Cannabis Inventory Variance: Investigating Discrepancies and Reporting Thresholds

A cycle count that does not match the system is the start of an investigation, not the end of one, and how a licensed producer handles that gap is what an auditor and a regulator will judge.

Cannabis inventory variance is the difference between what your seed-to-sale system says you hold and what a physical count actually finds on the shelf. Every cannabis operation produces variance: weight loss as flower dries, lab samples pulled and not logged, transfers entered late, simple data-entry slips. The variance itself is normal. What separates a compliant licensed producer from a suspended one is the discipline that follows: confirm the number with a second count, document it with a reason code, investigate the root cause, and report it when a threshold is crossed. This page is the working playbook for what to do the moment a cycle count uncovers a discrepancy, the eight reason codes that classify it, and the reporting thresholds (including California’s reported “significant discrepancy” rule) that turn an internal note into a regulator notification.

This is one spoke of our wider cannabis cycle count guide for licensed producers, which covers the full count lifecycle from planning through continuous improvement across US states, Canada, and EU GMP markets.

What Is Cannabis Inventory Variance and Why Does It Matter?

Cannabis inventory variance is any measurable gap between the system-recorded quantity for a product and the quantity a physical count records. On a count sheet it shows up as the difference between System Qty and Physical Count: a vault holding 85 pre-rolls in the system but 83 on the shelf carries a variance of negative 2. In most industries a two-unit gap is a rounding note. In cannabis it is a compliance event, because regulators treat unexplained product as product that may have been diverted to the illegal market.

The reason variance carries this weight is the zero-tolerance posture regulators take toward unaccounted inventory. A discrepancy you can explain (a sample pulled for the lab, breakage during handling, moisture loss in drying) is an adjustment with a documented reason. A discrepancy you cannot explain is, in the eyes of an inspector, a potential diversion. The consequences of getting this wrong are not abstract. In California, fines for selling a recalled product can reach $10,000 per unit. In Colorado, inventory and recordkeeping issues comprised 16% of all compliance violations in 2022, the single largest category. And a Canadian cultivator had its licence suspended after an inspection found critical inventory-control failures that Health Canada described as necessary to prevent diversion to the illegal market.

Inventory accuracy in cannabis isn’t optional, it’s the foundation of every compliant operation.

Variance also has an operational cost that runs alongside the compliance risk. Undetected discrepancies mean theft and shrinkage go unnoticed, recalled products cannot be fully traced, orders cannot be reliably fulfilled, and purchasing and production decisions are made on numbers that are quietly wrong. The point of cycle counting is to surface variance early and small, while it is still cheap to investigate. For the prevention side of that equation, the loss-and-diversion controls, see our companion spoke on cannabis inventory shrinkage.

How Do You Confirm a Cannabis Inventory Variance Before Acting?

The first rule of handling a cannabis inventory variance is that the initial count is a signal, not a fact. Before any adjustment is entered or any report is filed, the discrepancy must be confirmed by a second physical count performed by a different person. This double-count is the single most important control in variance handling, and it exists because the most common cause of a discrepancy is a counting or data-entry error, not missing product.

The re-count is deliberately performed by someone other than the original counter. Segregation of duties here does two things: it removes the confirmation bias of a counter who already “knows” the answer, and it provides an independent witness to the variance for the audit trail. The cleanest version of this is a blind count, where the second counter is given the product identifier and location but not the expected quantity, so the number they report is uninfluenced by the system figure. Our spoke on cannabis blind count covers the counting methods and two-person rule in full.

The confirmation step resolves into one of three outcomes:

  • The re-count matches the original physical count. The variance is real. Proceed to documentation and investigation.
  • The re-count matches the system quantity. The original count was the error. Correct the count sheet, note the miscount, and no inventory adjustment is needed.
  • The re-count produces a third number. The product or the count method needs scrutiny: check whether the unit of measure is consistent, whether partial units or sub-lots were missed, and whether two locations were conflated. Count a third time before treating any figure as confirmed.

Only a confirmed variance moves forward. This discipline keeps a momentary slip from becoming a filed adjustment that later has to be unwound, and it keeps the investigation focused on genuine gaps rather than counting noise.

