# How to set tolerance thresholds for packaging

> A step-by-step guide to setting tolerance thresholds for packaging and corrugate invoices, by contract clause type, with a revisit cadence. Read the full guide.

Source: https://valuexpa.com/insights/how-to-set-tolerance-thresholds-for-packaging-and-corrugate
Publisher: ValueXPA (https://valuexpa.com)
Updated: 2026-09-06

---

Margin drift is the gap between what a vendor contract says and what the invoice actually charges. In packaging and corrugate, that gap tends to hide in places a standard invoice review does not look: a resin surcharge that outlived the index move that justified it, a board grade billed above the volume tier the purchase order actually qualifies for, or a minimum order charge applied to a shipment that met the minimum.

A tolerance threshold is the line that separates a variance worth investigating from one worth ignoring. Getting that line right, category by category and clause by clause, is what makes [a review process sustainable](/guides/the-quarterly-margin-drift-review-a-control-design-pattern) instead of either useless or exhausting.

## Executive Summary

Packaging and corrugate invoices pass [three-way match on price and quantity](/guides/the-three-way-match-gap-what-your-erp-structurally-cannot) but rarely get checked against the actual contract terms behind the price: resin or index adders, minimum order quantities, and board grade tiers. A tolerance threshold is the rule that decides when a variance is worth a human looking at it. Set it too tight and AP drowns in noise it dismisses without reading. Set it too loose and the same drift clears every cycle.

The mechanism that causes the drift is a price file that only captures a base unit price while the contract carries several conditional terms alongside it: an index adjustment clause, a grade tier tied to volume, and a minimum charge per order. None of those live as a field the ERP checks automatically. A tolerance threshold has to be built against the contract term, not against last month's invoice, or it just re-approves whatever the vendor billed last time.

What changes it is treating threshold-setting as a design exercise with inputs: variance type, dollar materiality, and false-positive cost, each set deliberately rather than copied from a freight or MRO template. The steps below give a repeatable method for setting and revisiting those thresholds.

## 1. What is a tolerance threshold in a packaging invoice review?

**A tolerance threshold is a pre-set variance limit, in dollars or percent, that decides whether a packaging or corrugate invoice line gets flagged for manual review or clears automatically. Below the threshold, the system assumes the difference between contract price and invoice price is rounding or unit conversion. Above it, someone checks whether the vendor applied the wrong grade, the wrong tier, or a surcharge that should have expired. The threshold is a control setting, not a contract term itself.**

A tolerance threshold sits between the contract and the invoice, not inside either one. The contract states a rate card, a grade tier, and any surcharge or minimum charge clauses. The invoice states what the vendor billed. The threshold is the rule your AP team or your review software applies to the gap between the two: is this difference small enough to auto-approve, or large enough to route to a person.

Packaging and corrugate carry more clause types per invoice than most indirect categories: base linerboard price, grade or flute substitution, freight-included versus freight-added terms, minimum order quantities, and in some contracts a containerboard index adjustment. Each of those clause types drifts differently, so a single blanket threshold for the whole category treats a rounding difference on freight-included pricing the same as a full grade substitution, which is a different-sized problem entirely.

Setting the threshold well means setting it per clause type, not per vendor and not per category as a whole. That distinction is the subject of the rest of this page.

## 2. How do you decide what threshold to set for each contract clause?

**Set the threshold from two inputs: the dollar materiality of a typical line under that clause, and the cost of a false positive landing on someone's desk. A high-volume, low-unit-price clause like base linerboard pricing tolerates a tight percentage threshold because small percentage drift still adds up across volume. A clause that fires rarely, like a minimum order charge, should flag on any occurrence at all rather than carry a percentage tolerance, because there is no volume to average the.**

Two inputs decide the setting, and both are things you already know about your own spend, not a benchmark you need from anywhere else.

The first is materiality: how much dollar variance does a given percentage tolerance actually let through, given the volume on that clause. A 2% tolerance on a high-volume linerboard price line can pass more absolute dollars in a year than a 20% tolerance on a rarely-invoiced surcharge line, because volume multiplies the miss.

The second is the cost of investigating a false positive. A grade substitution flag needs someone to pull the purchase order and the spec sheet, which takes real time, so it should only fire when the substitution is real. A minimum order charge is binary: either the shipment met the minimum or it did not, so there is no tolerance band to set at all, only a yes or no check.

