The most valuable percentage point in the business
An operator processing substantial deposit volume can usually find one or two percentage points of acceptance available through work that costs comparatively little. Those points fall directly to revenue, because the customers involved are already acquired, already interested and already attempting to give the operator money.
Set against the cost of acquiring an equivalent amount of new revenue through marketing, acceptance work is among the highest-return activity available. It is also unglamorous, technical and easy to defer, which is why the opportunity persists at many operators.
Starting with the data
Aggregate acceptance rate is nearly useless for diagnosis. An operator holding steady at 85% overall may have a specific issuer declining nearly everything, a market performing far below its potential or a payment method quietly failing, all invisible in the average.
Useful acceptance reporting segments by market, since regulatory and issuer environments differ fundamentally; by issuer, which is where the sharpest variation sits; by payment method; by provider or route, so that provider performance can be compared for equivalent traffic; by transaction value band, since high-value transactions decline at different rates; by customer status, separating first deposits from repeat deposits because the first is commercially critical and behaves differently; and by decline reason code.
That last dimension is the starting point for everything else.
Reading decline codes
Declines fall into categories that require entirely different responses.
Insufficient funds reflects the customer's actual balance. Retrying will not help. The useful response is product-level: offering a lower amount or an alternative method.
Do not honour is the most frustrating category, because it is a generic issuer refusal with no stated reason. It frequently indicates issuer policy on gambling transactions or an issuer risk model decision. High volumes of this code from specific issuers point at a structural problem requiring alternative methods rather than technical fixes.
Authentication required or failed indicates that additional verification was needed and either was not provided or failed. This is often recoverable through correct implementation of the authentication flow.
Invalid card details indicates data entry error and is addressable through better form design, validation and clearer error messaging.
Expired card is straightforward and is addressable by prompting the customer to update.
Suspected fraud indicates an issuer or provider risk decision. Retrying is inappropriate.
Restricted card or transaction not permitted frequently indicates a card type or issuer that will not process gambling transactions at all, which is a method mix problem.
Technical and timeout errors are genuinely transient and are the clearest candidates for automatic retry.
The distinction that matters operationally is soft against hard. Soft declines may succeed on retry. Hard declines will not, and repeatedly retrying them breaches card scheme rules, wastes processing fees and can contribute to the merchant risk profile that causes worse problems later.
The levers
Routing. Different providers achieve different acceptance on the same traffic, because of their acquirer relationships, their local presence and how their transactions are presented to issuers. Systematically comparing provider performance for equivalent traffic, and shifting volume accordingly, is the single most reliable acceptance improvement available. This requires running enough traffic through each provider to compare meaningfully, which is an argument for maintaining multiple live routes rather than a primary and a dormant backup.
Local acquiring. Transactions processed through an acquirer in the customer's own country are frequently authorised at higher rates than cross-border ones, because issuers treat domestic transactions more favourably. Establishing local acquiring in significant markets is one of the more substantial improvements available and one of the more demanding to arrange.
Retry and cascading. Automatically retrying eligible declines through an alternative route recovers a meaningful proportion. The design requirements are that only soft declines are retried, that retry counts are limited in line with scheme rules, that the retry uses a genuinely different route rather than repeating the same one, and that the customer is not charged twice.
Authentication implementation. Where strong customer authentication applies, implementing it well matters considerably. A poorly implemented flow produces abandonment and failed authentications; a well-implemented one authenticates smoothly and, where exemptions are available and appropriate, avoids adding a step at all.
Data quality. Transactions submitted with complete and accurate data are authorised more often. Address information, correct merchant descriptors and appropriate transaction categorisation all influence issuer decisions.
Method mix. Where card acceptance is structurally poor in a market, the answer is usually a different method rather than better card processing. Adding the dominant local method typically produces a larger improvement than any amount of card optimisation.
The product side of acceptance
A significant proportion of deposit failures are not payment failures at all, and payment teams working in isolation miss them.
Form design determines how often customers enter details incorrectly. Field validation, appropriate keyboard types on mobile, clear formatting guidance and sensible error messages all reduce avoidable failures.
