The short answer
Transaction matching software pairs individual records from two or more sources, such as a network settlement file and the core postings behind it, using rules on amount, date and reference. Items that pair are cleared without anyone touching them. Items that do not pair go to an exception queue with the reason attached. At a US bank the hard part is not the matching itself. It is that the records come from very different systems: Nacha files, Fed reports, card network settlement, correspondent MT940 and camt.053 statements, and a core that summarizes everything into GL batches.
Most transaction matching products are modules inside a financial close suite, sold to corporate accounting teams and licensed per user or by entity. A bank's operations floor needs something different: daily matching across rails, at volumes measured in hundreds of thousands of items a month, worked by a team that changes with the morning queue. That is the job this page is about.
What a bank actually matches every day
Each pair below is its own reconciliation with its own keys and its own typical break. A useful demo loads at least three of them from your real files.
| Source A | Source B | Matching key | Typical break |
|---|---|---|---|
| FedACH settlement on the Fed statement | Core ACH postings by batch | Settlement date, window and batch total | Same Day window posted to the next day, returns netted into the batch |
| Visa or Mastercard settlement report | Core card GL postings | Settlement date, BIN, net amount after interchange | Interchange and fees netted, chargebacks and Reg E credits on different days |
| ATM switch and network settlement | ATM cash GL and terminal balancing | Terminal, business date, dispensed amount | Cutover times, reversals, cash loads recorded late |
| Federal Reserve statement of account and FIRD detail | Due from Fed GL | Value date, amount, IMAD or service code | Service charges, cash orders, wires posted after cutoff |
| Correspondent MT940 or camt.053 | Nostro or due from bank GL | Value date, amount, reference | Charges deducted in flight, FX rate source, value date across time zones |
| Wire system outgoing and incoming | Core postings and Fed account | End to end reference and amount | Returned wires credited to the Fed account but not posted to the customer |
Transaction matching rules that work on bank data
The rule set decides how much of the queue clears on its own. These are the rule types a bank needs, and what each one is for.
One to one
One record on each side, same amount, same date window, same reference. Wires, individual returns and most correspondent entries clear this way. It is the easy 60 to 70 percent of most estates, and every product on this page does it.
One to many and many to one
One FedACH settlement credit against several hundred core postings, or one card network net settlement against the day's card GL activity. The rule groups the many side by settlement date, window or batch and compares the total. This is where spreadsheet matching breaks down, because a VLOOKUP cannot see a group.
Many to many
Several partial settlements against several postings, for example split cash orders or a correspondent that settles one instruction in pieces. The rule has to search combinations inside a window, and the software has to show which records it grouped so a reviewer can follow the logic.
Tolerances and expected differences
Interchange netted out of card settlement, correspondent charges, rounding on FX: these are known differences, not errors. Good matching models them as expected outcomes with their own rule, so a $0.42 interchange difference clears as interchange and a real $0.42 break still shows up. Widening a tolerance to make the noise go away hides real breaks, and examiners ask about it.
Timing windows
FedACH settles at 8:30 a.m. ET for forward items and in three Same Day windows, settling at 1:00, 5:00 and 6:00 p.m. ET. Card networks settle on their own calendar. A matching rule for a bank needs a window per rail, so an item that will pair tomorrow is shown as in transit, not as a break.
Transaction matching modules compared
Most buyers searching for transaction matching are really choosing between the matching module of a close suite they already own and a separate platform. The table is written to make that choice honest. Prices are shown only where the vendor or a named benchmark publishes them.
| Product | Built for | How matching is sold | Price signal |
|---|---|---|---|
| BankAutomation | Daily bank operations: ACH, card, ATM, Fed, correspondent and suspense | Per institution, by matched items and sources | Published: $1,200, $3,900, $9,500 and $14,900 a month, annual billing at half |
| Oracle Account Reconciliation (ARCS) Transaction Matching | Oracle EPM Cloud customers | Included only in EPM Enterprise, per named user | List $500 per named user a month on Enterprise |
| OneStream Transaction Matching | OneStream consolidation customers | Released to existing customers at no added license cost; feeds built by partners | Interactive user licenses and partner services |
| BlackLine Transaction Matching | Corporate finance and shared services | Module in the BlackLine subscription, priced partly on volume | Vendr median $40,125 a year for BlackLine overall |
| FloQast Close AI | Corporate accounting teams on an ERP | AI matching inside the Close AI package | Vendr median $24,481 a year for FloQast overall |
| Trintech Cadency Match, Adra Matcher, ReconNET | Enterprise close (Cadency), mid-market (Adra), banks and credit unions (ReconNET) | Modules quoted per customer | Quote only. Rivals estimate Adra Matcher at $15K to $140K a year by volume |
| SmartStream Smart Reconciliations | Large banks with a dedicated reconciliation function | Enterprise license or managed service | Quote only |
Where the close suites win is the month end. If your bank already pays for Oracle EPM Enterprise or OneStream, their matching modules are the cheapest way to certify GL balances, and our onestream account reconciliation alternative page explains when to keep OneStream for exactly that. Where they struggle is the daily floor: every Nacha file, card network report and Fed statement has to be mapped by a partner, and every analyst who works breaks needs a license. The longer comparison, with 16 platforms and prices, is in the reconciliation software comparison for banks, and the Oracle side is in Oracle ARCS alternatives for banks.