### A. Percentage-based clauses

Base unit price, grade tier pricing, and freight-included adders behave as continuous variance: the invoice price sits some percentage above or below the contracted price. These clauses take a percentage threshold, and the percentage should tighten as invoice volume on that line item rises, because a small percentage compounds over more shipments.

Start with a range and adjust it after the first few review cycles rather than trying to derive an exact number up front. If a threshold on base linerboard price produces mostly noise, meaning the flagged lines turn out to be legitimate index adjustments, widen it. If it produces mostly real findings, tighten it and expect a larger flagged volume until the underlying pricing gets corrected.

### B. Binary clauses

Minimum order quantities, surcharge sunset dates, and rebate tier triggers are not continuous. The shipment either met the minimum or it did not; the surcharge either has an active justification or it does not. These clauses do not need a percentage threshold. They need a flag that fires on any occurrence, because there is no acceptable partial version of a minimum charged on an order that met the minimum.

Treat binary clauses as a checklist item in the review, separate from the percentage-variance report, so they do not get averaged into a tolerance band that was designed for a different kind of drift.

## 3. How does board grade substitution affect the threshold you set?

**Board grade substitution changes the entire unit price, not a small percentage of it, so it should be caught as a category match rather than a variance percentage. A tolerance threshold built only on dollar variance can miss a substitution if the substituted grade happens to price close to the contracted one. Build a separate check that compares the billed grade code against the contracted grade code directly, independent of what the resulting price difference looks like.**

Grade substitution is a structural mismatch, not a pricing error, and a percentage tolerance is the wrong tool for catching it. A vendor might bill a lighter flute or a lower basis weight board than the contract specifies, at a price close enough to the contracted grade's price that a percentage variance threshold does not fire.

The fix is a direct match on the grade or flute code between the purchase order or contract line and the invoice line, checked as its own control, separate from the price variance percentage. This is closer to a compliance check than a tolerance setting: the codes either match or they do not.

Where the contract allows substitution under defined conditions, such as an out-of-stock provision, the check should confirm the condition was documented, not just that a substitution occurred. That documentation requirement belongs in the contract language itself, and reviewing it against the invoice is [a form of contract term matching](/answers/tolerance-thresholds-vs-contract-term-matching) rather than a pure price comparison.

## 4. Should tolerance thresholds account for containerboard index movements?

**Yes, where the contract ties pricing to a published index, the threshold should compare the invoice adjustment against that index's actual published movement, not against a fixed percentage. Per the US Bureau of Labor Statistics Producer Price Index for converted paper and paperboard products, series WPU0915, read September 6, 2026, the July 2026 index value was 325.968, up 2.8% year over year. A surcharge or price adjustment that moves faster or slower than the index it cites is.**

Some packaging and corrugate contracts tie a portion of price to a published containerboard or converted paper index rather than a fixed rate. When that clause exists, a fixed percentage tolerance is the wrong reference point, because the index itself moves and the contract's whole purpose is to track it.

The correct check compares the invoice's adjustment against the index's own published change over the same period. If the index that governs the clause moved 2.8% year over year as of the July 2026 reading, per BLS series WPU0915 read September 6, 2026, an invoice adjustment that moved further or in the opposite direction is the variance worth routing for review, regardless of what a generic percentage band would have said.

This only applies where the contract actually references an index by name. Do not apply an index-based check to a fixed-rate contract, and do not assume a given contract references this particular series without confirming the clause language first.

## 5. How often should packaging tolerance thresholds be revisited?

**Revisit thresholds whenever the underlying contract renews, whenever a vendor's invoice volume on a given clause changes materially, and on a fixed calendar cadence regardless of either trigger, because a threshold set once and never checked again drifts out of alignment with the contract it was built against. A quarterly review of flagged-versus-cleared ratios tells you whether the threshold is still doing its job or has become noise.**

A threshold is a setting, not a fact, and settings decay. Three triggers should prompt a revisit.

Contract renewal is the most obvious: a new rate card, a new grade tier structure, or a new surcharge clause invalidates the threshold that was built against the old terms. Reviewing thresholds as a fixed step in contract renewal, rather than as an afterthought once invoices start arriving under the new terms, closes that gap before it opens.

