When Dunning Is the Wrong Metric

What Dunning Actually Measures

Dunning recovery is a lagging indicator. It counts attempts to fix a payment that already failed. A high dunning recovery rate (60%+) feels like operational success - the system caught the problem and fixed it. But that framing inverts the real question: why did the payment fail in the first place?

For Shopify DTC brands, payment failures fall into three categories: transient (temporary card block, network timeout), customer-initiated (card expired, customer canceled), and systemic (billing address mismatch, processor rules, fraud filter). Dunning can recover transient and some customer-initiated failures. It cannot fix systemic problems.

The metric becomes misleading when teams optimize for dunning recovery without measuring failure rate. A brand might achieve 50% dunning recovery while failure rate climbs from 3% to 8% - a net loss masked by a single metric.

Failure Rate vs. Recovery Rate

Operators must track both. Failure rate is the percentage of all billing attempts that fail on first try. Recovery rate is the percentage of those failures that dunning converts back to successful charges.

Example: 10,000 monthly billing attempts. 300 fail on first try (3% failure rate). Dunning recovers 150 of those (50% recovery rate). Net result: 150 recovered, 150 permanently lost.

If failure rate rises to 5% (500 failures) but recovery stays at 50% (250 recovered), total revenue loss doubles even though recovery rate is unchanged. Teams optimizing only recovery rate miss this shift entirely.

Failure rate threshold: 2-3% is healthy for most DTC recurring billing. Above 5% indicates processor, card, or customer data problems.
Recovery rate threshold: 30-40% is realistic. Above 50% suggests either excellent dunning strategy or that failures are mostly transient.
Track both weekly. Plot failure rate and recovery rate on separate axes to spot divergence.
Set alerts: if failure rate rises 1% week-over-week, investigate before dunning volume increases.

When High Dunning Volume Is a Red Flag

Dunning volume (total failed charges per month) is often invisible in dashboards focused on recovery rate. A brand might celebrate 45% recovery while dunning volume grows 20% month-over-month.

This pattern signals one of three problems: (1) card data decay - customers' cards are expiring or being replaced faster than expected; (2) processor rule changes - the payment processor tightened fraud filters or 3D Secure requirements; (3) customer acquisition quality - new cohorts have higher churn or invalid payment methods.

The correct response is not better dunning. It's to identify which problem and fix it upstream. Dunning is a band-aid. If the wound is growing, the band-aid is the wrong metric.

Diagnostic: Isolate Failure Causes

Before optimizing dunning, categorize failures by decline reason. Most payment processors return a decline code (e.g., 'insufficient funds', 'card expired', 'lost/stolen card', 'processor error').

Pull 30 days of failed charges. Bucket by decline code. Calculate failure rate and dunning recovery rate for each bucket separately.

If 60%+ of failures are 'card expired', the problem is card data refresh, not dunning strategy. If 40%+ are 'processor error' or 'network timeout', the problem is processor reliability or retry logic, not customer payment behavior. If 50%+ are 'insufficient funds', the problem is customer selection or economic downturn, not billing mechanics.

Card expired / invalid card: Fix with card update flows and proactive card refresh. Dunning recovery on these is low (10-20%).
Insufficient funds: Dunning can help (retry after 3-5 days). Recovery rate typically 25-35%.
Processor error / timeout: Fix with retry logic and processor redundancy. Dunning recovery on these is high (60-80%) because the failure is transient.
Lost/stolen card: Dunning recovery is near zero. Focus on customer communication and payment method updates.

The Dunning Trap: Diminishing Returns

Aggressive dunning (10+ retry attempts over 30 days) can recover an additional 5-10% of failures. But each retry attempt increases customer friction and churn risk. A customer who receives 8 dunning emails is more likely to cancel the subscription than to update their card.

Measure net revenue impact, not recovery rate. If dunning recovers $10,000 but causes 50 cancellations worth $15,000 in lifetime value, the strategy is net negative.

Threshold: if dunning recovery rate is below 25%, or if failure rate is above 5%, stop optimizing dunning strategy and fix the upstream problem instead.

When to Ignore Dunning Metrics

Dunning metrics should be ignored entirely in three scenarios. First: during payment processor migrations or changes. Failure rates and recovery rates will shift; historical benchmarks are useless. Focus on absolute failure rate and decline code distribution instead.

Second: after major customer acquisition campaigns. New cohorts often have different card validity rates and churn patterns. Dunning recovery rates from the previous cohort don't apply. Wait 60 days for the cohort to stabilize before benchmarking.

Third: when failure rate is rising faster than recovery rate can offset. If failure rate is climbing 2% per month and recovery rate is flat, dunning is not the bottleneck. The billing system, processor, or customer base is the problem.

Operator Checklist

Use this checklist to determine whether dunning is the right metric to optimize or a symptom of a deeper problem.

Is failure rate stable (2-3%) or rising? If rising, fix the root cause before tuning dunning.
Have you categorized failures by decline code? If not, you cannot diagnose the problem.
Is recovery rate above 30%? If below 30%, dunning strategy is not the constraint.
Are you measuring net revenue impact (recovery minus churn cost) or just recovery rate? If just recovery rate, you're missing the full picture.
Has failure rate or recovery rate changed in the last 30 days? If yes, investigate processor, card data, or cohort changes before optimizing dunning.
Is dunning volume (total failed charges) growing faster than subscription base? If yes, failure rate is rising and dunning is a symptom.

Questions

FAQ

What's a good dunning recovery rate for a Shopify DTC subscription brand?

30-40% is realistic and healthy. Recovery rates above 50% suggest either exceptional dunning strategy or that most failures are transient (processor errors, network timeouts). Rates below 20% indicate that failures are mostly permanent (expired cards, insufficient funds, customer cancellation) and dunning is not the solution.

Should we send more dunning emails to improve recovery?

Only if failure rate is stable and recovery rate is below 30%. If either metric is moving in the wrong direction, more emails will increase churn without recovering enough revenue to offset it. Measure net revenue impact (recovery value minus churn cost) before scaling dunning volume.

How do we know if a payment failure is transient or permanent?

The decline code from the payment processor tells you. 'Processor error', 'network timeout', and 'try again later' are transient - dunning can recover these. 'Card expired', 'invalid card', 'insufficient funds', and 'lost/stolen card' are permanent or customer-initiated - dunning recovery is low. Pull your processor's decline code report and bucket failures by code.

When should we stop optimizing dunning?

When failure rate rises above 5%, when recovery rate falls below 25%, or when dunning volume is growing faster than your subscription base. At those thresholds, dunning is a symptom of a broken billing system, not a recoverable problem. Investigate processor changes, card data quality, customer acquisition cohort quality, or fraud filter rules instead.

Want this on your account?

Thirty minutes. Bring the number that keeps you up.

Book a demo