Recovery Stack · 2026 Technical Guide
Stripe's default recovery behavior recovers 30–40% of failed payments. Recurflux's 5-layer stack recovers 65–75%. The gap between those two numbers represents tens of thousands of dollars in annual MRR — and it is closed entirely by architectural decisions about how each layer of the recovery pipeline is built. This guide covers every component, every configuration option, every integration requirement, and the expected recovery impact of each layer — quantified.
Quick answer
The Recurflux recovery stack is a 5-layer technical architecture that recovers 65–75% of failed Stripe payments, versus the 30–40% recovered by Stripe's default behavior. The five layers: decline-code-specific retry logic, multi-channel dunning sequences, proactive card expiry detection, payment intelligence scoring, and cancellation flow integration. Each layer compounds the others — the full stack outperforms any single layer by 3–5×. This guide documents every component, configuration option, and expected recovery contribution.
35%
Gap between default and full-stack recovery
The difference between $22K and $27.5K/month recovered on $100K MRR
5×
Multiplicative advantage of all 5 layers combined
Each layer amplifies every other — compounding, not additive
30 min
Setup time for all 5 layers
No engineering resources, no code changes, no production deploys
Why layers beat point solutions
Why layers beat point solutions
The five layers are not independent — they are multiplicatively interdependent. Each layer feeds intelligence into the next. A point solution recovers a fraction of what a coordinated system recovers.
| Approach | What It Handles | Recovery Rate |
|---|---|---|
| Stripe default retries only | Fixed-interval retries, no code intelligence | 30–40% |
| Smart retries only (no prevention / dunning) | Code-specific retry timing | 45–55% |
| Dunning only (no smart retries) | Customer communication after failure | 35–50% |
| Smart retries + dunning (no prevention / dispute) | Combined retry + communication | 55–65% |
| Full 5-layer Recurflux stack | Prevention + Retry + Dunning + Dispute + Analytics | 65–75% |
The $5,500/month gap
On a $100K MRR SaaS, the difference between 55–65% and 65–75% recovery is the difference between recovering $22,000/month and $27,500/month. That $5,500 monthly gap is created entirely by the analytics and prevention layers — the two layers most founders skip.
Layer 1 · Prevention
Layer 1 · Prevention
Prevention is the highest-ROI layer because it generates zero downstream costs. A failure that never occurs requires no dunning emails, no retry processing fees, no support tickets, no dispute risk. Every failure prevented is pure recovered revenue with no operational overhead.
Recurflux scans your entire active subscription base daily for upcoming card expirations. When a card is within 30 days of expiry, the pre-expiry notification sequence activates automatically — each notification linking directly to the Hosted Payment Portal.
| Notification | Timing | Tone | Message |
|---|---|---|---|
| Reminder | 30 days before expiry | Friendly, low urgency | "Your card expires next month — update to keep service active" |
| Escalation | 15 days before expiry | Clear urgency | "Your card expires in 15 days — one click to update" |
| Final Notice | 7 days before expiry | High urgency | "Action required: card expires in 7 days" |
When banks reissue cards with new numbers or expiration dates, Stripe's CAU automatically refreshes stored credentials before the next charge — preventing the failure silently. Recurflux integrates CAU natively, processing updated card information before every retry attempt.
Visa Account Updater
~80% issuer participation
Mastercard ABU
~70% issuer participation
American Express
Limited coverage
| Prevention Mechanism | Without Recurflux | With Recurflux |
|---|---|---|
| Card Health Monitoring (30/15/7) | Occur at billing cycle | Prevented 30 days early |
| CAU Integration | 100% fail on reissued cards | Auto-resolved silently |
| Combined prevention layer | ~15% prevented | 30% prevented |
Layer 2 · Smart Retry
Layer 2 · Smart Retry
Every failed payment passes through four sequential decision stages before a retry is scheduled. This is where Recurflux's differential over default Stripe behavior is most dramatic — the gap between 30–40% and 55–70% recovery lives entirely in architectural decisions made here.
The moment Stripe fires invoice.payment_failed, Recurflux classifies the decline code into one of four recovery buckets. This classification is adaptive — when a code that is typically Scheduled Retry shows zero recovery across 3+ cycles for a specific customer, the system escalates automatically.
processing_error, issuer_not_available, reenter_transaction
Retry within minutes — 70–85% recovery
insufficient_funds, do_not_honor, card_declined, try_again_later, card_velocity_exceeded
Code-specific cadence — 45–65% recovery
expired_card, incorrect_cvc, incorrect_number, incorrect_zip
Dunning sequence — 35–55% recovery
fraud, stolen_card, lost_card, pickup_card
Generic decline notice — 0–5% recovery
For Scheduled Retry codes, Stage 2 calculates the precise optimal retry moment using decline code baseline cadence, customer historical payment patterns, timezone and bank hours, and card network retry guidelines.