A material change in invoice volume on a given clause is the second trigger. A clause that used to generate a handful of invoices a month and now generates many changes the materiality math behind the threshold, even if nothing about the contract terms themselves changed.

The third trigger is calendar-based and needed regardless of the other two, because drift in the review process itself, not just the contract, is easy to miss. A quarterly look at how many flagged lines turned into real findings versus how many were dismissed as false positives tells you directly whether the threshold is set correctly. A high dismissal rate says widen it. A low flag rate on a category known to carry contract complexity says tighten it.

## 6. What tools and roles are needed to enforce a packaging tolerance threshold?

**Enforcement needs three things regardless of what system runs it: a current price file that reflects the actual contract terms, a defined owner who reviews flagged exceptions on a set cadence, and a documented escalation path for confirmed drift back to the vendor. None of these require new software. A spreadsheet-based check works if someone owns it, and a system-based check fails without an owner just as easily as a spreadsheet does.**

The steps below assume three roles exist, whatever tool executes them.

- Someone owns the price file and keeps it synced to the current contract, including grade tiers, surcharge clauses, and any index reference. An out-of-date price file makes every downstream threshold meaningless, because the comparison point itself is wrong.

- Someone reviews flagged exceptions on a defined cadence: a shorter cycle for high-volume clauses, a longer one for low-volume binary checks like minimum order charges. Flagged lines that sit unreviewed for a quarter defeat the purpose of flagging them at all.

- Someone owns escalation to the vendor once a flagged line is confirmed as real drift, with a documented record of the contract clause cited and the invoice line affected. Without this step, confirmed findings accumulate without ever being recovered or corrected going forward.

A spreadsheet can carry all three roles for a smaller packaging spend base. As invoice volume grows, the review step is usually the first one that needs a system rather than a person scanning a report, simply because the volume of lines to check outgrows what one person can review on a fixed schedule.

- **Own the price file:** Keep it synced to the current contract terms, grade tiers, and any index reference so the comparison point is accurate.

- **Review flagged exceptions:** Work through flagged lines on a set cadence rather than letting them accumulate unreviewed.

- **Escalate confirmed drift:** Document the contract clause and invoice line, then send it back to the vendor for correction or credit.

For the wider pattern this sits inside, start with the [margin drift](/guides/contract-compliance-controls-p2p) guide.

## 7. Frequently Asked Questions (People Also Ask)

### What is a reasonable starting tolerance percentage for base linerboard pricing?

There is no single correct starting number, since it depends on your invoice volume and the materiality that volume produces. Start with a band you can defend, review the flag-to-finding ratio after a full cycle, and adjust from there. The goal is a threshold that catches real drift without generating more flagged lines than your review cadence can actually work through.

### Can a tolerance threshold catch a board grade substitution?

Not reliably on its own. A substitution can price close enough to the contracted grade that a percentage variance threshold does not fire. Catching substitution requires a direct match between the billed grade or flute code and the contracted code, run as its own check separate from price variance.

### Should freight-included and freight-added packaging invoices use the same threshold?

No. Freight-included pricing bundles a variable cost into the unit price, so its variance behaves differently than freight-added pricing, where the base price and the freight line can each drift independently. Set separate thresholds for each pricing structure rather than one blended setting.

### How do minimum order quantity charges fit into a tolerance framework?

They do not fit a percentage tolerance at all. A minimum order charge is either correctly applied to a shipment that fell short of the minimum, or it is not. Treat it as a binary checklist item in the review rather than trying to fit it into a variance band.

### Does a containerboard index clause change how often I should review thresholds?

Yes. An index-linked clause should be checked against the index's own published movement each time the index updates, not on the same fixed calendar as a flat-rate clause. Per BLS series WPU0915, the relevant index changes are published on a regular schedule, so an index-linked check needs to run at least as often as that publication.

### Who should own the packaging price file used for threshold comparisons?

Ownership should sit with whoever is accountable for the underlying contract, typically procurement or a category manager, with AP consuming the file rather than maintaining it. A price file maintained by the same team that processes invoices against it risks never being updated to catch drift that team is not positioned to notice.

