The processing statement is the best prospecting document in payments – if you can decode it fast. In the hosted tool, statement and Schedule A files are processed in the browser and are not uploaded to the Statement Forensics backend. Internet access is required to sign in and load the authorized rate vintage, and the audit then checks the statement against the applicable published or program-specific evidence for that month.
It reads your own book exactly as well as it reads a competitor's. Somebody is going to run this audit on the accounts you hold – the only question is whether it is you, or the agent trying to take them. The hosted tool is the product, and access is granted through the existing Statement Forensics application rather than a separate download or install flow.
Full cost stack in one pass: interchange at published rates, assessments, the incumbent's markup, and padded lines – named, with monthly dollars. Reconciles to the penny or tells you it can't. Lines the schedule can't confirm are flagged as unverified – program-priced or simply not in the table – never guessed into a number you'd have to defend.
On a real 64-category Chase interchange-plus statement the audit prices each line it can support with named evidence, then shows the source and evidence tier. Visa and Mastercard values are network-published where the backend evidence supports that statement; Discover values are identified as program or plan-guide results when they are derived from a processor guide rather than a public brand schedule; American Express remains disclosed with its OptBlue participation and fee limits. Where a category genuinely lacks defensible source support, it is identified rather than guessed.
Interchange is the card networks’ cost and no processor’s to mark up. When it is billed above the published rate inside a line presented as pass-through, that difference is margin the ISO keeps — and because it never appears as disclosed margin, it is generally outside the residual you are paid on. This is the one check that protects your own economics as directly as your merchant’s. On the Chase statement it found seven such lines. You cannot ask for a share of margin nobody has told you exists.
Repricings named and priced across months — including increases pre-announced in the statement messages (“flagged before it landed” is a door-opening sentence) — plus sponsor-bank and terms changes surfaced with effective dates. When a merchant’s bill moves, the audit decomposes exactly why. “Here’s what changed, in dollars” is a conversation nobody else on the account can have.
On recovery, ERR and billback pricing the processor’s margin is built into the rate printed on each interchange line rather than billed as a fee. Priced against the published schedules, one real statement disclosed 0.31% and was actually charging 0.50% — 38% of the markup appeared as a rate nowhere on the page. Not an overcharge, and the audit does not call it one. But a 31 basis point quote and a 50 basis point reality are not the same deal, and until now nobody could show the difference.
Where a statement prints a reason against each dispute, the audit groups them and names who owns the fix. Fraud is an authentication problem. “Not as described” is a fulfillment problem that no gateway setting touches. A chargeback on a declined authorization is the merchant’s own process and close to indefensible in a representment. Most merchants have been sold fraud tools for what turns out to be a shipping problem. That conversation is worth more than the dollars in it — and it lands on the cover page, not buried in an appendix.
Every account you hold is a statement someone else can read. The only open question is whether you get to it first, or whether the findings arrive in a competitor’s hand with a proposal attached. Run the same audit across your own portfolio and you learn what is on those statements while it is still a service call — in your framing, on your timing, before anyone else has framed it for you.
Comparing rates is the easy half and it is where most reviews stop — it is also the half a spreadsheet already does. The same statement announces its own repricing in a paragraph nobody reads, reports the ratios the card networks measure the account on, prices authorizations that never settle, and carries fees with no published rate behind them. And an account can be indefensibly priced as a whole while every line on it passes — the rate anomaly, measured against that account’s own network floor rather than an average. A tool that only compares lines cannot produce that finding at all.
Every check is gated on data, so the tool knows which ones it could not run and why. That silence is turned into a page headed “Questions to take back to your processor”: the documents to ask for, and separately the things no export can settle — the rate the acquirer license fee is passed through at, the FANF figure for this merchant’s own profile, whether an announced increase is points or a percentage, the MCC the account is boarded under. Each carries the reason the statement cannot answer it, and none of them is phrased as an accusation.
It is a downgrade pattern nobody watched, a fee that arrived with a rate cycle and was never explained, an MCC that was wrong at boarding, a dispute ratio drifting toward a monitoring threshold. You did not price any of those. But you are the one who gets asked about them — and the question usually arrives with somebody else’s business card behind it.
