How BIN-Level Analysis Helps Improve Authorization Rates Across PSPs

The weekly payments review is going well. Global authorization rate: 92.2%, up 0.2 percentage points from the previous week. Nothing in the headline number demands attention, so the team moves on.

But buried in the same data, cards from one issuer in one market have been approving at 74% for six consecutive weeks. The cohort accounts for just 3% of transaction volume, so its performance barely registers in the blended rate.

The headline rate can hide where approvals are being lost

That is the problem with looking at the overall authorization rate in isolation. The 92.2% figure is accurate, but it compresses every issuer, card type, country and BIN range into one number. A small segment can perform dramatically worse than the rest of the book and still leave the headline rate looking healthy.

The next step is to break that number apart:

• Which issuers are declining more cards than expected?

• Is the problem concentrated in a particular card type or market?

• Is the same pattern showing up across your payment service providers (PSPs)?

Up front, so you're not left waiting for it: this post covers four ways to break down your authorization data, the mistakes that make that analysis lie to you, and where our own tooling fits. Everything before that last part works whatever you run it in.

Why the overall rate hides money

Everyone wants to know what a good authorization rate is. It's the wrong question. A blended benchmark tells you little about where your own approvals are being lost.

Visa Consulting & Analytics reports that U.S. approval rates are below 87%, the lowest of any region globally. Its top three global decline reasons by amount are insufficient funds, suspected fraud, and exceeding the approval amount.

Your own distribution matters more than that benchmark. Consider this an illustrative example, not customer data:

The 3% cohort approves 20 points below the roughly 94% rate achieved by the rest of the book, yet its small volume barely changes the 92.2% headline.

Here's what that gap looks like financially. On 1,000,000 card transactions a year, 3% means 30,000 attempts. At 74%, 22,200 are approved; at 94%, 28,200 would be approved. That 6,000-attempt gap represents roughly €480,000 of declined volume that comparable traffic would have approved, assuming an €80 average order value.

At 100,000 annual transactions, the same 3% segment represents roughly €48,000 of declined volume that comparable traffic would have approved.

You cannot assume every declined attempt is recoverable, but the comparison shows what a small underperforming cohort can conceal. And as that cohort grows? Today's 3% problem can become tomorrow's market-level approval gap.

How big the leak is has nothing to do with how visible it is in the headline number.

What a BIN gives you to work with

Once you enrich transaction data with BIN attributes, you can segment an approval-rate problem much more precisely. A single BIN can give you dimensions such as:

IXOPAY's BIN Lookup provides 70+ data points per card, including issuing bank, card type, card level, BIN country and card segment, with data refreshed weekly in partnership with the card brands.

Visa's own BIN Attribute Sharing Service makes the same case. It provides merchants with issuer name, product ID and account funding source specifically to “improve authorization rates, reduce fraud and improve the general checkout experience.”

The catch is that your analysis is only as reliable as the BIN data behind it. Visa mandated support for 8-digit BINs in 2022, and says it “will no longer be possible to rely on the first 6 digits of the PAN” for processing decisions. Visa also warns that “BIN tables purchased off the internet are only approximately 30% accurate.”

A BIN is therefore a key for turning one approval rate into several actionable segments, provided the enrichment is current and trustworthy.

The four lenses

Each lens exposes a different layer of the authorization problem, helping you move from a broad performance gap to the specific segment you can act on.

1. Issuer level — where “do not honor” stops being a dead end

Visa Consulting & Analytics reports that insufficient funds, suspected fraud and exceeding the approval amount are the three largest global decline reasons.

Say Issuer A accounts for 50,000 of your monthly transactions. Its approval rate falls from 94% to 89% over six months, but a large share of the additional declines are returned as “do not honor.” On its own, that code tells you very little. Break the data down by issuer and compare the decline pattern over time, and you may find that the increase is concentrated in the same issuer while suspected-fraud and approval-amount declines are also rising.

That shift gives you a stronger indication that the issuer has changed its decisioning. The issuer may have tightened a fraud rule or changed how it applies transaction limits. You will rarely get a notification when that happens, so your transaction data may be the first signal.

Now compare that issuer across your PSPs. If the same decline pattern appears regardless of provider, the problem is more likely issuer-side. If one PSP performs materially better for comparable Issuer A traffic, routing may give you a way to recover some of the lost approvals.

Use the finding to:

• Test alternative routing for the affected issuer where another PSP shows stronger acceptance.

• Tune retries around soft declines and avoid repeatedly retrying hard declines, which can count against you with the issuer.

• Take persistent patterns to your acquirer with issuer-level evidence, when you have a channel into the relationship.

For a deeper look at decline responses, see our guide to decoding and understanding payment declines.