### What happens if the tolerance threshold is set too tight?

A threshold set too tight generates more flagged lines than the review cadence can work through. When that happens, reviewers tend to start clearing flags without real investigation just to keep up, which defeats the threshold's purpose more thoroughly than a threshold set too loose would.

### Is a percentage threshold or a fixed dollar threshold better for corrugate invoices?

Each fits a different clause type. A percentage threshold suits high-volume, continuous-variance clauses like base unit pricing. A fixed dollar threshold can suit a clause with a stable per-unit cost regardless of volume. Match the threshold type to how the clause actually behaves rather than defaulting to one across the whole category.

### Does this process require an ERP capability my system does not have?

No. Three-way match in most ERPs checks the invoice against the purchase order and receipt, not against grade codes or index clauses. The threshold checks described here run alongside that match, often in a spreadsheet or a separate exception report, not inside the ERP's native matching logic.

### Is contract complexity quietly draining your operating margin?

A small systematic drift between your negotiated contracts and your actual vendor billing compounds quietly across a year of invoices. Stop guessing at your exposure and run a targeted audit.

**[Take the Free Screener → https://valuexpa.com/margin-drift-screener](https://valuexpa.com/margin-drift-screener)**

## Executive Summary

Packaging and corrugate invoices pass [three-way match on price and quantity](/guides/the-three-way-match-gap-what-your-erp-structurally-cannot) but rarely get checked against the actual contract terms behind the price: resin or index adders, minimum order quantities, and board grade tiers. A tolerance threshold is the rule that decides when a variance is worth a human looking at it. Set it too tight and AP drowns in noise it dismisses without reading. Set it too loose and the same drift clears every cycle. The mechanism that causes the drift is a price file that only captures a base unit price while the contract carries several conditional terms alongside it: an index adjustment clause, a grade tier tied to volume, and a minimum charge per order. None of those live as a field the ERP checks automatically. A tolerance threshold has to be built against the contract term, not against last month's invoice, or it just re-approves whatever the vendor billed last time. What changes it is treating threshold-setting as a design exercise with inputs: variance type, dollar materiality, and false-positive cost, each set deliberately rather than copied from a freight or MRO template. The steps below give a repeatable method for setting and revisiting those thresholds.

## 1. What is a tolerance threshold in a packaging invoice review?

A tolerance threshold is a pre-set variance limit, in dollars or percent, that decides whether a packaging or corrugate invoice line gets flagged for manual review or clears automatically. Below the threshold, the system assumes the difference between contract price and invoice price is rounding or unit conversion. Above it, someone checks whether the vendor applied the wrong grade, the wrong tier, or a surcharge that should have expired. The threshold is a control setting, not a contract term itself. A tolerance threshold sits between the contract and the invoice, not inside either one. The contract states a rate card, a grade tier, and any surcharge or minimum charge clauses. The invoice states what the vendor billed. The threshold is the rule your AP team or your review software applies to the gap between the two: is this difference small enough to auto-approve, or large enough to route to a person. Packaging and corrugate carry more clause types per invoice than most indirect categories: base linerboard price, grade or flute substitution, freight-included versus freight-added terms, minimum order quantities, and in some contracts a containerboard index adjustment. Each of those clause types drifts differently, so a single blanket threshold for the whole category treats a rounding difference on freight-included pricing the same as a full grade substitution, which is a different-sized problem entirely. Setting the threshold well means setting it per clause type, not per vendor and not per category as a whole. That distinction is the subject of the rest of this page.