Why spreadsheet matching stops scaling
Most community banks and credit unions start with Excel: export the core GL, paste in the settlement report, match with formulas, and highlight what is left. It works until volume or rails grow. Same Day ACH added three settlement windows. FedNow added a rail that settles every day of the year. Card programs added interchange netting and disputes. Each change adds a sheet and a person, and the evidence an examiner wants (who matched what, under which rule, approved by whom) lives in nobody's file.
- One to many and many to many matches are done by eye, so the same group is rebuilt every morning.
- Tolerances are whatever the preparer decided that day, with no record of the rule.
- Aged items are found at month end instead of the day they appear.
- When the person who built the workbook is out, the reconciliation waits.
Who buys transaction matching software for a bank
- Deposit and payment operations managers at community banks and credit unions matching FedACH, card and ATM settlement to the core in spreadsheets every morning.
- Controllers who certify due from Fed, due from banks and clearing accounts and keep finding the same aged items at month end.
- Banks on Oracle EPM, OneStream or BlackLine whose matching module was set up for finance and never reached the operations floor.
- Institutions holding a Trintech, SmartStream or FIS renewal who want a priced option for daily cash matching.
- Sponsor banks and card issuers whose program volumes outgrew the workbook.
A worked example from card settlement
Input
- Mastercard net settlement · 2026-10-05 · 184,226.18 USD · 1 entry
- Core card GL postings · 2026-10-05 · 2,317 items totaling 186,104.62 USD
- Interchange and network fees on the settlement report · 1,878.44 USD
Output
- Matched 2,317 postings grouped to one settlement, many to one, by settlement date and BIN.
- Expected 1,878.44 USD cleared as interchange and fees under the approved interchange rule.
- Break 0.00 USD left. Had a chargeback posted a day late, it would show here as in transit with its own age.
The arithmetic is the easy part: 186,104.62 less 1,878.44 is 184,226.18. What the software adds is that the grouping, the interchange rule and its version are stored on the match, so the controller and the examiner can see why 2,318 records cleared without anyone touching them. The resolver at the top of this page runs the same logic on sample card data.
Getting transaction matching live
Nothing about your core, your Fed connection or your card processor changes. The rollout reads files you already receive.
Step 01
Pick the first three reconciliations
Usually FedACH settlement, card settlement and the Fed account, because they carry the volume and the morning deadline.
Step 02
Load your real files
Nacha and Fed reports, card network and processor settlement, correspondent statements, and the core GL extract. No sample data.
Step 03
Write and approve the rules
Keys, grouping, tolerances and timing windows per rail. One person proposes each rule, a second approves it, and the version is stored on every match.
Step 04
Run a month side by side
Include a month end and a payroll cycle, compare against the spreadsheet, then switch. Most banks add the next reconciliations in the second month.
What BankAutomation costs
Pricing is published and priced per institution, not per seat. Operations is $1,200 a month for 250,000 matched items, three source connections and up to 10 users, which covers FedACH, the Fed account and the core at most community banks. Platform is $3,900 a month for 2,000,000 matched items, unlimited sources, up to 50 users, SSO and maker-checker approvals. Institution is $9,500 a month with unlimited items and users for multi-entity groups, and Enterprise is $14,900 a month for private or VPC deployment. Annual billing halves each monthly rate. Every limit is on the pricing page.
Matching is the engine under every workload on this site. The rail by rail detail is on ACH reconciliation software, debit card reconciliation software, ATM reconciliation software and federal reserve account reconciliation software, and what happens to the items that do not match is covered by payment exception management software.
The resolver above runs on sample USD card settlement data. It stores nothing you paste into it, so you can see how items are grouped, cleared and classified before you talk to anyone.
Questions about transaction matching software
What is transaction matching software?
Transaction matching software pairs individual records from two or more systems, such as a card settlement file and the core postings behind it, using rules on amount, date and reference. Matched items clear automatically and unmatched items go to an exception queue with the reason attached, so people only work the differences.
What are transaction matching rules?
Transaction matching rules tell the software which records belong together: which fields must agree, how much of a difference is tolerated, how many days apart the records can be, and whether records may be grouped one to many or many to many. Banks also need rules for expected differences such as interchange and correspondent charges, approved by a second person.
Is transaction matching only on balance sheet accounts?
No. Matching works on any two sets of transactions that should agree, whether or not a balance sheet account sits behind them. At a bank most matching does end in a balance sheet account, such as due from the Fed, due from banks, ACH settlement or card clearing, because the purpose is to prove that the GL balance is made of real, explained items.
What is the difference between transaction matching and account reconciliation?
Transaction matching pairs individual items between sources. Account reconciliation proves that an account balance is correct and signed off, usually at month end. Matching feeds reconciliation: once items are matched, the unmatched remainder explains the difference between the GL and the statement. Banks need matching daily and reconciliation sign-off monthly.
How much does transaction matching software cost?
Most vendors quote. Oracle lists Transaction Matching only in EPM Enterprise at $500 per named user a month, Vendr reports a $40,125 median for BlackLine overall, and Trintech and SmartStream quote per customer. BankAutomation publishes per institution pricing: $1,200 a month for 250,000 items, $3,900 for 2,000,000 items and $9,500 for unlimited volume.
Can transaction matching replace our spreadsheets without changing the core?
Yes. Matching software reads the files and reports your core, Fed connection, card processor and correspondents already produce. Nothing is posted back to the core and no money moves. Most banks run one month side by side with the spreadsheet, including a month end, before switching.