Interchange changes in April and October, and processor-side increases tend to ride the same window — announced as a paragraph on page one that reads like a notice rather than a price change. Load the month before and the month after and the audit names each increase, prices it, and says whether anything pre-announced it. That is the cheapest retention call you will make all year, and the calendar sets it for you.
When the audit finds nothing wrong, that is the report: priced fairly, reconciled to the statement’s own totals, in writing and dated. A merchant holding that document has an answer ready the next time someone calls with a rate. Nothing else you can send does that job, because nothing else is checkable — every figure resolves to a published schedule the merchant can look up.
Card-testing attacks, network integrity fines billed per faulty authorization, and dispute ratios approaching monitoring-program thresholds – detected, priced, and delivered as a one-page brief before the program letter arrives. You become the agent who caught it.
A detailed working report for you – every line, rate, and month compared – and a plain-English version to hand the merchant. Same numbers, two vocabularies. Toggle and export.
Load a Schedule A – six common formats parse automatically, including multi-currency revenue-share grids whose rates repeat per currency, and manual entry always works – and verify buy rates and proposals against the same published-rate engine your audit uses. The audit also surfaces average ticket, average monthly volume, and highest ticket when stated – the underwriting figures a new-account submission needs, already reconciled. And when the new account’s first statement arrives, enter the quoted terms and get a verdict – green “promise kept” within tolerance, or the overcharge in dollars, with wrong-markup and wrong-model checks.
Fallback-bucket transactions – incomplete commercial data, late settlement, non-qualified – named with cause, fix, and dollars where provable. Visa's 2026 commercial-data overhaul (CEDP) makes this urgent: accounts still mapped to the retired Level 2 program are repriced above no-data rates, and the audit flags the stale setup with effective dates.
A repricing is only provable against a baseline. Audit each account’s current statement before anything moves and you hold the reference: every rate, every fee, every line, dated. When the next schedule lands you are not arguing about whether it went up — you are showing by how much, in dollars, per month. An account you have baselined is an account a competitor cannot surprise you on.
Quote the deal, then prove it was kept. Load the merchant’s first statement on the new pricing — not the one you quoted from — and the audit reprices that month’s actual mix at the terms you quoted against what was really charged: a promise-kept verdict, or the gap in dollars, plus checks that catch boarding at the wrong disclosed markup or on a bundled model when interchange-plus was sold. It is a confirmation, not a proposal, and it is about a different statement from the one that won the deal. It exports as its own one-page document — the merchant’s name, the period, your details, the verdict — with your cost basis, margin and split never on it. And because it runs inside a complete audit of that new statement, it also reports what else was found: a kept promise and a clean statement are two different findings.
Integration playbooks look alike: pricing migrates to the acquirer’s schedule and new line items appear — announced in the statement message nobody reads. Load two or three months and the audit names each increase, prices it, and matches it to the message that pre-announced it; an increase with no notice anywhere is flagged as a silent change. Every affected account in your book is a retention event. Every one outside it is a door.
A statement that bills interchange under its own tier labels rather than named published programs has no computable network floor — and a spread quoted against a floor that cannot be measured counts the merchant’s entire interchange bill as processor-side. The audit refuses to do that. Instead it splits the bill into the part that is known exactly (itemized fees carrying no interchange at all) and the part the published schedules can only narrow to a range, and reports total processor-side cost as a range with a source behind each end — lowest and most expensive published program the bucket could have carried, per bucket, per brand, debit never mixed with credit. Brands with no publishable interchange are excluded by name rather than guessed. A defensible range beats an indefensible number, and a named statement or a portal export closes it to a figure.
Load up to three and every finding in the document is the period’s number — spread above the floor, padded-line count, volume by brand, the cost stack, the returns and disputes screen, and the cover sentence. Per-statement detail is not lost: it lives in an appendix at the back, each section badged with the statement it belongs to. Headline totals are the sum of each statement’s own printed totals, never a merged dataset. And the returns and dispute ratios are re-tested on the summed months rather than averaged — an account can sit under threshold on a quiet month and over it across the quarter, and the quarter is what an acquirer measures.
