Five Types of Invoice Matching (and When Each One Matters)

Most AP teams pick one matching policy and apply it to every invoice that comes through the door. Three-way matching everywhere is the usual default. It feels prudent. But matching is a control, not a ritual, and matching every invoice at the same level costs more than it saves.

A purchase order for $50,000 of manufacturing equipment and a recurring $29 SaaS subscription are not the same risk. Treating them identically means you either burn AP hours verifying invoices that do not need it, or leave high-value spend under-checked. The five types of invoice matching each answer a different question. Knowing which question to ask for each purchase is the skill that separates a fast AP function from a slow one.

What invoice matching actually verifies

Invoice matching is the process of comparing an incoming vendor invoice against supporting documents to verify information is correct before payment is issued. The three core documents are the purchase order, the invoice, and the goods receipt. Differences among these documents are called matching discrepancies, and those discrepancies are compared against tolerances that determine whether the invoice passes automatically or gets flagged for review.

The key fields checked during matching include supplier name, vendor code, quantity, purchase amount, PO number, address, line items, and product descriptions. The match confirms you are paying for what you ordered, at the agreed price, for goods or services you actually received. When it works, payment flows without anyone touching the invoice. When it does not, the invoice lands in an exception queue where AP time disappears.

Type 1: Invoice totals matching

Invoice totals matching is the lightest form of verification. It compares total amounts on the invoice to total amounts on the purchase order and includes the least amount of detail, which minimizes the staff time required for review.

Six totals are compared, including balance, total discount, charges, and sales tax. Each field has a variance percentage measured against expected totals calculated from the PO prices, charges, and sales tax information combined with quantities from the invoice. If the allowable tolerance is set to 20% and the total discount shows a 100% variance, that triggers a discrepancy flag even though the overall totals might match.

Totals matching works for high-volume, low-risk transactions where line-item verification is not worth the effort. It catches large overcharges and billing errors but will miss a supplier charging the right total for the wrong mix of items. For a $50 office supply order, that trade-off is fine. For a $15,000 equipment invoice, it is not.

Type 2: Two-way matching

Two-way matching compares the purchase order and the invoice, confirming that quantities, prices, and totals agree within tolerance. It answers one question: were we billed correctly for what we ordered?

This is the appropriate choice when there is no physical delivery to confirm. Services, subscriptions, utilities, and software licenses have no goods receipt to match against, so two-way matching is the natural fit. It also makes sense for low-value, high-volume purchases where the labour cost of obtaining and checking a goods receipt exceeds the risk of the spend.

Rillion describes it as ideal for service-based invoices or straightforward transactions where goods receipts are not involved. Two-way matching catches billing errors and overcharges against the PO. What it cannot catch is being billed for goods that were never delivered. An invoice for 200 units matches a PO for 200 units. If only 160 arrived, two-way matching does not know. For physical goods of any meaningful value, that gap matters.

Type 3: Three-way matching

Three-way matching adds the goods receipt note as the third document in the comparison. It verifies that the PO, the goods receipt, and the invoice all agree on quantities, prices, and totals. The question shifts from “were we billed for what we ordered?” to “were we billed for what we actually received?”

This closes the gap two-way matching leaves. The classic example: an invoice for 200 units matches the PO for 200 units and passes a two-way match. But only 160 units were delivered. Three-way matching catches the 40-unit gap.

Three-way matching is the most commonly automated type because it balances control with efficiency. It is the standard for physical goods, high-value orders, and regulated spend where paying for undelivered items carries real cost. If the purchase involves a warehouse receiving dock, three-way matching is the minimum sensible level.

Type 4: Four-way matching

Four-way matching adds an inspection or acceptance report as the fourth document. It confirms the goods received were of acceptable quality or condition. The question becomes: were we billed for what we received and accepted as meeting our standards?

This is the most rigorous type and is reserved for industries where quality control is essential, such as construction and high-end manufacturing. A shipment that matches the PO on quantity and price but fails a quality inspection would pass three-way matching and fail four-way. In industries where substandard materials carry real liability, that distinction matters.

Four-way matching is expensive in both time and process overhead. Applying it to every invoice is overkill. Applying it where quality failures create liability or rework costs is exactly what the control is designed for.

Type 5: Charges matching

Charges matching is a distinct type in ERP systems like Microsoft Dynamics 365 that compares charges information on the invoice to charges information on the purchase order. It covers freight, surcharges, handling fees, and other extras that sit outside the unit price and quantity fields.

These charges are where discrepancies accumulate quietly. A supplier quotes $12 per unit at order time but adds a $200 freight surcharge on the invoice that was not on the PO. Unit price matching does not catch it because the unit price is correct. Totals matching might catch it only if the surcharge pushes the total beyond tolerance. Charges matching catches it directly by comparing the charges line items between the two documents.

Dynamics 365 recognizes four matching types: invoice totals matching, two-way matching, three-way matching, and charges matching. Charges matching is not a separate document check but a field-level verification that plugs a specific gap the broader matching types can miss.

Tolerances, exceptions, and where matching breaks

Discrepancies come in two forms. A quantity deviation is when invoice details do not match supporting documents in terms of amount. A price deviation is when the price on the invoice does not match the purchase order. Both trigger the same outcome: an invoice hold. Payment stops until the discrepancy is resolved or manually released.

Matching tolerances are thresholds that allow small, immaterial discrepancies to pass automatically. They can be set as a percentage, a flat dollar amount, or both. Without tolerances, minor rounding differences, shipping variances, and unit-of-measure conversions create exceptions on nearly every transaction. The matching process drowns in noise and slows to a crawl.

The exception queue, not the match itself, is where AP time actually gets lost. The single biggest driver of AP cycle delays is exceptions with no named owner. A mismatch sits in a generic queue because nobody knows it is their responsibility. The invoice misses its payment window. The vendor escalates. Time burns on a problem that clear routing would have resolved quickly.

Best-in-class companies achieve a 97.1% first-time invoice match rate. The gap between that and a typical AP team's rate is not the matching logic. It is the combination of well-set tolerances and clear exception ownership.

Automation changes the economics here. When matching software flags mismatches automatically instead of requiring line-by-line comparison, the AP team sees only the transactions that genuinely need a human decision. Zerentry automates invoice data entry with AI OCR, extracting vendor, amount, VAT, and line items and syncing to Xero or QuickBooks, which shifts the workflow from manual verification to exception handling. The duplicate detection built into every plan catches duplicates before they reach the payment run.

How to choose the right type for each purchase

The primary matching decision turns on one question: did physical goods arrive that need their receipt independently confirmed before payment? Services, subscriptions, utilities, and intangible spend have no goods receipt, making two-way matching the appropriate level.

Applying three-way or four-way matching to every invoice is too rigid and too expensive. It imposes the labour of obtaining and checking extra documents, delays payments, and generates exceptions from documents that were never necessary. Applying only two-way matching to everything leaves the business exposed on physical goods and high-value spend. Both uniform approaches are wrong.

The right approach is risk-based matching. Match the type to the purchase:

For low-value purchases, the math is worth running. A 2/10 net 30 early-payment discount is worth roughly a 36% annualized return. If over-matching delays payment past the discount window, the cost of the matching exceeds the value of the matching.

The goal is not to eliminate every possible discrepancy. It is to apply verification effort where it reduces real risk and not waste it where it does not.

FAQ