I like to treat the procurement cycle as the point where approvals, budgets, receiving, and payables all collide. When the workflow is disciplined, you get cleaner spend data, better liability recognition, and fewer payment disputes. When it is not, even routine purchases create noise in the close. In the sections below, I break the process into the stages that matter and show where accounting teams can tighten control without slowing the business down.
Key points at a glance
- The workflow starts with a documented need and an approved requisition, not with an invoice.
- Accounting cares about coding, budget checks, receiving, matching, and cutoff as much as procurement does.
- Goods usually need a physical receipt; services often need milestone or completion approval.
- Three-way matching is the most reliable default control for most goods purchases.
- Weak vendor master data and missing receipts are two of the fastest ways to create payment exceptions.
- A clean process improves cash discipline, auditability, and the speed of the month-end close.
What the workflow really covers in accounting
In practice, I separate procurement from simple buying because the accounting implications are different. The broader workflow, often called source-to-pay or procure-to-pay, covers the business need, supplier selection, ordering, receiving, and payment, while finance cares about whether each step creates a clean audit trail and the right liability at the right time.
That is why I do not think of this as paperwork. I think of it as a control system: the request proves the need, the approval proves authority, the receiving event proves delivery or acceptance, and the invoice proves what the supplier wants to be paid. If one of those records is missing, the numbers in the ledger become harder to trust.
In a US organization, the details also matter for tax, budget ownership, and period-end reporting. A service contract, a subscription renewal, and a pallet of office supplies all move through the same basic flow, but the accounting treatment is not identical. That distinction leads naturally into the stages themselves.
The stages from request to payment
The most useful way to map the process is to ask what record each stage creates and who owns it. When I review a workflow, I want one clear owner per stage and one clear document to support the decision.
| Stage | What happens | Accounting impact | Typical failure |
|---|---|---|---|
| Need identified | Someone documents why the organization needs a good or service. | Creates the first budget signal and starts the spend forecast. | Verbal requests and no written basis. |
| Requisition approved | The request is coded, reviewed, and authorized by the right manager. | May create an encumbrance, which reserves funds before cash moves. | Wrong cost center, wrong project, or no approval trail. |
| Supplier selected | The team confirms the vendor, quote, contract, or other commercial terms. | Supports price reasonableness and contract compliance. | Off-contract buying or unclear terms. |
| Purchase order issued | A formal order goes to the supplier. | Turns the commitment into a record that AP can follow. | Work starts before the PO exists. |
| Receipt or acceptance | Goods are received, or services are signed off as complete. | Supports liability recognition and helps liquidate the encumbrance. | Late receipts, missing sign-offs, or incomplete service evidence. |
| Invoice matched | The invoice is checked against the PO and the receipt. | Confirms the payable amount, coding, and cutoff treatment. | Price variances, quantity mismatches, or duplicate invoices. |
| Payment posted | AP approves the invoice and disburses cash. | Clears accounts payable and records the cash outflow. | Duplicate payment or wrong period posting. |
| Closeout completed | Open commitments are reviewed and the order is closed. | Improves budget accuracy and reduces stale encumbrances. | Orphaned PO balances that never get cleared. |
For goods, the receiving step is physical. For services, I usually look for a milestone approval, timesheet sign-off, service-entry sheet, or completion note that proves the work happened. That small difference is why service invoices often need more judgment than packaged goods, and it is also why month-end accruals matter when the invoice arrives later than the work.
Once those handoffs are clear, the accounting questions become much easier to answer.
Where accounting adds control without slowing purchasing
This is the point where a good finance team earns its keep. The best accounting controls do not block legitimate spend; they make sure the spend is coded, timed, and recognized correctly from the start.
I like to focus on four decisions early in the flow. First, the request should point to the right account, cost center, fund, or project. Second, the team should know whether the item is an expense, a prepaid item, or a capital asset that belongs on the balance sheet. Third, tax treatment and vendor classification should be clear, especially for services that may affect 1099 reporting or sales tax handling. Fourth, the organization should decide whether the purchase belongs in the current period or in a month-end accrual because the goods or services were already received.