Documenting Variance: The Adjustment Reason Codes

A confirmed variance must be recorded against a standardized adjustment reason code. Reason codes turn a bare number into an explanation an auditor can follow, and they are what allows a regulator to distinguish a routine sample pull from a suspected theft at a glance. The set below is the operational convention used in the cycle count template; it is an industry and operational standard rather than a single national mandate, so confirm your own seed-to-sale system and SOPs use matching codes.

Code Meaning When to use it
SAMP Lab Sample Product pulled for testing (potency, microbial, pesticide) that was not logged out at the time. A frequent and benign source of negative variance.
DMG Damage Units broken, crushed, or spoiled during handling or storage and removed from sellable stock.
WL Weight Loss Moisture loss as flower or biomass continues to dry. Expected and recurring; track the trend rather than each instance in isolation.
ACCT Data Entry Error A keying mistake, wrong unit of measure, or transposed figure in the system. The most common discrepancy source overall.
THEFT Suspected Theft A confirmed shortfall with no benign explanation after investigation. Triggers escalation and, where required, regulator and law-enforcement notification.
XFER Transfer Not Recorded Product physically moved between rooms, facilities, or stages but not entered as a transfer in the system.
WASTE Destruction / Waste Product destroyed or sent to waste under the facility’s destruction SOP but not yet reflected in inventory.
OTHER Other A documented cause that does not fit the codes above. Always paired with a free-text note explaining the circumstance.

Alongside the code, the record should capture the date and time, the product and batch, the system quantity, the physical count, the calculated variance, who counted, who verified, and a short note. A worked example from the template reads: OG Kush 1g Pre-roll, batch BL-2026-007, Shelf 4, system 85, count 83, variance negative 2, code SAMP, note “lab sample not logged.” That single row is a complete, defensible adjustment. The reason code is mandatory; an adjustment with no code is exactly the unexplained loss a regulator reads as diversion. For the fillable sheet and the six-step count process behind it, see our cannabis cycle count template.

Never close a variance without a reason code

An adjustment entered with no reason code, or a blanket “OTHER” with no note, is functionally an unexplained loss. Auditors and regulators read unexplained loss as potential diversion, and the burden of proof falls on the licensee to show otherwise. Every confirmed variance gets a code, a note, and two names (counter and verifier) before it is closed. Build a history of these records: California auditors specifically expect to see a documented trail of cycle count records, not just a current snapshot.

Want the complete cannabis cycle count playbook?

The free Cannabis Cycle Count Guide bundles the fillable count template, the eight adjustment reason codes, the ABC and location-rotation counting schedules, jurisdiction-by-jurisdiction reconciliation rules for US states, Canada CTLS, and EU GMP, plus the investigation and reporting workflow this page describes.

Download Free Guide

How Do You Investigate the Root Cause of a Variance?

Assigning a reason code is the conclusion of an investigation, not a substitute for one. Before settling on a code, work three lines of inquiry in order, because most confirmed variances resolve at the first or second before you ever reach the question of theft.

  1. Check recent transactions for data-entry errors. Pull the seed-to-sale history for the affected batch over the days since the last count. Look for transfers entered against the wrong batch, a sale or destruction logged twice or not at all, a unit-of-measure mismatch (grams entered where units were meant), and sample pulls that never got logged out. Data entry is the most common discrepancy source, so this is where the majority of variances are explained. If you find the entry error, the code is ACCT or XFER and the gap closes here.
  2. Inspect storage for misplaced items. Product is frequently not missing, just shelved in the wrong location, mixed into an adjacent batch, or sitting in a quarantine or sampling area that the count missed. Physically search the surrounding locations, the staging and packaging areas, and any holding zones. Misplaced product that is found reconciles the variance with no adjustment at all.
  3. Scrutinize handling and process loss. If transactions and storage are clean, examine the physical handling: moisture loss in drying (code WL), breakage during processing or packaging (code DMG), trim or waste removed but not yet recorded (code WASTE). These are real, explainable losses tied to how the product was handled, and each carries its own code.

