The early payment discounts companies are losing without realizing it
Senior Manager Content Marketing at Medius
- Introduction
- The discount window closes before most AP teams know it opened
- Why do companies with cash on hand still miss early payment discounts?
- The leakage is recurring, not occasional
- Where exactly does the time go?
- Visibility is the second failure, and it's the quieter one
- What it takes to capture discounts consistently
- A discount you can't act on isn't a discount
- Frequently asked questions
Hear what's covered in this article:
Many finance teams have already decided to take early payment discounts. Treasury approved the policy, and the cash is there. But a lot of those discounts still go uncollected, because the invoice reaches payment a few days after the window closes. Below, we look at where those days go and what it takes to get them back.
The discount window closes before most AP teams know it opened
Early payment discounts are one of the few savings opportunities in accounts payable that require no negotiation. The terms are already in the contract, the supplier has already agreed to them, and nothing has to be renegotiated for the savings to materialize. The only condition is that payment goes out inside the discount window.
That condition is where it breaks down. A standard 2/10 net 30 arrangement gives accounts payable 10 days to receive the invoice, code it, match it against the purchase order, resolve anything that does not line up, route it for approval, and release payment. Ardent Partners puts the average invoice cycle time at 11.6 days in its 2026 State of AP report, which leaves zero margin in a normal week.
Plenty of invoices are not clean. The same report puts the average exception rate at 19.9%, so close to one in five invoices stops somewhere for a quantity mismatch, a missing PO line, a coding question, or an approver who is out of office. An exception rarely costs hours. It costs days, because it pulls the invoice out of the automated path and drops it into a person's queue, where it waits behind everything else in that queue.
What makes this difficult to see is that nothing visibly fails. No one declines the discount. No alert fires, no approval is denied, and no one is told that a decision was made. The invoice reaches payment a few days late, the discount quietly stops applying, and the payment goes out at full value. Multiply that across a year of invoices and the company has run an unmanaged cash program it never approved. A structured invoice approval workflow is the first place that program becomes visible.
Why do companies with cash on hand still miss early payment discounts?
Because deciding to take a discount and being able to act on one are two separate functions, owned by two different parts of the business, and they are rarely reconciled against each other.
Treasury owns the decision. A company with a strong cash position, a modest cost of capital, and no near-term liquidity pressure will approve early payment as a standing policy. The math is not close, and in most organizations the policy is already written down somewhere.
Accounts payable owns the execution. Whether one specific invoice reaches payment inside its window depends on the day it arrived, how quickly it was captured, whether it matched, how many approvers it needed, whether any of them were traveling, and when the next payment run is scheduled. None of those are treasury decisions. Most of them are invisible to treasury, which sees a cash forecast and a payment file, not a queue.
The result is a company that says yes at the policy level and no at the invoice level, thousands of times a year, with no mechanism anywhere that notices the contradiction. Cash position is not the binding constraint. Cycle time is.
Your working capital strategy is only as strong as its execution.
When local teams operate in silos and treasury lacks visibility, global targets stay on paper. This playbook gives enterprise finance leaders the framework to close the gap between strategy and action.
A company sitting on significant liquidity and a company running tight both lose the discount for the same reason, which is that the invoice was not ready in time.
The gap is worth closing because the amounts are not trivial. A 2% discount on an invoice paid 20 days early returns more on that cash than almost anything else it could be doing over the same stretch, which is why treasury signs off on the policy without much debate. Terms can be pushed further still, since sliding-scale arrangements let the discount rate track the exact payment date rather than a fixed cutoff. Dynamic discounting depends on the same operational foundation, though, so a team that cannot hit a fixed 10-day window will not do better against a sliding one.
The leakage is recurring, not occasional
What this costs comes down to three numbers, and most organizations have never multiplied them together.
Take a company processing 12,000 invoices a year at an average invoice value of $2,500, which is $30 million in annual AP spend. Assume a fifth of that spend sits with suppliers offering 2/10 net 30. That share is a planning assumption rather than a published benchmark, because no reliable figure exists for how much B2B spend carries discount terms, and any source claiming otherwise is worth checking.
That puts $6 million under discount terms and $120,000 of discount value available every year. Capture half of it and the other half walks out the door, annually, without appearing as a line item anywhere.
Whether half is optimistic depends on the organization. APQC's open standards benchmark for the percentage of invoices paid within the discount period reports a median of 14.9%. That measure counts all invoices processed rather than only discount-eligible ones, so it is not a capture rate in the strict sense. It is still a reasonable indication of how infrequently invoices land inside a window at a typical organization.
The arithmetic scales with volume, so an enterprise operation running several times that invoice count is giving up proportionally more. Substitute real numbers and the output is usually larger than finance expects, for a reason worth sitting with: this is not a one-time write-off. It reprices every year, it grows with spend, and unlike most margin initiatives it requires no supplier negotiation to recover. The money is already allocated in the contract. The only question is whether the process can move fast enough to collect it.
Where exactly does the time go?
The average cycle time is not one delay. It is five smaller ones that compound, and each is addressable independently.
An invoice emailed to a shared inbox on Friday afternoon may not be entered until Monday or Tuesday. The discount clock usually starts at the invoice date, not the date the company noticed it, so the window has already been running for several days before the invoice exists in any system. Paper and PDF invoices extend this further, since someone has to key or validate the data before anything downstream can begin.
Invoices that arrive without a purchase order need a cost center, a GL account, and often a project code. When that routing depends on one person's knowledge of which business unit owns which supplier, the invoice waits for that person's availability rather than for a system.
