Top-performing procurement teams process a purchase order in five hours, according to the American Productivity and Quality Center. Bottom performers take two to three weeks. The difference is not the software. It is the number of people who must say yes before a purchase can happen.
Every additional approval layer adds two to four days to the purchase cycle. The direct cost shows up in missed volume discounts, rush shipping fees, and production delays. The indirect cost — the procurement staff hours spent chasing signatures, the business stakeholders who stop requesting purchases because the process is too slow — is larger and harder to measure. Both are real.
The approval tax is invisible because nobody measures it
Procurement dashboards track savings, compliance rates, and contract coverage. They rarely track the cost of the approval process itself. When a manager waits four days for a $2,000 software license approval, nobody logs the value of the lost productivity. When a negotiated volume discount expires because legal review added a week to the cycle, the savings report still shows the discounted rate — even though the discount was never applied.
The approval chain is the one procurement cost that increases with every policy designed to reduce risk. Each new threshold, each additional reviewer, each compliance checkpoint is individually reasonable. Together, they form a tax that nobody's budget accounts for.
Parallel approvals cut cycle time by half without removing anyone
The most common approval structure is sequential: the budget owner approves, then procurement reviews, then legal, then finance. Each step waits for the previous one to finish. A 2026 analysis by Orolabs found that switching to parallel approvals — legal and finance reviewing simultaneously rather than one after another — reduces cycle time by 40 to 60 percent without removing any approval step.
Parallel approval is not a software feature. It is a policy decision. Most procurement systems support it. Most procurement policies forbid it because the approval policy was written when paper forms moved through in trays, and nobody updated it when the system changed.
Where to cut: approval levels that add cost without reducing risk
Not all approvals are waste. A legal review on a multi-year contract with termination liability is justified. A finance review on a purchase that exceeds the departmental budget matters. The problem is the approvals that exist because the policy template said so, not because anyone analyzed whether they prevent the thing they are supposed to prevent.
Procurement teams that audit their approval chains typically find three categories of unnecessary review:
- Redundant reviewers. Two people in the same department approving the same purchase. If they share a budget and a manager, one approval is enough.
- Threshold creep. The policy sets a $5,000 approval threshold. Five years later, inflation means the same threshold now catches purchases that are 30% smaller in real terms. The policy drifts into micromanagement without anyone noticing.
- Legacy sign-offs. A compliance checkpoint added during a specific risk event in 2019. The event passed. The checkpoint stayed. Nobody remembers why it exists.
A two-tier approval model that actually works
The organizations with the fastest procurement cycles share a pattern. They use two approval levels for purchases under $50,000: the budget owner and one procurement reviewer. Three levels is the maximum, reserved for contracts above $250,000 or multi-year commitments. Every additional level beyond three adds cost without meaningfully reducing risk.
Thresholds must be adjusted annually for inflation. A $5,000 threshold set in 2020 should be closer to $6,500 today. If the policy does not index its thresholds, it tightens every year without anyone deciding to tighten it.
The specific numbers matter less than the principle: the cost of the approval process must not exceed the risk of the purchase. For a $2,000 software license, a four-person approval chain costs more in staff time and delay than the license itself. The math does not care about policy intentions.
How much do approval delays actually cost?
Every additional approval layer adds two to four days to the purchase-to-order cycle. Top-performing procurement teams complete the entire PO process in five hours, according to APQC benchmarks. Bottom performers take weeks. The direct costs include missed volume discounts, rush shipping fees, and production delays. Indirect costs from staff time spent chasing approvals and stakeholders who bypass procurement entirely add another layer.
What is the right number of approval levels for a purchase?
Two levels is sufficient for most purchases under $50,000: the budget owner and one procurement reviewer. Three levels should be the absolute maximum, reserved for contracts above $250,000 or those with multi-year commitments. Every additional level beyond three adds cost without meaningfully reducing risk — the marginal risk reduction from a fourth approver is near zero in most organizations.
Does procurement software solve the approval bottleneck?
Software automates the routing but does not change the number of approvers. Parallel approval workflows, where legal and finance review simultaneously rather than sequentially, can cut cycle time by 40-60% without removing any approval step. But the largest gains come from eliminating unnecessary approvals entirely — removing a reviewer costs nothing and saves more time than any software feature.
What approval thresholds should a procurement policy set?
Under $5,000: budget owner only. $5,000 to $50,000: budget owner plus one procurement reviewer. $50,000 to $250,000: add a second reviewer. Above $250,000: three levels maximum. Index these thresholds to inflation annually. The $5,000 threshold set in 2020 should be approximately $6,500 today. Without indexing, the policy silently tightens every year.
Sources
- APQC — Procurement Benchmarks: Purchase Order Cycle Time
- Orolabs — How to Reduce Procurement Cycle Times Using Automated Parallel Approvals (2026)
- Procurify — How to Improve Purchase Order Cycle Time
- Planergy — PO Cycle Time: Calculate and Improve It
- Akirolabs — Optimizing Procurement Cycle Time: Key Insights and Strategies