Only when all three lines are exhausted, no entry error, no misplaced product, no explainable handling loss, does an unexplained shortfall remain. That residue is what gets escalated as suspected theft or diversion (code THEFT) and triggers the reporting obligations below. Document each line of inquiry as you work it, because the investigation trail is itself audit evidence: it shows an inspector that an unexplained loss was unexplained only after a genuine search, not by default. For how this investigation feeds the broader reconciliation cadence each jurisdiction requires, see our spoke on cannabis inventory reconciliation.

What Are the Reporting Thresholds for Cannabis Inventory Variance?

A confirmed, investigated variance may cross a threshold that turns an internal adjustment into a mandatory report to the regulator. These thresholds are set jurisdiction by jurisdiction, and the most widely cited is California’s. As California’s reported rule, a “significant discrepancy” is a difference of more than 3% of the licensee’s average monthly sales. When an audit or reconciliation finds a variance over that 3% figure, the licensee must report it to the state, and often to law enforcement, within 24 hours. Treat that 3% over average monthly sales and the 24-hour clock as California’s specific stated rule; do not assume the same number applies in every state, because reconciliation frequencies and reporting triggers differ by jurisdiction.

The reconciliation frequencies that govern how often these checks must happen also vary. The table below reproduces the US-state figures as each state’s reported rule. They are minimum requirements; best practice is more frequent cycle counting to stay well ahead of the deadline.

Jurisdiction Reconciliation frequency Key note on variance
California At least every 30 days Variance over 3% of average monthly sales reported within 24 hours (California’s reported “significant discrepancy” rule).
Colorado Effectively daily (retail) Inventory issues were 16% of all compliance violations in 2022, the largest category.
Oregon At least every 14 days Strict seed-to-sale tracking via Metrc.
Washington At least every 30 days Monthly reconciliation is the standard.

The verification of California’s threshold should be made against the regulator directly. Confirm the current rule on the California Department of Cannabis Control site before relying on the 3% and 24-hour figures, as state regulations are amended over time.

Outside the US states, the reporting picture differs again. In Canada, inventory tracking is centralized through Health Canada’s Cannabis Tracking System: every licensed cultivator and processor submits a comprehensive monthly inventory report through the Cannabis Tracking and Licensing System (CTLS) by the 15th of each month, even if there was no activity. A failure to report, or data that raises accuracy concerns, can be referred to the Compliance and Enforcement Directorate. Many Canadian licensed producers run internal weekly or biweekly cycle counts precisely to avoid an end-of-month surprise on that filing. You can confirm the Canadian reporting framework with Health Canada. In EU GMP markets, full traceability and frequent inventory audits are required for pharmaceutical-grade compliance, with producers expected to account for every milligram and keep a per-batch audit trail.

What Happens When a Cannabis Inventory Variance Cannot Be Explained?

An unexplained variance is the most serious outcome in cannabis inventory management, because regulators interpret unaccounted product as product that may have been diverted to the illegal market. The posture is zero tolerance: the burden sits with the licensee to account for every gram, and a shortfall that survives a genuine investigation is treated as a potential diversion until proven otherwise.

The consequences of unexplained loss escalate quickly and stack on top of one another:

  • Fines and monetary penalties that scale with the product involved; in California, selling a recalled product can reach $10,000 per unit.
  • Licence suspension or revocation when inventory-control failures are found to threaten diversion prevention. A Canadian cultivator had its licence suspended on exactly this basis.
  • Mandatory law-enforcement reporting where the variance crosses a reporting threshold, such as California’s significant-discrepancy rule.
  • Failed audits and reputational damage that follow once a pattern of unexplained loss appears in the record.

The defence against all of this is the same discipline applied early and consistently: count on a rotating schedule so variances surface small, confirm every discrepancy with a second person, code and document each one, investigate before you escalate, and report when a threshold is crossed. A variance that is caught at two units, investigated the same day, and closed with a documented reason code is a non-event. The same variance left to compound across a reporting period, with no record of how it arose, is the kind of finding that suspends a licence.

This is where purpose-built seed-to-sale software does the heavy lifting. GrowerIQ generates count sheets, lets staff enter counts digitally, and automatically flags discrepancies as they arise. Every inventory action is logged with the user, a timestamp, and a reason, producing the complete audit trail a regulator expects, across all inventory types from live plants to finished goods to waste. GrowerIQ is cannabis seed-to-sale and operations software used by 200+ licensed facilities across 9 countries, and the variance workflow on this page is built into how it handles counts and adjustments.

