Understanding Why a Merchant Might 'Do Not Honor Decline' and What to Do

“Do Not Honor” is the single most frustrating decline code in payments, and it exists as a deliberate black box that hides the real reason a transaction failed. It maps to ISO 8583 response code 05, travels through Data Element 39 of the authorization message, and accounts for 10–60% of all payment declines depending on geography and card network. For subscription businesses, where failed payments drive 20–40% of total churn, understanding this code’s mechanics is the difference between recovered revenue and silent customer loss.

Key Takeaways

• Code 05 is a catch-all: Adyen’s data shows roughly 50% of “Do Not Honor” declines are actually insufficient funds in disguise, while VisaNet actively remaps suspected fraud (code 59) to code 05 before merchants ever see it.

• Every processor surfaces it differently. Stripe exposes the raw network code, Adyen uses ML to infer the real reason, and Square hides the code entirely. Only three major processors give you the raw ISO 8583 response at all.

• Card networks impose hard retry limits with real penalties. Visa allows 15 retries per 30 days ($0.10–$0.25 per excess), Mastercard allows 35 retries per 30 days (fees up to $0.50 each), and Amex has no published framework yet uses code 05 for up to 92% of all declines.

• Subscription MCCs face a 3.5x decline penalty. MCC 5968 (subscription/continuity) has a 259% higher decline rate than MCC 5734 (computer software), driven entirely by issuer risk scoring.

• No platform publicly breaks down recovery rates by specific decline code. This is the biggest data gap in the industry.

What Actually Happens Inside the Payment Chain When Code 05 Fires

When a cardholder’s payment is declined with “Do Not Honor,” the decline originates at the issuing bank’s authorization system and travels back through the card network to the acquirer and finally to the merchant. The technical path is: merchant system → acquirer/processor → card network (VisaNet or Mastercard Banknet) → issuing bank → response back through the same chain. The issuer populates Data Element 39 (DE 39) of the ISO 8583 authorization response message (MTI 0110) with the two-character code “05.”

The information asymmetry is stark. The issuer sees everything: the cardholder’s full transaction history, available balance, risk score, velocity patterns, geographic behaviour, and device fingerprint. The merchant receives only “05: Do Not Honor.” No balance information, no fraud score, no account status. This is by design.

Three factors make code 05 the most common decline. First, many issuers run legacy authorization systems that lack granular reason code mapping, so they default to 05 for anything they can’t categorise cleanly. Second, card networks have historically not enforced specific code usage, though both Visa and Mastercard now penalise issuers whose Category 4/generic declines exceed 5% of total decline volume. Third, and most revealing: VisaNet actively remaps response code 59 (“Suspected Fraud”) to code 05 (“Do Not Honor”) before passing the response to the acquirer. This legacy behaviour exists to prevent confrontational situations at point-of-sale terminals, but it means merchants processing Visa transactions can never distinguish a fraud flag from a genuinely ambiguous decline.

Adyen’s analysis of global payment data confirms that approximately 50% of “Do Not Honor” declines are actually insufficient funds in disguise. The remaining cases split among AVS/CVC mismatches (no standard refusal code exists for AVS failures, so issuers default to 05), suspected fraud (especially on Visa due to the 59→05 remapping), velocity limit triggers, and genuinely unknown factors.

How Code 05 Compares to Nearby ISO 8583 Decline Codes

Compared to nearby ISO 8583 codes, code 05 is deliberately vague. Code 04 (“Pick Up Card”) and code 07 (“Pick Up Card, Special Conditions: Fraud”) are hard declines that should never be retried. Code 51 (“Insufficient Funds”) is specific and actionable. Code 01 (“Refer to Card Issuer”) requests manual voice authorization. Code 05 sits in the middle as a retryable soft decline with no actionable information attached.

How Each Payment Processor Surfaces This Decline

Despite all receiving the same underlying network response code 05, every major processor translates it into its own taxonomy. This matters for developers building retry logic, because without the raw code you cannot distinguish “Do Not Honor” from dozens of other decline reasons at the network level.