| Decline Code | Stripe Default | Recurflux Smart Retry | Recovery Improvement |
|---|---|---|---|
| insufficient_funds | 24h, 3d, 7d | 24h, payday-aligned, 7d, 14d | +15–25% |
| do_not_honor | 24h, 3d, 7d | 24h, 72h, 7d, 14d | +10–20% |
| processing_error | 24h, 3d | Immediate → 1h → 4h | +25–35% |
| expired_card | 24h, 3d | CAU check → 24h email | +20–30% |
| card_velocity_exceeded | 24h | 24h, 48h + cap check | +10–15% |
| authentication_required | Manual | Auto-trigger 3DS flow | +30–50% |
Positive signals (expedite retry)
Negative signals (defer or cancel retry)
Layer 3 · Dunning
Layer 3 · Dunning
When retries cannot resolve a failure, the dunning layer converts technical billing failures into human recovery conversations. Recurflux operates across four channels simultaneously — email, SMS, in-app banners, and the Hosted Payment Portal.
| Step | Timing | Channels | Tone | Content |
|---|---|---|---|---|
| 1 | Within 24 hours | Friendly | Payment notice + 1-click update link | |
| 2 | Day 3 | Email + In-App | Gentle reminder | Subscription at risk + direct portal link |
| 3 | Day 7 | Email + In-App + Pause Offer | Escalated | Access suspension warning + 30-day pause option |
Doubles the recovery window to 30 days — capturing an additional 15–20% of failures that resolve between Day 7 and Day 30, typically from customers experiencing temporary financial disruption.
| Step | Timing | Channels | Content |
|---|---|---|---|
| 1 | Within 24 hours | Payment notice + update link | |
| 2 | Day 3 | Email + In-App | Subscription at risk + portal |
| 3 | Day 7 | Email + In-App | Access warning + pause offer |
| 4 | Day 14 | Email + SMS | Final payment notice + pause |
| 5 | Day 21 | Email + SMS | Last chance before pause |
| 6 | Day 30 | Paused — reactivation link |
The Subscription Pause Innovation
When dunning exhausts, the default outcome is cancellation — permanently destroying the customer relationship with a 5–15% re-acquisition rate. Recurflux offers a pause instead: 14, 30, or 60 days, account preserved, data accessible, reactivation is one click.
| Outcome | Short-Term Revenue | 12-Month Revenue (at $100 ARPU) |
|---|---|---|
| Cancel | $0 | $0–$150 (5–15% re-acquire) |
| 30-day Pause | $0 (pause month) | $770–$880 (70–80% reactivate) |
For 20 dunning-exhaustion events/month at $100 ARPU, the annual LTV difference between cancel-by-default and pause-by-default is $148,800–$175,200 in preserved revenue.
Layer 4 · Dispute Intel
Layer 4 · Dispute Intel
Chargebacks are not just a revenue problem — they are an existential threat to your Stripe account. Every dispute costs a flat $15 fee, counts toward network monitoring thresholds regardless of outcome, and when thresholds are breached, creates account restriction and termination risk.
$15
Per dispute fee
Regardless of outcome (US)
0.9%
Visa VDMP threshold
Stripe acts at ~0.5–0.7% estimated
1.0%
Mastercard ECP threshold
Breach it → Excessive Chargeback Program
Within minutes of a dispute being filed, Recurflux automatically assembles a complete evidence package from six data sources — formatted to Visa, Mastercard, or Amex submission requirements. Manual evidence assembly takes 2–5 hours per dispute. Recurflux reduces it to seconds.
Email communication logs
Portal visit analytics + login timestamps
Full retry history timeline
Subscription agreement + terms acceptance
Refund / cancellation policy disclosure
Complete billing history
4-signal classification on every dispute before you decide how to respond:
| Signal | Interpretation | Context |
|---|---|---|
| Usage pattern during disputed period | Logged in = fight the dispute | Proves service was actively used |
| Prior cancellation attempt history | Recent attempt = assess carefully | May indicate genuine dissatisfaction |
| Support contact before dispute | No contact = likely friendly fraud | No warning = surprise claim |
| Dispute timing vs. renewal date | Post-renewal spike = renewal surprise | Common friendly fraud pattern |
Layer 5 · Analytics
Layer 5 · Analytics
The analytics layer is not a passive reporting dashboard — it is the feedback loop that continuously improves every preceding layer and provides the unambiguous ROI proof that justifies every dollar of payment recovery investment.
Revenue returned through successful retries and dunning conversions this billing cycle — updated in real time as each retry fires and each dunning email converts.
Active failures currently in the retry or dunning queue — showing exactly how much MRR is still recoverable versus already lost.
On first connection, Recurflux pulls 3 months of failed charges to establish your true recovery baseline before the optimization stack activates.
What your MRR, recovery rate, and LTV would look like without Recurflux — calculated in real time. Closes every internal ROI debate permanently.
The 90-day baseline — why it matters
The 90-day historical sync provides three simultaneous functions: immediate visibility into how much MRR you've lost in the past 3 months (often a number that triggers urgent action), accurate benchmarking so every improvement is measured against your actual historical performance, and ML training data that gives the retry timing models 3 months of your specific customer payment patterns to optimize against from day one.