2. Card type and tier — where approval and cost collide

Your card type analysis can show whether a persistent approval problem is concentrated in a particular card cohort, and whether improving acceptance for that cohort is economically worthwhile.

• European Union: The Interchange Fee Regulation caps interchange fees on consumer debit at 0.2% and consumer credit at 0.3%. Commercial cards fall outside those caps, which is why scheme-set commercial rates run several times higher.

• United States: Regulation II caps debit interchange for covered issuers, while credit interchange has no federal cap.

Now look for a cohort such as prepaid or commercial cards with a persistent approval gap. If its declines are concentrated in limit or restriction-type responses rather than fraud-related responses, the likely intervention is different. You may need to examine how those cards are routed, when failed payments are retried, or how authentication is being applied.

There is a second number to keep beside the approval rate: cost.

Optimizing routing purely for approval can increase your cost per transaction. A route that improves approval by two percentage points while costing 40 basis points more may or may not create more value. You can only answer that question when approval and cost sit in the same view.

Use the finding to segment retry and 3DS strategies by card type, then evaluate any routing change against total acceptance cost rather than approval rate alone.

3. Issuing country — the widest gaps, and the clearest explanation

Country-level analysis gives you the broadest view of why the same payment can perform differently across markets.

• Cross-border risk is materially higher. The joint EBA/ECB 2025 report on payment fraud, covering 2024 data from the European Economic Area (EEA), found that cross-border card fraud within the EEA ran at roughly seven times the domestic rate. Outside the EEA, it was roughly 17 times higher.

• Card-not-present (CNP) transactions carry disproportionate fraud risk. CNP fraud was about 13 times higher by value than card-present fraud. CNP transactions accounted for roughly 83% of card fraud value while representing only around 28% of card payment value.

• SCA changes the risk picture. Only 64% of card payment value carried SCA in 2024, and non-SCA fraud rates were about twice as high within the EEA.

That risk profile helps explain what you see in your authorization data. Issuers have their own exposure to cross-border and CNP risk, and that risk feeds into their decisioning. When your cross-border cohort approves worse than domestic traffic, there's a reason behind the gap. More importantly, the size of the gap can vary significantly by issuing country.

If the volume justifies it, test local acquiring. Where appropriate, offer local payment methods that avoid the card networks altogether. Tune your SCA exemption strategy by region. And include cross-border fees and FX markups when you calculate the business case.

4. BIN range — the most granular view, and the one that needs the most care

This is where you find the specific pockets behind the broader patterns. A country may look weak because of one issuer. An issuer may look weak because of one card product. A card type may look weak because a handful of BIN ranges are driving the result.

That is why BIN-level analysis can turn a broad observation into a routing decision. Look for BIN ranges with persistent underperformance after you have controlled for provider, country, payment method and authentication flow. If the same ranges continue to lag across comparable traffic, you have a finding worth investigating.

The potential impact of segment-level changes is real. Mastercard reports that Checkout.com merchants using network tokens saw a 10.3 percentage-point increase in approval rates and a 7.2% increase in gross sales revenue. That is a result reported for one named group, not a typical outcome you should expect from every tokenization program. It does, however, show why the payment details behind an underperforming segment are worth investigating.

• Route by BIN. Build BIN-level routing rules for ranges that consistently underperform on comparable traffic.

• Tune 3DS by BIN. Make BIN-specific 3DS decisions where the data shows a different authentication approach is warranted.

• Adjust retry timing by BIN. Change retry timing for particular ranges when their decline patterns suggest a different retry strategy.

The important caveat is that this is also where analysts can fool themselves most easily. A BIN-level result is only useful if you have enough volume, clean transaction data, consistent provider definitions and current BIN enrichment behind it.

Five ways this goes wrong

BIN-level analysis can give you a much better view of payment performance. It can also produce a convincing answer to the wrong question. Check these five failure modes before acting on a finding.

1. Not enough volume to conclude anything

A BIN with 40 transactions and a 70% authorization rate has too little data to support a routing change. A few additional approvals or declines can shift the rate by several points.

Set a minimum-volume floor and confidence check before acting.

If a BIN has only a few dozen attempts, flag it for monitoring. Once it has enough observations, compare its rate with the relevant baseline across multiple periods. Persistent underperformance will give you a stronger basis for intervention than a single weak result.

2. Outliers wearing the costume of a trend

A BIN can clear your volume floor and still be dominated by a handful of cards. One customer retrying twelve times after an insufficient-funds decline, or dunning logic repeatedly charging a card that was never going to approve, can drag down the segment rate without saying anything about the BIN itself.

Look at the distribution before the average. Compare attempts per unique card and first-attempt authorization separately from the retry-inclusive rate. If the gap disappears when retries are stripped out, fix the dunning schedule before changing the routing table.