Stripe

Stripe returns decline_code: "do_not_honor" within a card_declined error, with outcome.type: "issuer_declined" (an issuer decline) and the raw network code exposed as outcome.network_decline_code: "05". On the charge’s outcome object, Stripe also exposes a network_advice_code field, which is a 2–4 digit code from the issuer offering retry guidance. Stripe classifies all payment failures into three buckets: issuer_declined (where “Do Not Honor” falls), blocked (Stripe Radar or Adaptive Acceptance), and invalid (bad API call). Stripe’s official recommendation for this code is blunt: “The customer needs to contact their card issuer for more information.”

Braintree / PayPal

Braintree uses its own numeric system: processor response code 2000 with text “Do Not Honor,” explicitly classified as processor_response_type: "soft_declined". Braintree exposes the raw network response through network_response_code and network_response_text fields.

Adyen

Adyen takes a different approach entirely, mapping code 05 to refusalReasonCode: 27 with refusalReason: "Declined Non Generic." This naming is intentional. Adyen’s documentation explains that code 27 “maps all those response codes that cannot be reliably mapped,” distinguishing generic declines from more specific ones. Adyen also offers an “Inferred Refusal Reason” feature that uses machine learning across its platform-wide data to identify the likely actual cause behind a “Do Not Honor” (e.g., insufficient funds, suspected fraud, AVS mismatch).

Square and Authorize.net

Square collapses the code into GENERIC_DECLINE within its PAYMENT_METHOD_ERROR category, using descriptive string enums rather than numeric codes. Square does not expose raw ISO 8583 network codes to merchants. Authorize.net is even less granular, returning only Response Code 2 (“Declined”) with Response Reason Code 2 (“This transaction has been declined”), with no “Do Not Honor” sub-code or network code visibility.

Only Stripe, Braintree, and Adyen expose the raw network decline code to merchants. For developers building retry logic, this is a critical distinction.

Card Network Retry Rules: Visa vs. Mastercard vs. Amex

The card networks diverge sharply in how they categorise code 05 and what they allow merchants to do about it.

Visa

Visa classifies response code 05 as Category 4 (Generic Response Codes), permitting up to 15 retry attempts within a rolling 30-day window. The fee structure for violations, enforced since April 2021 with escalation in April 2022, charges $0.10 per domestic excess attempt and $0.25 per cross-border excess attempt beyond the 15th retry. Visa also prohibits acquirers from changing key data elements (MCC, POS environment, electronic commerce indicator) between retry attempts, and violations incur additional fees. Worth noting: Visa pressures issuers to keep Category 4 usage below 5% of total declines, with penalty fees for issuers who exceed this threshold.

Mastercard

Mastercard uses a Merchant Advice Code (MAC) system instead of Visa’s four-category framework. MACs are appended by Mastercard’s network to the issuer’s decline response in DE 48, Subelement 84. For code 05 specifically, the MAC value depends on the underlying reason the issuer declined. Mastercard allows 35 retries within 30 days (more than double Visa’s limit) but imposes escalating fees that have increased aggressively: $0.15 per excess transaction (November 2023), $0.30 (January 2024), and $0.50 (January 2025). Mastercard also prohibits any retry of transactions declined with MAC 03 (“Do Not Try Again”) or MAC 21 (“Payment Canceled”). Since October 2023, Mastercard’s Authorization Optimizer uses 300+ AI-driven features to provide optimal retry timing for recurring transactions, claiming an average 18% increase in conversions for merchants following its guidance.

American Express

American Express operates as a closed-loop network, acting as both card network and issuer, which gives it full visibility into both sides of every transaction. Amex uses standard ISO 8583 response code 05 but does not publish a structured retry fee framework comparable to Visa or Mastercard. In practice, merchants should limit retries to 2–3 attempts. The most striking Amex data point: according to Churnkey’s analysis of $2B+ in subscription revenue, as high as 92% of all Amex declines use “Do Not Honor” as the response code, compared to just 11% for Visa and 4% for Mastercard. Amex relies on code 05 as a catch-all far more heavily than open-loop networks.

