Ask ten AP managers how often they run payments and you'll get three answers: weekly, bi-weekly, daily. Each defends their cadence with the same reasons — cash flow visibility, processing efficiency, vendor satisfaction. Most are picking the wrong tradeoff, but not in the way they think. Their cadence isn't broken; they picked a cadence before answering two more important questions.

Payment run scheduling looks simple and hides a lot of consequence. Pick the wrong cadence and you miss discount windows, push half your vendors into late territory, and burn the AP team's week chasing exceptions. Pick the right cadence on top of two broken upstream decisions and the same things happen anyway. The schedule is downstream. This piece walks the full sequence: the two upstream decisions that set your cadence, the cutoff design that makes any cadence work, the dual-authorization flow that keeps the run safe, how automation captures discounts you are currently missing, and the alerts that run the batch for you.

The cadence trap: why "how often" is the wrong starting question

Three patterns show up over and over in AP teams that think they have a scheduling problem:

"Vendors are complaining we pay late, so we need to run payments more often." Usually false. The invoices triggering complaints almost always cleared the AP queue 5-10 days before their due date but got paid 3-7 days after. That's an approval queue collapse, not a cadence problem. Running payments daily won't help if approvers respond in 4 days; you'll produce more partial batches and the same late tail.

"We're running daily because too many vendors need fast payment." Usually a segmentation gap. When every vendor is in the same lane, the urgent ones force the cadence for everyone. Same-day rails also carry per-transaction fees several times higher than standard ACH.

"Bi-weekly saves us time." Usually pushes you past every cutoff. A bi-weekly Wednesday cadence puts approval cutoff Tuesday afternoon. A controller out sick or hitting a multi-step approval blows through the cutoff and the invoice slides 14 days, not 2.

Cadence sits on top of two earlier decisions: how you segment vendors, and how you design cutoffs.

Decision 1: vendor segmentation

Most AP teams run a single cadence across all vendors. Split the vendor base into three tiers instead, because vendor payment behavior clusters into three patterns that need different treatment.

Tier A — Standard vendors. Long-tail recurring spend on Net 30 or Net 45, no time-sensitive obligations. Pay weekly in a single batch on standard ACH. Roughly 80% of vendor count and about 45% of payment value.

Tier B — Time-sensitive vendors. Utilities and telecoms with short windows before interruption, contractors on Net 7 or Net 15, regulated payments, and strategic suppliers with early payment discounts on the table. Run twice-weekly batches or exact-due-date payments. About 15% of vendor count, 45% of value.

Tier C — Exceptions. Broken supply requiring expedited reissue, COD arrangements, customer refunds processed through AP. Same-day rail with a hard volume cap. Should be 5% of count and 10% of value. If your Tier C is bigger, you have a queue problem dressed up as an exception problem.

Three Vendor Tiers, Three Cadences

Typical mid-market split. Tier A vendors dominate vendor count but only carry about half the payment value. Tier B is the hidden risk band — fewer vendors, equal value share, tight payment timing. Tier C is the same-day exception lane with a hard volume cap.

The tier determines the rail, the frequency, the approval path, and how exceptions are handled. Teams running daily payments for everyone are usually misclassifying Tier A as Tier B because nobody drew the lines.

Decision 2: cutoff design — the 24-hour rule

Cutoffs are where most payment run schedules break, and they break invisibly. The schedule runs on time; the work inside it doesn't.

A payment run has two cutoffs. The bank cutoff is set by your bank: ACH typically closes around 5pm ET, same-day ACH around 2:45pm ET, domestic wires around 5-6pm ET. The internal approval cutoff is when you stop accepting newly approved invoices into the current run. Most teams set the approval cutoff at or just before the bank cutoff. That's the mistake.

Your approval cutoff should sit at least 24 hours before your bank cutoff, because that gap is what absorbs exceptions. When an approver rejects an invoice or kicks it to a different cost center, the team needs time to handle it inside the same run. With a 1-hour gap, every late rejection slides invoices 7-14 days. With 24 hours, they slide one day.

Walk it backwards from a 5pm Wednesday ACH settlement:

Approval cutoff: 5pm Tuesday
Pre-stage and review: Tuesday afternoon
Last call to approvers: Tuesday morning, with an automated nudge
Invoice intake for this run: closes Monday EOD

Bi-weekly runs are the hardest to get right, because the long gap means a missed approval slides a full two weeks.

