1,000, it is 19%.

Why payments actually fail.

Across 5.4 million failed payments in Churnkey's dataset, one cause dominates — and it is a timing problem, not a lost customer (Churnkey):

Read the mix the way an operator should: the largest bucket (insufficient funds) is usually recoverable with timing; the do-not-retry buckets should never be retried at all; and expired cards need a new card, not another attempt. A single fixed retry schedule treats all three the same. Our decline codes guide works through the handling code by code.

What recovery actually recovers.

In Churnkey's published comparison, Stripe's Smart Retries alone recovered 51% of failed payments in their sample, at an average of 5.5 days to recovery. Layering decline-aware precision retries on top recovered 55% and recovered 0.7 days faster. Two lessons in that: the processor baseline is genuinely strong — roughly half of failures are recoverable by retries alone — and the layer above it earns its keep on the remaining half, where decline-specific handling and outreach decide the outcome (Churnkey).

Outreach adds incrementally on top of retries: in the same dataset, each dunning email in a sequence contributed roughly 1–2 percentage points of additional recovery, with the first email the strongest (2.8% recovery, 55.8% open rate) and later emails decaying. Recurly, for its part, publishes that its dunning campaigns recover an industry-leading 25% of at-risk revenue, and that its revenue optimization tooling recovers an average of 70% of failed subscription renewals — vendor-published figures, stated here as theirs (Recurly).

Benchmark your own store against these.

Three numbers tell you where you stand. Your involuntary share: of the subscribers you lost last quarter, how many left over a failed payment rather than a cancel click? If it's in the 20–40% band, you're typical — and that slice is the recoverable one. Your recovery rate: of the payments that failed, what share did you eventually collect? The published retry-only baseline is roughly half. Your failure mix: if insufficient funds dominates, timing is your lever; if expired and replaced cards dominate, card updates and outreach are.

Put your own volume through the math with our failed payment recovery calculator — you set the recovery-rate assumption, the formula is printed on the page, and nothing is claimed for you. And if you connect Swipe Again, measurement is the product: every recovered dollar is attributed to the attempt or message that won it, under the counting rules on how we count.

Frequently asked questions

Where does this data come from? Two published sources: Recurly's research and press materials (their billing network and State of Subscriptions report), and Churnkey's involuntary churn benchmark analysis, which combines Stripe's subscription data — $1.4 trillion processed across 200 million subscriptions in 2024 — with 5.4 million failed payments from Churnkey's own dataset. Every figure on this page links to its source.

Why are there no Swipe Again numbers? Because we don't have them yet. Swipe Again is a new product; publishing invented or modeled "results" would be exactly the kind of claim this page exists to avoid. When our own recovery data is large enough to be meaningful, it will appear here, labeled as first-party.

How current are these benchmarks? The industry and decline-mix tables are from Churnkey's analysis of 2024 Stripe data, published in 2025. The $129 billion estimate is Recurly's, published January 2025 for the year 2025. The reclaimed-revenue figures are from Recurly's 2026 State of Subscriptions. This page is reviewed when the sources publish updates.

What's a good recovery rate? Against the published baseline — roughly half of failed payments recovered by processor retries alone — a dedicated recovery system should be judged on what it adds on the harder half: decline-specific scheduling, card updates, and multi-channel outreach. Judge any vendor, including us, on recovered dollars attributed in a trial, not on a rate quoted in marketing.