Skip to content

Payments Economics 10 min read

What an Auth Rate Point Is Worth: Authorisation Optimisation Economics

One auth rate point is a revenue number, not an engineering metric. How to calculate what it's worth, what drives declines, and which interventions move it.

By Shaun Toh Published Updated
TL;DR

Calculate the illustrative gross value of one percentage point of authorization improvement, then separate approvals from completed orders, contribution margin and retry costs.

The authorisation rate is the most undervalued number in most payments reporting dashboards. It sits next to the MDR line items, often reported by the PSP as a single percentage, and receives a fraction of the optimisation attention that goes into conversion rate, chargeback rate, or processing cost. This is a mistake — and it has a precise financial cost.

An approval improvement can have commercial value, but a declined attempt is not necessarily a lost unique order. The customer may succeed on another method or later attempt. Estimate incremental completed sales and contribution after fulfillment, fees, refunds, fraud losses and implementation costs.

The Revenue Calculation

To calculate what one percentage point of auth rate improvement is worth, you need three numbers:

  1. Monthly transaction volume (attempts, not approvals)
  2. Current authorisation rate
  3. Average transaction value

Formula: 1% × monthly transaction attempts × average transaction value = illustrative additional approved value

Hypothetical examples by scale:

Monthly attemptsAverage order valueOne-percentage-point improvement in approved value
1,000$80$800/month
5,000$80$4,000/month
20,000$120$24,000/month
100,000$60$60,000/month

At 20,000 monthly transactions and $120 average order value, moving from 88% to 90% auth rate is worth $48,000/month — $576,000 annually. That number is comparable to significant engineering investments or major PSP contract renegotiations. It rarely gets comparable management attention.

The arithmetic below holds average attempted value and traffic constant. It estimates additional approved value, not net revenue or profit. Deduplicate orders and check capture and fulfillment before assigning an economic benefit.

The Anatomy of a Decline

Not all declines are equal. Understanding the decline reason code is the prerequisite for any optimisation.

Hard declines are issuer determinations that will not change on retry:

  • 14 Invalid Card Number — the card number failed the Luhn check or does not exist.
  • 41 Lost Card / 43 Stolen Card — the cardholder has reported the card. Do not retry the same reported lost/stolen credential.
  • 54 Expired Card — the card expiry date has passed.
  • 57 Transaction Not Permitted to Cardholder — the cardholder's account restrictions prohibit this transaction type.

Retrying hard declines wastes attempts. For expired or invalid cards, seek updated credentials. For generic 05 Do Not Honour, the cause and retry eligibility are unknown from the code alone: check the PSP's detailed response and current network restrictions, then seek issuer assistance or another payment method when appropriate.

Soft declines are potentially recoverable:

  • 51 Insufficient Funds — the cardholder's available balance is insufficient at this moment. May resolve with a retry on a different day.
  • 61 Exceeds Withdrawal Amount Limit — temporary limit on the account; may resolve.
  • 65 Exceeds Withdrawal Frequency Limit — frequency-based limit; wait before retrying.
  • 91 Issuer Unavailable — technical issue on the issuer side. Check provider retry advice and payment state before another attempt.

Classifying declines helps choose a permitted next action. Its economic effect must be measured on your own eligible cohort rather than assigned a fixed percentage-point uplift.

Four Areas to Investigate

Beyond decline codes, investigate these four areas against your own outcomes:

1. Network Tokens

Network tokens are a credential-management mechanism. Confirm supported lifecycle updates and issuer coverage with the provider rather than assuming every replacement card remains usable.

Test network tokens against a comparable credential cohort and measure approval, capture and completed-order outcomes. No generic uplift is established here.

Prioritise token implementation using your own credential failures, supported issuer coverage, integration cost and measured outcome.

2. Card Account Updater

Card Account Updater (CAU, also called Visa VAU or Mastercard ABU) is a network service that proactively updates stored card details when a card is reissued — new expiry, new account number after cancellation and reissue. For subscription businesses, it reduces declines driven by 54 expired card and account-change scenarios.

Confirm account-updater coverage, event semantics and commercial terms with your provider. An updated credential does not guarantee approval. The credential lifecycle reference explains operational checks.

