Payment Chaos

Payment chaos: one day, 96 providers, 1,198 ways to say the same thing.

We analysed one anonymised day of a merchant’s payment traffic: 1.5 million exchanges with 96 payment providers. The same payment status arrived under 39 different JSON paths and as 318 different values; amounts came in major or minor units; time came in 5 formats; 54.2% of statement lines did not carry the merchant’s ID.

One status, many spellings

  • "SUCCESS"
  • "confirmed"
  • "paid"
  • "success"
  • "completed"
  • "approved"
SUCCEEDED

318 → 9

Raw status values reduced to meanings, under 39 JSON paths — one anonymised merchant day. Spellings above are illustrative.

Reconora Research · One anonymised merchant day. The merchant and its providers are not named. One day is one observation, not an industry benchmark. Open data: JSON · CSV · Report (PDF)

Key findings

message shapes
1,198
distinct JSON key paths
3,110
raw status values
318
provider × error-code pairs
211
Source: Reconora Research, one anonymised merchant day
  1. Statuses. “Payment successful” is a different field and a different word at almost every provider: 39 paths, 318 values, which reduce to 9 meanings. The same word can mean different things at one provider depending on the message.
  2. Amounts. 190 message rules carry amounts in major units (150.00), 51 in minor units (15000). The unit cannot be guessed from the number — a wrong guess is a hundredfold error.
  3. Time. Only 32 of 278 message types carry the provider’s own timestamp, in 5 formats including epoch seconds and strings without a time zone. A missing zone moves payments into the neighbouring day.
  4. Identifiers. The provider may echo your ID, return only its own, or both. 54.2% of statement lines had no merchant ID — they can only be matched through the provider’s ID.
  5. Errors. 211 distinct provider-and-code pairs in one day; many codes are numeric with no reference, some arrive only as text or an HTML error page.
  6. Topologies. The same fact arrives by callback, by polling and by repeated polling: 401,255 repeated observations had to be folded so that nothing is counted twice. Orchestrators hide the end provider inside the message.
  7. Statements. 35 statement formats: headers not on the first row, “150 AZN” in one cell, thousand separators, date and time in different columns, each format with its own time zone.
  8. Noise. 329,997 records in the log were not provider data at all; 9,635 (0.6%) could not be read and were quarantined with a reason.

What it means for finance

A reconciliation built on lookups and manual mapping is a bet that none of these differences hits your money. In the reconciliation scope of the measured day they produced 809 discrepancies and left 10.1% of turnover unverifiable. See how Reconora handles it · Canonical model · The measured day, outcome by outcome

Scope and limitations

The Payment Chaos report describes one anonymised merchant day. It is a single observation (n = 1): one merchant, one day, the providers that merchant used. The figures show what really happens in a multi-provider payment flow, with exact counts. They do not show how often it happens across merchants, verticals or seasons, and they are not an industry benchmark.

Two scopes are used, and every figure belongs to one of them:

  • The full day — every message the merchant exchanged with its 96 providers: 1.5 million records. Counts of message shapes, key paths, status values, time formats, error codes and the record balance describe the full day.
  • The reconciliation scope — the merchant’s deposit log (114,065 transactions) reconciled with notifications and statements. Shares of turnover, discrepancies, statement-only payments and the spreadsheet comparison describe this scope.

No money amounts, currencies list, merchant name, geography or date are published. How every number is produced

What each figure counts

FigureValueWhat it countsScope
Records1.5 millionMessages and log lines read in the dayFull day
Providers96Payment providers with traffic that dayFull day
Message shapes1,198Distinct structures of provider messagesFull day
Key paths3,110Distinct JSON paths across all messagesFull day
Raw status values318Distinct status words and codesFull day
Canonical meanings9Meanings those values reduce toFull day
Statement lines without the merchant’s ID54.2%Share of statement linesFull day
Repeated observations folded401,255The same fact seen again by callback or pollingFull day
Matched75.9%Share of turnover confirmed by statementsReconciliation scope
Nothing to reconcile against10.1%Share of turnover with no complete statement to checkReconciliation scope
Discrepancies809Payments in the discrepancy queueReconciliation scope
Players charged, not credited320Payments declined in the log, successful in the statementReconciliation scope
Providers without a statement26Providers that sent no statement at allReconciliation scope

Method

Figures come from a full-day run of Reconora’s engine on anonymised data. Each figure is recorded once, with its internal source and its scope, and the site will not build if a page carries a number that is not in that record. Percentages are shares of messages, rules or turnover in scope; a figure that was not measured is not shown. Methodology

Data and changelog

The structural figures are open: JSON and CSV, each figure with a label, a unit and its scope. You may quote them with a link to this page.

VersionDateChange
v1.12026-10-18Labels, units and scopes for every figure; CSV copy; scope and limitations section
v1.02026-10-16First publication

Further days may be added later; until then this report stays a single-day observation.

Cite this report

Reconora Research (2026). Payment Chaos: how 96 payment providers describe the same payment. reconora.com/payment-chaos

Try it on one day of your data

One day of your data, reconciled, and the report stays with you whether or not you continue.

By Reconora Research · Updated