## 2. How do you decide what threshold to set for each contract clause?

Set the threshold from two inputs: the dollar materiality of a typical line under that clause, and the cost of a false positive landing on someone's desk. A high-volume, low-unit-price clause like base linerboard pricing tolerates a tight percentage threshold because small percentage drift still adds up across volume. A clause that fires rarely, like a minimum order charge, should flag on any occurrence at all rather than carry a percentage tolerance, because there is no volume to average the. Two inputs decide the setting, and both are things you already know about your own spend, not a benchmark you need from anywhere else. The first is materiality: how much dollar variance does a given percentage tolerance actually let through, given the volume on that clause. A 2% tolerance on a high-volume linerboard price line can pass more absolute dollars in a year than a 20% tolerance on a rarely-invoiced surcharge line, because volume multiplies the miss. The second is the cost of investigating a false positive. A grade substitution flag needs someone to pull the purchase order and the spec sheet, which takes real time, so it should only fire when the substitution is real. A minimum order charge is binary: either the shipment met the minimum or it did not, so there is no tolerance band to set at all, only a yes or no check. ### A. Percentage-based clauses Base unit price, grade tier pricing, and freight-included adders behave as continuous variance: the invoice price sits some percentage above or below the contracted price. These clauses take a percentage threshold, and the percentage should tighten as invoice volume on that line item rises, because a small percentage compounds over more shipments. Start with a range and adjust it after the first few review cycles rather than trying to derive an exact number up front. If a threshold on base linerboard price produces mostly noise, meaning the flagged lines turn out to be legitimate index adjustments, widen it. If it produces mostly real findings, tighten it and expect a larger flagged volume until the underlying pricing gets corrected. ### B. Binary clauses Minimum order quantities, surcharge sunset dates, and rebate tier triggers are not continuous. The shipment either met the minimum or it did not; the surcharge either has an active justification or it does not. These clauses do not need a percentage threshold. They need a flag that fires on any occurrence, because there is no acceptable partial version of a minimum charged on an order that met the minimum. Treat binary clauses as a checklist item in the review, separate from the percentage-variance report, so they do not get averaged into a tolerance band that was designed for a different kind of drift.

## 3. How does board grade substitution affect the threshold you set?

Board grade substitution changes the entire unit price, not a small percentage of it, so it should be caught as a category match rather than a variance percentage. A tolerance threshold built only on dollar variance can miss a substitution if the substituted grade happens to price close to the contracted one. Build a separate check that compares the billed grade code against the contracted grade code directly, independent of what the resulting price difference looks like. Grade substitution is a structural mismatch, not a pricing error, and a percentage tolerance is the wrong tool for catching it. A vendor might bill a lighter flute or a lower basis weight board than the contract specifies, at a price close enough to the contracted grade's price that a percentage variance threshold does not fire. The fix is a direct match on the grade or flute code between the purchase order or contract line and the invoice line, checked as its own control, separate from the price variance percentage. This is closer to a compliance check than a tolerance setting: the codes either match or they do not. Where the contract allows substitution under defined conditions, such as an out-of-stock provision, the check should confirm the condition was documented, not just that a substitution occurred. That documentation requirement belongs in the contract language itself, and reviewing it against the invoice is [a form of contract term matching](/answers/tolerance-thresholds-vs-contract-term-matching) rather than a pure price comparison.

## 4. Should tolerance thresholds account for containerboard index movements?

Yes, where the contract ties pricing to a published index, the threshold should compare the invoice adjustment against that index's actual published movement, not against a fixed percentage. Per the US Bureau of Labor Statistics Producer Price Index for converted paper and paperboard products, series WPU0915, read September 6, 2026, the July 2026 index value was 325.968, up 2.8% year over year. A surcharge or price adjustment that moves faster or slower than the index it cites is. Some packaging and corrugate contracts tie a portion of price to a published containerboard or converted paper index rather than a fixed rate. When that clause exists, a fixed percentage tolerance is the wrong reference point, because the index itself moves and the contract's whole purpose is to track it. The correct check compares the invoice's adjustment against the index's own published change over the same period. If the index that governs the clause moved 2.8% year over year as of the July 2026 reading, per BLS series WPU0915 read September 6, 2026, an invoice adjustment that moved further or in the opposite direction is the variance worth routing for review, regardless of what a generic percentage band would have said. This only applies where the contract actually references an index by name. Do not apply an index-based check to a fixed-rate contract, and do not assume a given contract references this particular series without confirming the clause language first.