Also check whether a decline spike is concentrated on a few cards. That can indicate card testing, which calls for a risk response.

3. Tangled causes

A BIN can look weak because most of its traffic goes through one weak provider, a difficult market or a different 3DS flow. Before calling it a BIN problem, compare the same BIN across providers while controlling for market, payment method and authentication.

Act only when the finding survives those comparisons.

4. Comparing providers without normalizing first

PSPs can report authorization attempts, retries, timestamps and decline reasons differently. One provider may count a retry as a new transaction while another does not.

Map provider data to the same definitions before comparing rates. Otherwise, you may optimize for a reporting difference rather than a payment-performance difference.

5. A snapshot of something that keeps moving

Issuer policies change. BIN ranges get reassigned. Scheme rules are updated, and BIN reference data can become stale.

Last quarter's BIN analysis is historical evidence. For live decisions, you need current enrichment and repeated measurement.

Making it repeatable

Even a well-designed BIN analysis will go stale if you have to rebuild it manually every time a payments review comes around. To make the analysis useful for ongoing optimization, build the underlying data layer around the questions you want to answer:

• Normalized transaction data: Map authorization attempts, decline reasons, retries, timestamps and outcomes to the same definitions across every PSP. This will give you a consistent basis for provider and segment comparisons.

• One authoritative BIN reference: Apply the same BIN attributes to every transaction, regardless of which PSP processed it. PSPs expose BIN data at different levels of detail and refresh it on different schedules. Using each provider’s version can therefore produce inconsistent segment definitions. IXOPAY BIN Lookup provides 70+ attributes per card, refreshed weekly in partnership with the card brands. The data is available through Payments Intelligence or by API, so the same enrichment can support your analysis, routing and risk decisions.

• Per-segment baselines: Track expected authorization performance for each meaningful cohort. That lets you identify a BIN drifting from its normal rate rather than waiting for it to become an obvious outlier.

• Segment-level anomaly detection: Monitor individual BINs, issuers and other cohorts alongside the overall rate. A 92% headline can remain stable while one segment deteriorates for weeks.

• Prioritized findings: Rank segments by the volume and value of approvals at risk, so you know which problem deserves investigation first.

• Recurring checks: Run the analysis automatically and surface meaningful changes, without rebuilding a spreadsheet for every review.

Every payment question should not become a reporting project.

Where IXOPAY fits

The analysis above only works when the data underneath it is consistent and easy to interrogate. That is where IXOPAY Payments Intelligence fits.

1. Start with reliable BIN data. IXOPAY provides the authoritative BIN reference behind the analysis, so the same enrichment can be applied regardless of which PSP processed the transaction.

2. Compare the BIN data your PSPs return with IXOPAY's reference data. Payments Intelligence holds both IXOPAY's BIN data and the original BIN data returned by each PSP. That can give you a second view of the same card data. You can identify cases where PSPs classify the same BIN differently or assign different fee information to it. This is particularly useful when you're comparing provider performance, or weighing a higher authorization rate against higher processing costs. You can investigate whether a provider difference reflects payment performance, or differences in the underlying data.

3. Give every segment an expected baseline. Payments Intelligence benchmarks BIN-level performance across authorization rates, disputes and fees, giving you an expected value to compare against. Machine-learning models then monitor two types of change: a BIN drifting from its own historical performance and a BIN moving away from its expected benchmark.

4. Turn the analysis into an ongoing workflow with IXO Nav. IXO Nav is the AI payments expert embedded in Payments Intelligence and connected to your payment data. Ask it which issuer or BIN range is dragging down your authorization rate, and it can surface the answer without sending you back to another spreadsheet.

From there, you can:

• Set scheduled monitoring so IXO Nav checks your BINs against defined benchmarks on a recurring basis.

• Set alerts for BIN-level performance changes that cross your thresholds, so deterioration surfaces without someone opening a dashboard.

• Run scenario simulations to model how the same BINs could perform through a different PSP before you move live traffic.

That closes the loop between finding an approval gap and deciding what to do about it.

From a healthy headline to an actionable gap

The next time your weekly review shows a 92.2% authorization rate, treat it as the start of a question rather than the answer to one.

Which issuers are below their baseline? Is the gap concentrated in commercial cards, a particular issuing country, or a handful of BIN ranges? Are those BINs performing differently across your PSPs?

None of that needs new data. The issuer, the card type, the country and the BIN are already in every transaction you process. What it needs is a view that keeps those dimensions intact instead of averaging them away — and the time to look before the next review comes around.

Payment teams are not short on data. They are short on time to turn that data into action.

Ready to see where your approval gaps actually sit? Request a demo.