Method presentation determines which methods customers choose. Displaying the methods most likely to succeed for a given customer, in order of likely success, improves outcomes before any transaction is attempted.
Amount handling matters where customers attempt amounts that will fail. Suggesting alternatives when a transaction declines for insufficient funds recovers customers who would otherwise abandon.
Error messaging is frequently the worst part of the flow. A generic failure message tells the customer nothing and leaves them with no next action. A message explaining what happened and what to try instead recovers a meaningful share.
Flow length matters because every additional step loses customers. A deposit requiring several screens, re-entry of details or navigation away from the game they were about to play will be abandoned by some proportion regardless of whether any transaction would have been authorised.
Saved methods improve repeat deposit acceptance and speed substantially, and the security and consent requirements around storing them are worth implementing properly.
Retry design in practice
Because retry is the lever most often implemented badly, some specific guidance.
Retry only soft declines, using the reason code rather than a general rule. Limit retry count in line with card scheme requirements, which restrict how many attempts may be made on a declined transaction. Use a different route on retry, since repeating the same route rarely changes the outcome. Apply appropriate timing, since some declines benefit from a short delay and others from an immediate alternative route. Ensure idempotency, so that a retry cannot result in the customer being charged twice, which is the failure that turns an acceptance improvement into a complaint volume increase. And measure recovery separately, so the value of retry is visible rather than absorbed into aggregate acceptance.
Making the case internally
Acceptance work competes for engineering and product attention against features that are more visible, so quantification matters.
Express the opportunity in revenue terms: current acceptance, achievable acceptance, deposit volume, and the resulting revenue difference. Include the acquisition cost recovered, since failed first deposits represent marketing spend already committed. Show market-level detail, because a poor performing market is a more compelling case than an aggregate figure. And compare against alternatives, since the honest comparison is between the cost of an acceptance project and the cost of acquiring equivalent revenue through marketing, and acceptance work usually wins that comparison decisively.
The reason this argument frequently needs making is that payment performance is invisible to most of the business. Nobody sees the customers who tried to deposit and could not, because they are not in any funnel report that gets circulated. Making them visible is generally the first step in getting the work prioritised.
A worked diagnosis
Working through a realistic case shows how the segments combine.
An operator notices overall deposit acceptance has fallen from 86% to 81% over six weeks. Nothing has changed in its own systems.
Segment by market. Four markets are stable. One has fallen from 84% to 64%. The aggregate decline is entirely attributable to that market, and everything else can be set aside.
Segment that market by method. Card acceptance has collapsed. Bank transfer and wallet acceptance are unchanged.
Segment cards by issuer. Two issuers, representing a large share of the market's cards, have moved from authorising most transactions to declining most of them.
Segment by reason code. The declines are almost entirely generic refusals rather than insufficient funds or authentication failures.
The diagnosis is now clear: two major issuers have changed their policy towards gambling transactions. Nothing the operator did caused it and no technical fix resolves it.
The available responses follow from that. Test whether an alternative acquirer, particularly a local one, presents transactions in a way those issuers treat differently. Promote the bank transfer and wallet methods that are unaffected, moving them up the method selection order for customers in that market. Add the dominant local method if it is not already offered. Communicate with customers whose deposits fail, explaining that an alternative method is likely to work. And raise it with the acquirer, since issuer policy shifts are sometimes negotiable at scheme level and the acquirer may have visibility the operator does not.
The general lesson is that the diagnosis determines the remedy. An operator that responded to the initial 5% aggregate decline by generically improving its deposit form would have spent effort on something entirely unrelated to the cause.
Measuring properly
Some definitional care is needed, because acceptance rate can be calculated several ways and the differences matter.
Attempts against successes is the basic measure and depends on what counts as an attempt. Transactions abandoned before submission, blocked by the operator's own fraud rules, or rejected for validation reasons are all sometimes included and sometimes not.
Customer-level success is frequently more useful than transaction-level success. A customer whose first attempt failed and whose second succeeded has deposited, and counting only the failure understates the outcome. Measuring the proportion of customers who successfully deposit after attempting is closer to the commercial question.