3. 3DS2 Frictionless Flows

3DS2 supplies authentication information. Check applicable scheme liability conditions separately; authentication is not an approval guarantee.

Track authentication attempts, frictionless outcomes, challenge completion, authorization and final completed orders separately.

Evaluate supported authentication and exemption paths under applicable rules. Merchant-initiated transactions and exemptions are not interchangeable categories; use the SCA strategy reference for that distinction.

4. Merchant Category Code and Billing Descriptor

The MCC (Merchant Category Code) and billing descriptor (what appears on the cardholder's statement) affect issuer approval models and cardholder dispute rates. Issuers use MCC-level risk signals to set approval thresholds — a merchant on the wrong MCC may face systematically lower approval rates because the issuer's model is calibrated for a riskier category.

Check that the billing descriptor identifies the merchant recognisably. Measure unrecognised-charge contacts and disputes before and after a change; no generic reduction or resulting approval uplift is assumed.

Multi-Acquirer Routing as an Auth Rate Lever

Evaluate multi-acquirer routing against integration and reconciliation costs, eligible traffic and a controlled test; there is no universal volume threshold here.

The mechanism: different acquirers have different approval relationships with issuers. Transactions declined by one acquirer/issuer pair may be approved when routed through a different acquirer — because the acquiring bank has different risk-model history with that issuing bank, or because the acquirer's BIN (Bank Identification Number) is associated with different risk signals in the issuer's model.

Compare routing outcomes on matched issuer-country cohorts and include duplicate prevention, fees and the effect on final unique-order completion.

Subscription-Specific Retry Logic

Subscription rebilling creates a distinct optimisation problem. When a subscription renewal attempt is declined, the revenue is not lost immediately — it may be recoverable through a retry at the right time with the right frequency.

Smart retry frameworks:

  • On potentially recoverable declines: Follow the provider's current retry guidance and network limits; a balance shortfall may resolve after a customer funding event. Do not apply one delay to all soft declines or treat generic 05 as a timed retry signal.
  • On 65 frequency limit: Check the provider response and applicable retry rules; the code does not establish a universal waiting period.
  • On 91 issuer unavailable: Confirm payment state and provider retry guidance; prevent duplicate processing.
  • Never: Retry hard declines on the same card without a card update event.
  • Cadence: Set permitted retry limits and customer communications for the provider and payment method; measure recovery and stop conditions.

Measure eligible declined unique orders, subsequent approved attempts and final captured orders separately. Recovery percentages depend on the denominator and observation window.

The retry mechanics above are the summary version. The full decisioning framework — decline-code-specific retry logic, Visa's scheme-level retry caps and non-compliance fees, how to evaluate a PSP's "smart retry" claim, what to change between attempts besides timing, and how to measure recovery rate against the right denominator — is covered in Failed Payment Recovery: Retry Decisioning After a Declined Charge.

The Measurement Framework

To measure your auth rate correctly and identify optimisation opportunities:

  1. Separate authorisation rate from conversion rate. Many analytics tools conflate checkouts abandoned before attempting payment with genuine authorisation declines. Authorisation rate should be calculated only on attempted transactions — those that sent an authorisation request to the issuer.

  2. Segment by card type and geography. Domestic auth rates versus cross-border auth rates should be tracked separately. Credit versus debit. Consumer versus corporate. Each segment has different drivers and benchmarks.

  3. Break down by decline code. If your PSP provides decline reason codes, build a weekly distribution and investigate material changes by issuer, BIN and transaction type. A high rate of generic 05 responses is a reason to seek provider detail, not evidence by itself of fraud or MCC misclassification.

  4. Use comparable baselines. Compare your own matched cohorts and explain any external benchmark population before using it as a target. A difference alone does not diagnose a structural defect.

Auth rate optimisation is not a one-time project — it is a continuous operations function for any business with significant recurring or card-not-present volume. The math on a single percentage point of improvement is large enough to justify dedicated engineering and product attention at relatively modest scale.

Shaun Toh By Shaun Toh · Director, Digital Payments · Razer

More Payments Economics briefings