## 5. How often should packaging tolerance thresholds be revisited?

Revisit thresholds whenever the underlying contract renews, whenever a vendor's invoice volume on a given clause changes materially, and on a fixed calendar cadence regardless of either trigger, because a threshold set once and never checked again drifts out of alignment with the contract it was built against. A quarterly review of flagged-versus-cleared ratios tells you whether the threshold is still doing its job or has become noise. A threshold is a setting, not a fact, and settings decay. Three triggers should prompt a revisit. Contract renewal is the most obvious: a new rate card, a new grade tier structure, or a new surcharge clause invalidates the threshold that was built against the old terms. Reviewing thresholds as a fixed step in contract renewal, rather than as an afterthought once invoices start arriving under the new terms, closes that gap before it opens. A material change in invoice volume on a given clause is the second trigger. A clause that used to generate a handful of invoices a month and now generates many changes the materiality math behind the threshold, even if nothing about the contract terms themselves changed. The third trigger is calendar-based and needed regardless of the other two, because drift in the review process itself, not just the contract, is easy to miss. A quarterly look at how many flagged lines turned into real findings versus how many were dismissed as false positives tells you directly whether the threshold is set correctly. A high dismissal rate says widen it. A low flag rate on a category known to carry contract complexity says tighten it.

## 6. What tools and roles are needed to enforce a packaging tolerance threshold?

Enforcement needs three things regardless of what system runs it: a current price file that reflects the actual contract terms, a defined owner who reviews flagged exceptions on a set cadence, and a documented escalation path for confirmed drift back to the vendor. None of these require new software. A spreadsheet-based check works if someone owns it, and a system-based check fails without an owner just as easily as a spreadsheet does. The steps below assume three roles exist, whatever tool executes them. 1. Someone owns the price file and keeps it synced to the current contract, including grade tiers, surcharge clauses, and any index reference. An out-of-date price file makes every downstream threshold meaningless, because the comparison point itself is wrong. 2. Someone reviews flagged exceptions on a defined cadence: a shorter cycle for high-volume clauses, a longer one for low-volume binary checks like minimum order charges. Flagged lines that sit unreviewed for a quarter defeat the purpose of flagging them at all. 3. Someone owns escalation to the vendor once a flagged line is confirmed as real drift, with a documented record of the contract clause cited and the invoice line affected. Without this step, confirmed findings accumulate without ever being recovered or corrected going forward. A spreadsheet can carry all three roles for a smaller packaging spend base. As invoice volume grows, the review step is usually the first one that needs a system rather than a person scanning a report, simply because the volume of lines to check outgrows what one person can review on a fixed schedule. 1. Own the price file: Keep it synced to the current contract terms, grade tiers, and any index reference so the comparison point is accurate. 2. Review flagged exceptions: Work through flagged lines on a set cadence rather than letting them accumulate unreviewed. 3. Escalate confirmed drift: Document the contract clause and invoice line, then send it back to the vendor for correction or credit. For the wider pattern this sits inside, start with the [margin drift](/guides/contract-compliance-controls-p2p) guide.

## Common questions

### What is a reasonable starting tolerance percentage for base linerboard pricing?

There is no single correct starting number, since it depends on your invoice volume and the materiality that volume produces. Start with a band you can defend, review the flag-to-finding ratio after a full cycle, and adjust from there. The goal is a threshold that catches real drift without generating more flagged lines than your review cadence can actually work through.

### Can a tolerance threshold catch a board grade substitution?

Not reliably on its own. A substitution can price close enough to the contracted grade that a percentage variance threshold does not fire. Catching substitution requires a direct match between the billed grade or flute code and the contracted code, run as its own check separate from price variance.

### Should freight-included and freight-added packaging invoices use the same threshold?

No. Freight-included pricing bundles a variable cost into the unit price, so its variance behaves differently than freight-added pricing, where the base price and the freight line can each drift independently. Set separate thresholds for each pricing structure rather than one blended setting.

### How do minimum order quantity charges fit into a tolerance framework?

They do not fit a percentage tolerance at all. A minimum order charge is either correctly applied to a shipment that fell short of the minimum, or it is not. Treat it as a binary checklist item in the review rather than trying to fit it into a variance band.

### Does a containerboard index clause change how often I should review thresholds?

Yes. An index-linked clause should be checked against the index's own published movement each time the index updates, not on the same fixed calendar as a flat-rate clause. Per BLS series WPU0915, the relevant index changes are published on a regular schedule, so an index-linked check needs to run at least as often as that publication.

---

ValueXPA runs a fixed-scope Margin Drift Diagnostic that validates every service vendor invoice against contract terms, for $100M+ US industrial manufacturers and distributors. Two to four weeks. The client retains 100% of recoveries. https://valuexpa.com/contact-us