First deposit acceptance deserves separate tracking, because it is the commercially critical case and because new customers behave differently from established ones.
Recovery rate measures what proportion of initial declines are subsequently converted, whether through automatic retry or through the customer trying again. This isolates the value of retry logic and messaging.
Time to successful deposit captures how much friction customers experience even when they eventually succeed, which aggregate acceptance conceals entirely.
The recommendation is to define these once, document them, and hold them constant, because acceptance figures quoted on different bases are a reliable source of internal disagreement that consumes more time than the underlying problem.
What not to do
A short list of counterproductive responses that recur.
Retrying hard declines. Wastes fees, will not succeed and breaches scheme retry rules.
Adding providers without measuring. More routes only help if traffic is allocated according to performance, and an unmeasured route is a route nobody knows is failing.
Optimising cards in a market where issuers are hostile. Effort spent here has a low ceiling; the answer is method mix.
Removing legitimate fraud controls to raise acceptance. Acceptance improves and chargebacks follow, which is worse on every dimension including acceptance eventually, since chargeback ratios affect merchant standing.
Presenting a single aggregate figure to management. Invites either complacency or generic pressure, neither of which produces the segmented work that actually helps.
Treating acceptance as a payments problem alone. A substantial share of the opportunity sits in the deposit flow, which belongs to product, and progress requires both functions working the same problem.
Fraud controls and the acceptance trade-off
A final consideration, because the two objectives genuinely conflict and pretending otherwise produces bad decisions.
Every fraud control an operator applies to deposits will decline some proportion of legitimate transactions. Tightening controls reduces fraud losses and reduces acceptance. Loosening them does the reverse. There is no setting that eliminates both problems, and the question is where the balance sits.
The framework for making that judgement is to compare the cost of fraud accepted against the cost of legitimate transactions declined. The first is visible: chargeback amounts, fees, investigation time and the effect on merchant ratios. The second is largely invisible, since a declined legitimate customer generates no record beyond a failed transaction, and their subsequent absence is not attributed to anything.
That asymmetry in visibility systematically biases operators towards over-blocking, because the cost of fraud appears in a report and the cost of false declines does not. Correcting for it requires deliberately estimating false decline cost, including the acquisition spend attached to those customers.
The practical approach is to segment rather than to set a single threshold. Controls appropriate for a first deposit from an unverified new customer are not appropriate for the fiftieth deposit from a verified customer with two years of history. Applying uniform rules across both populations means either accepting unnecessary fraud on the first or unnecessarily declining the second.
This connects to the broader theme of the course. Payment performance is not a single dial. It is a set of decisions about specific segments, methods, markets and customer states, and operators that manage it at that granularity consistently outperform those managing it in aggregate.
Where to start
For an operator that has not worked systematically on acceptance, the sequence that produces results fastest is reasonably consistent.
Build the reporting first. Acceptance segmented by market, method, provider, issuer, reason code and customer status. Without this, every subsequent step is guesswork, and most operators discover on building it that they had a specific problem they could not previously see.
Identify the largest single gap. Usually one market, method or issuer combination accounts for a disproportionate share of the shortfall, and addressing it produces more than a broad optimisation effort.
Check the deposit flow. Walk through it as a customer, on mobile, in each significant market. Form problems, poor error messages and unnecessary steps are frequently visible immediately and are cheap to fix relative to their effect.
Compare providers on live traffic. Route comparable traffic through each available provider and measure. Claimed acceptance rates are not evidence.
Implement or review retry logic. Confirm that only soft declines are retried, through different routes, within scheme limits, with no double-charge risk.
Review method mix per market. Establish whether the dominant local methods are offered and how they are presented in the selection order.
Then negotiate. Commercial terms matter and come last, because routing and method mix move the numbers more, and because negotiating from measured performance data produces better outcomes than negotiating from a general desire for lower rates.
Most operators working through this sequence find improvement available in the first two steps alone, which is a reasonable indication of how frequently the underlying data was never assembled.