The Real Numbers Behind “Do Not Honor” Revenue Loss

Published data on “Do Not Honor” specifically is scarce, as most platforms report aggregate recovery rates across all decline types. That said, several concrete data points emerge from industry research.

Prevalence by Geography and Network

Visa’s own Global Declines Transaction Analysis found that 76% of all global declined transaction volumes are tagged either “Insufficient Funds” or “Do Not Honor” combined. Churnkey’s 2024 analysis provides the most granular public breakdown: “Do Not Honor” is the second most common decline code in the US at approximately 10% of all declines, trailing only “Insufficient Funds.” The rate varies sharply by country. In India, 34% of all declines are “Do Not Honor,” compared to just 5% in Australia and 7% in the UK. Middle Eastern countries see disproportionately high rates, while European countries tend to use more specific decline codes.

Overall Decline Rates and Revenue Impact

Recurring payments face a 15% average decline rate according to Visa and Mastercard data, with some industries reaching 30%. Standard one-time purchases see only 4.5–5% decline rates, meaning recurring billing experiences roughly 3x higher decline rates. For SaaS businesses specifically, an average of 9% of recurring revenue is lost to failed payments annually. The broader false decline problem costs merchants an estimated $443 billion globally per year, with merchants losing 13x more revenue to false declines than to actual fraud.

Recovery Benchmarks Across Dunning Platforms

Recovery rates vary widely depending on the approach and platform. Stripe Smart Retries recovers approximately 57% of failed recurring payments, using ML trained on 500+ attributes across billions of data points, with a default of up to 8 retries within 2 weeks. Churn Buster reports an average 50.3% recovery rate across its customer base, with top performers reaching 94.5% and B2B SaaS companies with low churn typically achieving 70%+. Gravy (human-powered recovery) claims up to 80% recovery, noting that dunning emails alone recover only about 15%. Baremetrics Recover reports 40–60% recovery, with one client reducing involuntary churn from 12% to 2% in three months, recovering $50,000 in ARR. Recurly data from 1,400+ subscription sites shows nearly all industries achieve 40%+ recovery rates, with “Insufficient Funds” having the highest recovery rate and “Do Not Honor” showing moderate recovery within a shorter timeframe than fraud-related declines.

A critical gap in publicly available data: no platform publishes retry success rates broken down by “Do Not Honor” versus other decline codes. All recovery statistics are aggregated across decline types. The closest data point comes from Justt.ai’s 2025 estimate that 60–70% of soft declines (which include “Do Not Honor”) can be recovered through smart retries.

Subscription MCCs Face Much Higher Decline Rates

Merchant Category Code assignment directly impacts “Do Not Honor” decline rates through issuer risk scoring. The most striking data point comes from FlexPay (now Revaly): MCC 5968 (Direct Marketing: Continuity/Subscription) has a 259% higher decline rate than MCC 5734 (Computer Software Stores), representing a 3.5x difference driven entirely by MCC classification.

Issuers assign higher baseline risk scores to MCC 5968 because many merchants with elevated refund and chargeback rates operate under this code. However, merchant-specific performance matters: the New York Times averages a 93.4% approval rate under MCC 5968, well above the category average, demonstrating that issuers calibrate risk scoring over time based on individual merchant behaviour.

How MCCs Trigger “Do Not Honor”

Issuers can set spending controls by MCC, completely blocking transactions from certain business categories. When this happens, the typical response is “Do Not Honor” because the issuer blocks based on category without providing a specific explanation. High-risk MCCs including gambling (7995), adult entertainment (5967), subscription/continuity (5968), nutraceuticals, and debt collection all experience elevated decline rates. Transactions over $100 face more scrutiny from high-risk MCCs. The mechanism is asymmetric: issuer risk systems are quick to react to negative signals and slow to respond to positive data, meaning new merchants in high-risk categories face a sustained period of elevated declines before building issuer trust.

Why Recurring Billing Gets Hit Harder

Recurring billing triggers more “Do Not Honor” declines than one-time purchases for several structural reasons: cards go stale between billing cycles, stored credentials lack CVV for subsequent charges, and the absence of proper recurring transaction indicators in payment data increases issuer suspicion.