FAQ
FAQ
Recurflux's Card Health Monitoring scans your entire active subscription base daily and triggers an automated 3-step update sequence when a card is 30, 15, and 7 days from expiry. Each notification links directly to a white-labeled Hosted Payment Portal where customers can update in one click — no login required. The 30-day window is specifically validated: earlier notifications see low engagement because customers defer action, while later notifications leave insufficient time to update before the billing cycle.
Every failed payment is classified into one of four buckets: Immediate Retry (processing errors — retry within minutes, 70–85% recovery), Scheduled Retry (insufficient funds, do_not_honor — code-specific cadences, 45–65%), Customer Action (expired card, incorrect CVC — escalate to dunning, 35–55%), or No Recovery (fraud, stolen card — generic notice, 0–5%). This classification is adaptive: if a Scheduled Retry code shows zero recovery across 3+ billing cycles for a specific customer, the system automatically escalates to Customer Action.
For insufficient_funds declines, Recurflux schedules retry attempts to align with the customer's likely payday — typically the 1st or 15th of the month. The system analyzes each customer's historical payment success dates to identify individual patterns, then schedules retries accordingly. A customer who consistently pays on the 3rd triggers a retry on the 3rd rather than the generic 1st/15th default. This per-customer personalization delivers +15–25% recovery improvement over random-interval retries for insufficient_funds specifically.
The Rise 3-touch sequence covers Days 1, 3, and 7 — recovering ~60–70% of recoverable failures within the first week. The Surge 6-touch sequence extends to Day 30, adding SMS at Day 14, a second SMS at Day 21, and a reactivation email at Day 30. The extended window captures an additional 15–20% of failures that resolve between Day 7 and Day 30, typically from customers experiencing temporary financial disruption. Surge also adds full Email Template Editor with A/B subject line testing and High-Value Account Prioritization.
When dunning exhausts without resolution, the default outcome is cancellation — permanently ending the customer relationship with a 5–15% re-acquisition rate. Recurflux's pause instead preserves the account for 14, 30, or 60 days (depending on plan), keeps data accessible, and makes reactivation one click. The result is 70–80% of paused customers returning. For a SaaS company with 20 dunning-exhaustion events per month at $100 ARPU, the annual LTV difference between cancel-by-default and pause-by-default is $148,800–$175,200 in preserved revenue.
Exceeding card network retry frequency limits can trigger issuer-side blocks on your Stripe Merchant ID — causing increased decline rates for every customer on your account, not just the one whose card triggered the block. Recurflux enforces limits based on code type: aggressive for high-recovery processing errors (immediate/1h/4h), progressive for medium-recovery do_not_honor (24h/72h/7d/14d), and conservative for velocity codes (24–48h). After four unsuccessful retries on any code, the system escalates to dunning — the marginal recovery value of further retries is consistently negative at that point.
The counterfactual dashboard calculates — in real time — what your MRR, recovery rate, and LTV would look like if Recurflux had never been activated. It uses your 90-day historical baseline to establish a pre-Recurflux recovery rate, then projects that rate forward against your actual failure volume. The difference between the projected counterfactual and your actual recovered MRR is your attributable ROI. This closes the "would we have recovered that anyway?" debate for finance teams and investors by using your own pre-activation data as the benchmark.
Checkout Recovery (Surge plan only) targets a different failure mode than subscription dunning — customers who started but abandoned checkout before completing payment. It fires a 3-step automated sequence: a 1-hour "you left something behind" email, a 24-hour feature reminder, and a 72-hour final-chance notice. This operates entirely upstream of the subscription billing system, recovering prospects who showed strong purchase intent but encountered friction at checkout — converting failed trial signups into paying subscribers before they ever enter the subscription failure pipeline.
Run the Numbers
Run the Numbers
The recovery math changes at different MRR levels. Run your numbers before you decide.
ROI Calculator
At $30K MRR with a 4% failure rate, moving from 40% to 65% recovery is roughly $12,000/year.
Put your MRR in. See the exact ARR impact of the complete recovery stack at your scale.
Calculate my ROI →LTV Impact
The stack doesn't just recover a payment. It keeps the customer billing for another 14 months on average.
See how recovery rate improvement cascades into LTV — the number most tools ignore when quoting ROI.
See LTV impact →Churn Diagnosis
Before stacking solutions on top of churn, make sure you're treating the right kind.
Split your churn rate into involuntary vs. voluntary. The recovery stack fixes only one of them.
Split my churn →Connect Stripe in 60 seconds. Get your 90-day audit: MRR leakage, decline code distribution, recovery rate vs. benchmarks, and dispute exposure score.
Related Features
Related Features
Smart payment retry →
Code-specific retry cadences for 30+ decline codes — not generic Stripe retries.
Dunning email sequences →
Adaptive 5-step sequences that vary by decline code, tier, and subscription value.
Card health monitoring →
Pre-expiry scanning at 30/15/7 days to stop failures before they start.
Dispute protection →
Real-time Visa/Mastercard dispute tracking and 1-click evidence export.
Subscription pause →
Offer a pause instead of cancellation — 70–80% of paused customers return.
Win-back sequences →
Automated campaigns to re-engage lapsed subscribers.