Declined transactions are the most underreported revenue leak in e-commerce. Unlike a failed marketing campaign, which shows up clearly in acquisition metrics, a declined payment often disappears from reporting as a simple “failed transaction” — without exposing which part of the infrastructure caused it, whether it was recoverable, and how much revenue was permanently lost as a result.
For businesses processing at scale, the financial stakes are significant. A merchant processing €5 million per month at a 92% authorisation rate is losing €400,000 monthly to declines. Improving that rate to 95%, a realistic target with the right infrastructure, would recover €150,000 per month from existing traffic.
Understanding the Decline Taxonomy
Not all declines are equal, and not all are recoverable. The distinction matters because the correct response varies by type.
| Decline Type | Definition and Correct Response |
| Hard decline | Permanent rejection from the issuing bank: card cancelled, account closed, or flagged for fraud. Do not retry. Request an alternative payment method from the customer. |
| Soft decline | Temporary rejection that may resolve on retry: insufficient funds, transaction timeout, temporary issuer block. Retry with appropriate spacing; some issuers specify retry windows in the decline code. |
| False decline | Transaction blocked by the merchant’s or acquirer’s fraud system, despite being legitimate. The customer experiences a decline with no fault of their own. Requires calibration of the fraud rule to reduce frequency. |
| Issuer-side timeout | The issuing bank did not respond within the processing window. Often retryable; root cause may be routing inefficiency rather than a genuine card issue. |
| 3DS authentication failure | Customer failed or abandoned the authentication step. May indicate friction in the 3DS flow rather than genuine fraud — review frictionless rate and step-up challenge design. |
The Root Causes of Preventable Declines
Suboptimal Routing
Transaction routing determines which acquirer processes a given payment. Each acquirer has different relationships with issuing banks, different approval rates for specific BIN ranges, and different performance characteristics by geography. Routing every transaction through a single acquirer — regardless of card origin or type — sacrifices approval rates that intelligent routing would recover.
Missing or Incomplete Transaction Data
Issuers make authorisation decisions based on the data they receive with the transaction request. A request that includes device fingerprint, billing address, transaction history context, and 3DS data is treated very differently from a minimal request containing only card number, expiry, and amount. Enriched transaction data directly correlates with higher approval rates because it provides the issuer’s fraud system with more signals to distinguish legitimate transactions from fraudulent ones.
Miscalibrated Fraud Rules
Fraud prevention tools use rules and machine learning models to flag suspicious transactions. When these tools are configured too aggressively, for example, blocking all transactions from a specific country or all transactions above a certain value, they create false declines. A customer in Germany attempting a €300 purchase should not be blocked because someone in another geography attempted a fraudulent transaction of a similar value. Rule calibration requires ongoing attention, particularly as transaction patterns change with new geographies or product types.
No Retry Logic for Soft Declines
Many businesses treat a decline as terminal. In practice, a soft decline on the first attempt may resolve on retry, particularly for transactions declined due to temporary issuer timeouts or transient system issues. A structured retry strategy with appropriate timing intervals, different routing on retry, and clear logic for when not to retry can recover a meaningful share of soft-declined transactions.
Building a Decline Recovery Framework
An effective decline recovery framework has three components:
1. Real-Time Decline Categorisation
Every declined transaction should be categorised at the point of decline, not retrospectively in a weekly report. Real-time categorisation enables immediate retry decisions, prompts for alternative payment methods, and live routing adjustments.
2. Intelligent Retry Infrastructure
Retry logic should account for: the specific decline code (not all soft declines benefit from immediate retry), the routing path used on the original attempt (retry through a different acquirer when possible), and the time elapsed (some issuers impose retry cooldown windows). Sending the same transaction through the same route within seconds of a soft decline typically does not improve outcomes and may harden the decline.
3. Fraud Rule Audit and Calibration
Fraud rules should be reviewed regularly against false-decline-rate data. When a rule is blocking transactions that are subsequently verified as legitimate (for example, through manual review of declined orders that customer service follows up on), the rule needs recalibration. The goal is not to minimise fraud blocks, but to maintain an acceptable fraud rate while minimising false positives.
What Good Looks Like: Benchmark Targets by Segment
| Business Segment | Target Authorisation Rate Range |
| Digital goods / low average order value | 96–98% |
| General e-commerce, mid-ticket | 93–96% |
| High-ticket or luxury goods | 90–94% |
| Subscription / recurring billing | 88–93% (initial charge); 85–90% (renewal) |
| Cross-border transactions | 85–92% (highly geography-dependent) |
EVOXO provides merchants with granular decline analytics, intelligent multi-path routing, and configurable retry logic, giving operations and finance teams the tools to understand and improve payment performance at the transaction level.
| Key Takeaways |
| Authorisation rate is the primary payment performance KPI. 1% improvement on significant volume is material revenue. |
| Distinguish hard declines (do not retry), soft declines (structured retry), and false declines (rule calibration needed). |
| Enriched transaction data submission to issuers directly improves approval rates. |
| Intelligent retry logic, different acquirer, appropriate timing, recovers a meaningful share of soft declines. |
| Fraud rule calibration requires ongoing attention; benchmark false decline rate as a key metric. |