Close to one in five invoices is diverted into manual resolution. The underlying issue is often small, such as a minor quantity variance or a freight charge that was not on the PO, but resolution requires a round trip to a buyer or the supplier. Each round trip is measured in days, and most discount windows cannot absorb even one.
Sequential routing is the most expensive habit in the chain. If an invoice needs three approvers and each is given two days, the approval stage alone consumes six days of a 10-day window before the payment team has touched it. Parallel routing, delegation rules, and escalation on inactivity recover most of that time without changing anyone's authority.
A team that executes payments weekly can approve an invoice inside the window and still pay it outside one. The approval was on time. The calendar was not. Cadence is frequently the single largest contributor, and it is also the cheapest to change, because it is a scheduling decision rather than a process redesign.
None of these individually look like a discount problem. Together they are the discount problem, and they are the same delays that produce late payments and the supplier friction that follows.
Visibility is the second failure, and it's the quieter one
Speed alone does not capture discounts. A team can be fast and still miss windows if it cannot see which invoices have a deadline attached.
Discount terms usually arrive as text on the invoice, not as structured data. If the capture process does not extract the term into a field the workflow can read, the system has no way to distinguish an invoice worth prioritizing from one that can wait three weeks. Work is then prioritized by arrival date or by whoever is loudest, which is reasonable behavior given the information available, and it is the wrong behavior given what is at stake.
The second gap is forward visibility. Most AP teams can report what was paid. Far fewer can show which invoices currently in the queue carry a discount, how many days remain on each, and which are about to expire. Without that view, discount capture cannot be managed day to day. It can only be audited afterward, and by then the outcome is fixed.
The third gap is measurement. Capture rate is not a standard metric on most AP dashboards, so most organizations cannot say what theirs is. A loss that is never measured never enters a budget conversation, which is precisely why this leak persists in companies that scrutinize every other line item. Making it visible is a reporting problem before it is a process problem, and Medius Analytics exists to put capture rate, cycle time, and at-risk invoices on the same screen as the rest of AP performance.
If your AP report isn’t ready when your forecast is due, that’s a problem you can fix.
Outdated reporting slows decisions, misses discounts, and leaves you flying blind on liabilities. This guide shows finance teams how AI-driven automation transforms reporting from reactive to real-time.
What it takes to capture discounts consistently
Occasional capture happens when an invoice gets lucky. Consistent capture requires the window to be a known, tracked attribute of the invoice from the moment it arrives.
Extract terms at capture, not at payment
Discount terms need to become structured data the moment the invoice enters the system, so the workflow can act on them. A term discovered during payment review is a term discovered too late.
Prioritize by deadline rather than arrival
Once the window is a field, the queue can sort by time remaining instead of by date received. This single change moves discount-eligible invoices to the front without adding headcount.
Separate exception triage from the main flow
Discount-eligible exceptions should route immediately to a named owner with a deadline attached, rather than joining a general exception pool where a small freight variance can sit for days next to invoices that have no clock running.
Route approvals in parallel with escalation
Simultaneous routing with automatic escalation on inactivity compresses the approval stage from days to hours, and it is usually the largest single recovery available in a 10-day window.
Schedule payment for the last eligible day
Capturing the discount does not require paying immediately. Paying on the last day inside the window captures the full discount and keeps the cash in the business longer, which is the version treasury should want.
Report capture rate monthly
What gets measured gets defended in budget season. A capture rate trending up is one of the clearest returns AP can demonstrate.
Medius Accounts Payable Automation handles the first four of these by capturing terms as structured data, routing in parallel, and surfacing exceptions against their deadlines rather than their age. Medius Payments handles the fifth, scheduling execution for the last eligible day in the window so capture does not cost the company its float, with a Payment Optimization Agent that flags cash discount opportunities before the window closes rather than after.
A discount you can't act on isn't a discount
Every supplier contract with early payment terms is an open offer the company has already accepted on paper. Whether it is collected is decided by cycle time, routing, and visibility, which are operational variables, not financial ones. That is the part worth acting on, because it means the recovery does not depend on renegotiating anything or on the market cooperating.
Start by measuring capture rate against available discounts for the last 12 months. The number is usually lower than expected, and it is the only figure that makes the rest of this conversation concrete.
See how Medius helps AP teams capture early payment discounts consistently rather than occasionally.
Frequently asked questions about early payment discounts
It means the supplier will reduce the invoice by 2% if payment is made within 10 days, and the full balance is due within 30 days. The discount window usually runs from the invoice date rather than the date the invoice was received, which is why the clock is often already running by the time AP first sees it.
In most cases yes. The discount is earned for releasing cash 20 days earlier than required, and the return over that short period is well above what idle operating cash earns or what the cash costs the business. The exception is an organization under real liquidity pressure, where preserving cash on hand outweighs the saving.
There is no published industry target. APQC's open standards benchmark for the percentage of invoices paid within the discount period reports a median of 14.9%, though that measure counts all invoices processed rather than only discount-eligible ones. The more useful number is an organization's own capture rate measured against its own eligible invoices.
Because approval and payment are separate events on separate schedules. An invoice can clear approval inside the discount window and still wait for the next scheduled payment run, and on a weekly payment cycle that run can land after the window has closed. Nothing in the approval process flags it, because from the approver's point of view the invoice was handled on time.
The ones with the least time remaining in their discount window, not the ones that arrived first. This requires the discount term to be captured as structured data when the invoice enters the system, since a queue cannot sort by a deadline it does not store.