Frequently Asked Questions

What counts as a cannabis inventory variance?

A cannabis inventory variance is any confirmed difference between the quantity your seed-to-sale system records for a product and the quantity a physical count finds. On a cycle count sheet it is the gap between System Qty and Physical Count, expressed as a positive or negative figure (for example, a system of 85 against a count of 83 is a variance of negative 2). A variance only becomes meaningful once it has been confirmed by a second count, because the most common cause of an apparent discrepancy is a counting or data-entry error rather than genuinely missing product. Once confirmed, the variance must be documented with an adjustment reason code and investigated for root cause.

What is California’s significant discrepancy rule for cannabis inventory?

As California’s reported rule, a significant discrepancy is a variance of more than 3% of the licensee’s average monthly sales. When an audit or reconciliation uncovers a variance over that 3% threshold, the licensee must report it to the state, and often to law enforcement, within 24 hours. California licensees are also expected to reconcile inventory at least every 30 days. This 3% and 24-hour rule is specific to California as reported; other states set their own reconciliation frequencies and reporting triggers, so confirm the current requirement with the California Department of Cannabis Control and do not assume the same figures apply in another jurisdiction.

How should you investigate a cannabis inventory discrepancy?

Work three lines of inquiry in order before assigning a final reason code. First, check recent transactions in the seed-to-sale system for data-entry errors: mis-keyed quantities, transfers logged against the wrong batch, sample pulls never logged out, or sales recorded twice. Second, inspect storage for misplaced items, because product is often shelved in the wrong location, mixed into an adjacent batch, or sitting in a sampling or quarantine area the count missed. Third, scrutinize handling and process loss such as moisture loss in drying, breakage, or unrecorded waste. Most variances resolve at the first or second step. Only a shortfall that survives all three is treated as a suspected theft or diversion and escalated.

What are the adjustment reason codes for cannabis inventory variance?

The operational convention uses eight codes: SAMP (lab sample), DMG (damage), WL (weight loss), ACCT (data entry error), THEFT (suspected theft), XFER (transfer not recorded), WASTE (destruction or waste), and OTHER (a documented cause that fits none of the above, always paired with a note). Every confirmed variance is closed against one of these codes plus a short explanation and the names of the counter and the verifier. These codes are an industry and operational standard drawn from the cycle count template rather than a single national mandate, so align them with your own seed-to-sale system and SOPs. An adjustment entered with no code reads to an auditor as an unexplained loss.

Why is unexplained cannabis inventory loss treated so seriously?

Regulators interpret unaccounted cannabis as product that may have been diverted to the illegal market, and the posture toward it is zero tolerance: the licensee carries the burden of accounting for every gram. An unexplained shortfall can trigger fines (in California, selling a recalled product can reach $10,000 per unit), licence suspension or revocation, mandatory law-enforcement reporting, failed audits, and reputational damage. A Canadian cultivator had its licence suspended after inspectors found inventory-control failures that Health Canada described as necessary to prevent diversion. The protection is to catch and document variances early, while they are small and explainable.

Get the Full Cannabis Cycle Count Guide

The free guide packages the variance investigation workflow, the eight adjustment reason codes, the fillable count template, ABC and location-rotation schedules, and jurisdiction-by-jurisdiction reconciliation rules for US states, Canada CTLS, and EU GMP into one field-ready reference your team can run from day one.

Download Free Guide
Cannabis inventory variance cycle count guide cover

About GrowerIQ

GrowerIQ is a complete cannabis production management platform. Ours is the first platform to integrate your facility systems, including sensors, building controls, QMS, and ERP, into a single simplified interface.

GrowerIQ is changing the way producers use software - transforming a regulatory requirement into a robust platform to learn, analyze, and improve performance.

To find out more about GrowerIQ and how we can help, fill out the form to the right, start a chat, or contact us.

Start today.

Let us know how to reach you, and we'll get in touch to discuss your project.

GrowerIQ does not share, sell, rent, or trade personally identifiable information with third parties for promotional purposes. Privacy Policy