If the organization uses encumbrance accounting, the PO should reserve funds as soon as it is approved. That reserved amount should then be reduced or released as the receipt and invoice are processed. In my view, encumbrances are useful because they keep budget holders honest without waiting for cash to leave the bank.
The mistake I see most often is pushing all of this work to AP after the invoice arrives. By then, the team is cleaning up someone else’s decision instead of recording the transaction cleanly. That is why the next control layer, matching, matters so much.
Why matching rules keep the close honest
Matching is the gatekeeper between a vendor bill and a payment. The right rule depends on the risk of the purchase, but the principle is the same: do not pay for something until the documents tell the same story.
| Match type | What it checks | Best fit | Weak spot |
|---|---|---|---|
| 2-way match | Purchase order and invoice | Low-risk services, subscriptions, and simple recurring charges. | Does not confirm that anything was actually received. |
| 3-way match | Purchase order, receipt, and invoice | Most goods purchases and many deliverable-based services. | Requires timely receiving discipline. |
| 4-way match | Purchase order, receipt, invoice, and inspection or quality record | Regulated, safety-sensitive, or quality-checked items. | More administrative work and slower exceptions handling. |
For most physical purchases, three-way matching is the control I trust most. It keeps the invoice honest, confirms delivery, and gives AP a defensible basis for payment. A two-way match can still be fine for software subscriptions or other low-risk services, but I would not make it the default for goods just because it is faster.
The other control that matters here is segregation of duties. The same person should not create the vendor, approve the requisition, and release payment. When those powers sit in one place, the process starts to look efficient right up until something goes wrong.
That is why the exception workflow deserves as much attention as the normal one.
Common breakdowns that create cost and friction
Most process failures are not mysterious. They are usually the result of a few repeat problems that people have learned to tolerate.
- Buying before approval. Once work starts, the organization loses leverage and the paper trail gets messy fast.
- Off-contract spending. Maverick spend weakens pricing discipline and makes supplier management harder than it needs to be.
- Weak vendor master data. Duplicate vendor records, bad remittance details, and missing tax classification create avoidable AP cleanup.
- Missing receipts or acceptance records. Without them, invoices sit in exception queues and goods received not invoiced, or GRNI, becomes harder to clear at period end.
- Overly generic coding. If everything lands in a miscellaneous account, spend analysis becomes little more than guesswork.
- Orphaned open POs. Stale commitments distort available budget and make reports look less reliable than they should be.
- Exception handling with no root-cause review. If the same variance keeps happening, manual fixes are just hiding the real issue.
In my experience, the most expensive mistake is not the individual invoice exception. It is allowing the same exception pattern to repeat because nobody owns the underlying fix.
That is the point where a finance-ready operating model starts to separate strong organizations from merely busy ones.
What a finance-ready setup looks like in 2026
If I were standardizing this flow for a US organization today, I would focus on a few non-negotiables. Keep the approval matrix clear and tied to authority limits. Require a PO for anything above policy threshold. Make the receiving rule explicit for both goods and services. Give one person or team ownership of exception aging. Reconcile open commitments regularly so the budget view stays real.
- One approval matrix with dollar limits that people can actually follow.
- One receiving rule for goods and a separate acceptance rule for services.
- One vendor master standard so duplicates and bad tax data do not spread.
- One exception queue with a named owner and aging review.
- One month-end review of accrued liabilities and open commitments.
- One small dashboard with the measures that matter, such as first-pass match rate, open PO aging, and exception aging.
In 2026, the fastest gains usually come from better integration between procurement, AP, and the ERP rather than from adding more email approvals. If I had to prioritize only three improvements, I would start with the vendor master, the receiving rule, and the exception queue. Those three changes tend to improve auditability, shorten the close, and make every later control easier to trust.