The dual-authorization flow, step by step

Segmentation and cutoffs make the run predictable. Dual authorization makes it safe. The goal is that no single person can both create a payment instruction and release it.

Here is the concrete flow for a weekly Wednesday run:

Monday EOD — intake closes. The batch is assembled from approved, coded invoices. Any invoice touching a vendor whose bank details changed in the last 30 days is automatically flagged and pulled into a review lane.
Tuesday AM — preparer review. An AP specialist (the preparer) reconciles the batch against expected invoices and PO matches. The preparer can add or remove line items. The preparer cannot release the batch.
Tuesday PM — approval cutoff. The batch locks. Its total, line count, and any flagged items are captured in a snapshot. From this point, any change invalidates the approval and forces a re-review.
Tuesday PM — authorizer release. A second person with release authority (the authorizer, typically the controller) reviews the locked snapshot and releases it. This must be a different human from the preparer, enforced by the system rather than by policy.
Wednesday — bank submission and confirmation. The batch settles. Confirmations reconcile back against the locked snapshot, and any variance raises an alert rather than waiting for month-end.

Two rules make this real rather than theatrical. First, the lock has to bite: if the batch total changes by a cent after approval, the approval dies. Teams that allow post-approval edits have single authorization with extra steps. Second, the authorizer must not be able to edit. If the controller can add a line and release it in the same action, you are back to one person holding the whole path.

Set a release threshold rather than applying this to everything. A common structure: batches under $25K release on single authorization, batches above require the second signature, and any batch containing a changed-bank-detail vendor requires it regardless of amount. That last clause is the one that stops the most expensive failure mode, since business email compromise attacks the moment payment instructions change.

Batching tactics that actually work

Batch by payment method, not by vendor. Banks process one batch per method anyway.

Pre-stage at T-24h. The batch should exist 24 hours before submission with everything approved and coded. The hour before bank cutoff is for review, not assembly.

Don't let one stuck invoice block 200 others. A flagged invoice should be pulled from the batch, not block it, and roll into the next run once cleared.

Don't intermix legal entities. Mixing entities creates reconciliation pain, audit ambiguity, and intercompany cleanup at month-end close.

How automation captures the discounts you're missing

Most teams treat early payment discounts as a negotiation outcome. Operationally, they are a scheduling outcome — and they are usually lost to timing rather than to policy.

A discount like 2/10 Net 30 means you keep 2% by paying within 10 days. On $2M of eligible annual spend, full capture is $40,000. Missing half of it because invoices land between weekly runs costs $20,000 for no reason anyone chose.

Automated platforms close that gap in four ways:

Terms are read from the record, not remembered. Discount terms are parsed from the invoice and vendor contract and stored on the vendor master, so eligibility is a data field rather than institutional knowledge that leaves when someone changes jobs.
The discount deadline drives batch assignment. Instead of asking which weekly run an invoice falls into, the system asks which run pays it before the discount expires, and assigns it there. Discount-eligible invoices get pulled forward automatically.
The tradeoff is computed, not guessed. Paying 20 days early to save 2% is an annualized return in the mid-30s as a percentage, which beats almost any use of idle cash. When cash is tight, the system can compare that return against your cost of capital and flag the ones worth taking anyway.
Missed windows surface as a metric. Discount capture rate becomes a tracked number with a named cause per miss, which is the only way it improves.

The upstream effect matters too. Suppliers are under real liquidity pressure — C2FO's 2026 supply chain analysis found 28% of suppliers increasing borrowing despite high rates. Reliable early payment is worth more to those vendors than it costs you.

The alerts that run the batch

The batch should tell you what it needs rather than waiting to be checked. Route these to a dedicated AP channel, not to individual inboxes, so the record is shared and the handoffs are visible.

T-24h, approval nudge, to the channel:

Payment run WED-08-12 locks in 24 hours. 47 invoices, $312,480 staged. 6 invoices awaiting approval, $41,200. Approvers: @dana (4), @sam (2). Approve or the batch closes without them.

At lock, to the authorizer:

Payment run WED-08-12 is locked and ready for release. 51 invoices, $338,910. Prepared by @priya. 1 flagged item: ACME Corp bank details changed 9 days ago, verification completed by @priya on 2026-08-04. Release or reject.