Interchange depends on the merchant category code for every brand, not only Amex, and some gated families are structurally different rather than merely cheaper — a Visa utility credit transaction is a flat $0.75 with no percentage at all. The categories a statement bills prove the code the account is boarded under, because the networks gate those programs; that is stronger than asking, since it is what the acquirer submitted rather than what anyone believes. The direction is what makes it a conversation: the restricted programs are the cheaper ones — Mastercard’s restaurant program runs 1.19% where general retail credit runs 1.58%. A business that would qualify and is boarded generally pays that difference forever, and no processor can waive it. American Express is excluded from the reading on purpose: OptBlue industry is assigned by Amex under its own program and does not follow the Visa/Mastercard code, so counting it would tell a genuine restaurant with heavy Amex that it was miscoded.
Mastercard’s Global Merchant Audit Program replaces the Acquirer Chargeback Monitoring Program, effective 2027. The Excessive Chargeback floor steps down — 150 bps through 2028, then 130, 110 and 90 — while the 2.99% ceiling, the 100-chargeback minimum and the 300 bps HECM line stay put. Three changes matter more than the numbers: it is scored at sub-merchant level, it counts fraud without a chargeback, and it carries acquirer tiers with their own assessments and a Franchise Management review at twelve months. The audit reports a merchant who is compliant today but inside a later stage, naming the year — which is a conversation you can have in 2026 that nobody else on the account is having. Two caveats travel with it and belong in your pitch, not buried: the schedule came from an operator-supplied summary rather than Mastercard’s own publication, and one acquirer assessment figure falls where every other ladder rises — printed exactly as supplied and flagged as probably a transcription error rather than quietly corrected. Say both first; being the person who volunteers the weakness in their own source is worth more than the finding.
Mastercard charges each acquiring bank a license fee on Mastercard volume, and the rate passed through to a merchant depends on the sponsor bank and program. The observed benchmark range is 0.0039%–0.0100% of Mastercard volume across real statements and published processor schedules. The audit treats that range as a benchmark and a question to ask, not as an overcharge finding. It is named in context so the merchant can ask the right follow-up question instead of inferring a categorical conclusion from an unverified secondary value.
Independent. Built by Beech Edwards – 35+ years across processing and risk, no stake in any processor or platform – so the findings carry weight no sales deck can. Private by architecture. Everything runs locally in the browser on your machine; prospect statements are never uploaded to any server – an easy answer when a merchant asks where their data goes. Accountable to the penny. Every parse reconciles against the statement's own printed totals; a parse that can't anchor is flagged, never silently presented. Verified, not estimated. Every rate is checked against Visa and Mastercard’s published schedules and locked behind automated tests – the audit is deterministic math you can reproduce and defend, not a model’s guess made in real time. AI didn’t decide your merchant’s overcharge; verified math did – and the statement never leaves your laptop to be read by one. Every number survives contact with a processor. That is the point of the discipline underneath: interchange and assessments are published and get verified line by line; the Visa FANF depends on data one statement cannot show, so the finding is scoped honestly instead of invented; the Mastercard acquirer license fee and American Express OptBlue participation are acquirer-specific costs with no published rate, so they are benchmarked and the assumption is printed on the face of the report. Contracted markup is separated from padding and never added to it. Three kinds of cost, three standards of proof, stated rather than blurred — which is why a merchant can take the document to their processor without being embarrassed by it, and why a tool that treats every line the same is guessing on at least one of them. Evidence, not promises. The report shows what's fixed, what's negotiable, and what's avoidable – in dollars – and lets the merchant conclude. That restraint is why they believe it.
The tool ships with two built-in demos: a single sample statement, and a three-month batch that catches a pre-announced price increase landing, a PCI fee hike, a sponsor-bank change, a recurring Level 2 downgrade priced against the merchant's own better rate, and a card-testing attack emerging mid-period – with the one-page risk brief attached. Every screen behaves exactly as it does on a real statement.