What the Best Recovery Teams Actually Do Differently

The expert consensus across dunning platforms converges on a critical principle: front-load silent retries before contacting the customer. Churn Buster’s data shows 21% of payments can be recovered through retries alone before the first email is sent, and 15% of soft declines resolve silently without any customer contact. PYMNTS research found that credit card declines prompt 27% of subscribers to cancel or switch to a competitor, so premature customer communication about a “Do Not Honor” that would have resolved on retry can actually increase churn.

The Optimal Retry Strategy for Code 05

The optimal handling differs meaningfully from other decline types. For “Insufficient Funds” (code 51), retry timing should cluster around paydays on the 1st and 15th of the month. For “Expired Card” (code 54), retries are pointless; immediate customer contact for updated card details is correct. For “Do Not Honor,” the strategy is a hybrid: retry 2–3 times over 5–10 days at optimal times (weekday business hours, beginning of month), then escalate to customer outreach with empathetic, brand-consistent messaging that includes a one-click payment update link requiring no login. This medium-length window sits between the payday-optimised approach for insufficient funds and the immediate customer contact required for hard declines.

Technical Interventions That Reduce “Do Not Honor” Rates

Three technical interventions show measurable impact on “Do Not Honor” rates.

Network tokenisation delivers a 2.1% authorization rate uplift on Mastercard and 4.6% on Visa, with NMI reporting a 95% reduction in declines due to suspected fraud when using tokens. Tokens receive more favourable issuer treatment because they carry cryptographic validation and device-binding assurance, and they automatically update when cards are reissued.

Card updater services (Visa Account Updater, Mastercard Automatic Billing Updater) can reduce card-related involuntary churn by up to 30% by automatically obtaining new card credentials when cards are reissued.

Processor cascading, which involves routing a declined transaction to a different processor or attempting with a network token after a PAN-based decline fails, is recommended by Nuvei and others as a fallback strategy.

Stripe’s Adaptive Acceptance represents the most advanced automated approach, using real-time ML to adjust payment request parameters (formatting, metadata) within milliseconds before returning a response. It increased authorization rates by approximately 2.2% on average (some enterprises see around 4%) and reclaimed $6 billion in false declines across the Stripe network in 2024. The system also includes “Excessive Retry Prevention” that automatically prevents retry penalty fees from Visa and Mastercard.

Wrapping Up

“Do Not Honor” is not a single problem. It is an information vacuum that conceals at least four distinct issues (insufficient funds, suspected fraud, AVS failures, and MCC-based blocking) behind one opaque code. The most actionable insight for any team writing retry logic is that this code should be treated as a soft decline with a medium-length retry window of 5–14 days, positioned between the shorter payday-optimised window appropriate for insufficient funds and the immediate customer contact required for hard declines. Network tokenisation and card updater services are the highest-impact technical investments for reducing “Do Not Honor” volume proactively, while smart retry systems from Stripe, Mastercard, and specialised dunning platforms recover 50–70% of failed payments when properly configured.

Sources

Decline Prevalence and Revenue Impact

• Adyen Knowledge Hub: Do Not Honor: Adyen’s analysis showing ~50% of “Do Not Honor” declines are actually insufficient funds; VisaNet remapping of code 59 to 05. See also Chris Laumans’ breakdown in Finance Monthly.
• ClearSale: Understanding the True Cost of False Declines: $443 billion in global false declines projected by 2021, from the report “Balancing False Declines and Fraud Prevention.”
• Javelin Strategy & Research via FinTech Futures: False declines valued at $118 billion in the US alone, more than 13x total card fraud losses ($9 billion).
• PYMNTS x FlexPay: Subscription Churn from Payment Declines: 27% of subscribers cancel or switch providers after experiencing a payment decline.

MCC Impact on Decline Rates

• FlexPay (Revaly): How Does My MCC Impact My Business?: MCC 5968 has a 259% higher decline rate than MCC 5734, representing a 3.5x difference.

Recovery and Platform Data

• Stripe Blog: AI Enhancements to Adaptive Acceptance: Stripe Adaptive Acceptance recovered $6 billion in false declines across the Stripe network in 2024.