On a flagged exception, immediately:

Held from WED-08-12: invoice INV-4521, Northwind Ltd, $18,400. Reason: PO quantity mismatch (invoice 120 units, PO 100). Owner: @dana. Rolls to run FRI-08-14 if cleared by Thursday 5pm.

Post-settlement, to the channel:

Payment run WED-08-12 settled. 51 payments, $338,910. 2 returns pending. Discount captured: $2,140 across 7 invoices. On-time rate this week: 96%.

Three properties make these useful rather than noise. They name a person, they state a deadline, and they say what happens if nobody acts. An alert that says "6 invoices pending approval" with no owner and no consequence gets ignored by the second week.

Stop optimizing DPO

AP teams under cash pressure often try to optimize scheduling by stretching DPO. The instinct is reasonable and the metric is misleading.

DPO measures what you paid and when. It does not measure what suppliers did in response: pricing creep at renewal, deprioritized allocation when supply is tight, slower problem-solving, reduced willingness to extend credit. Suppliers absorb late payments invisibly, and the cost shows up 6-18 months later, off the AP scorecard.

Better targets: on-time payment rate (95%+), discount capture rate on eligible invoices (80%+), and exception rate (under 7%).

Implementation sequence (90 days)

Week 1-2: Pull six months of payment data. Tag every vendor by tier using payment terms, on-time rate, and historical urgency requests.
Week 3-4: Set cutoffs using the 24-hour rule. Document the exception policy and the dual-authorization threshold with named roles.
Week 5-8: Run the new schedule in parallel for one full cycle. Measure on-time rate, discount capture, and exception count.
Week 9-12: Cut over fully. Track KPIs monthly and revisit tier assignments quarterly.

Most teams see on-time rate climb from 70-80% to 92-96% and exception count drop by 40-60%. Discount capture moves slowest, because it depends on contract terms as well as scheduling.

FAQ

How often should we run payments?

For most mid-market companies: weekly for Tier A vendors, twice-weekly for Tier B, plus a same-day rail for exceptions. Weekly gives predictable cash flow visibility, allows a single approval cutoff, and matches how discount windows are structured. Daily payments add reconciliation cost and per-transaction fees without solving the approval-queue problems that drive urgency. Bi-weekly is structurally fragile because a missed approval slides 14 days instead of 7.

What does a proper dual-authorization payment run look like?

One person prepares the batch and a different person releases it, with the system enforcing the separation rather than policy. The batch locks at the approval cutoff and captures a snapshot of its total and line count; any change after that invalidates the approval. The authorizer reviews the locked snapshot and releases it but cannot edit it, because an authorizer who can add a line and release it in one action is single authorization with extra steps. Set a release threshold by amount, and require the second signature regardless of amount whenever the batch contains a vendor whose bank details changed recently.

What's the right cutoff time for payment runs?

Set your internal approval cutoff at least 24 hours before your bank's submission cutoff. For a 5pm ET ACH submission on Wednesday, approvals close 5pm Tuesday. The gap absorbs late-stage exceptions without sliding payments to the next cycle. Same-day ACH cutoffs are around 2:45pm ET; wire cutoffs vary by bank. Build a buffer for federal holidays.

How does automation improve early payment discount capture?

By making the discount deadline drive batch assignment rather than the weekly calendar. The platform parses discount terms onto the vendor master, assigns each eligible invoice to the last run that still pays it inside the window, computes the annualized return so the tradeoff against your cost of capital is explicit, and reports capture rate with a cause for every miss. Most teams lose discounts to timing rather than to a deliberate decision, which is exactly the failure automation removes.

Does AP automation eliminate the need for batched payment runs?

No, it makes batched runs work better. Banks charge per-batch fees that scale better than per-transaction, reconciliation is cleaner with fewer batches, and approval discipline is easier to enforce on a schedule. What changes is the work inside the batch: segmentation runs automatically, cutoffs are enforced by the system, and exceptions surface to the approver who can clear them rather than waiting for an AP clerk to notice.

Make Ken schedule and run your payments

Ken segments your vendors into the right tiers, enforces 24-hour cutoffs, holds the preparer and authorizer apart, and posts every batch alert to Slack with a named owner and a deadline — so payment runs assemble themselves on schedule instead of consuming your AP team's Friday afternoon. Start a free trial and ship a cleaner payment run schedule in 30 days.

Mastering Payment Run Scheduling: How to Optimize Cash Outflows Safely