Cannabis Lab Testing Operations

Operational guidance last updated: April 2026. This is the operator-side of cannabis lab testing — sampling, timeline management, failure handling, remediation decisions, and recall protocols. This is procurement guidance, not legal or medical compliance advice. Verify state regulatory rules with counsel before acting on a failed panel or a remediation decision.

Canonical COA interpretation / testing science: see coa-testing.md. This file covers the operator-procurement angle only per Phase 20 D-21. It does not duplicate panel interpretation, potency math, or annotated COA examples.

Lab selection / procurement toolkit: see lab-selection.md for evaluation rubric, RFP template, onboarding checklist, per-panel pricing benchmarks, and named labs by geography.

Summary

Audience: dispensary buyers, brand operations managers, compliance managers, and QA leads who receive COAs, move batches to the sales floor, and respond when a batch fails.
In scope: sample-to-sellable timeline, sampling workflow on the operator side, failure decision trees (retest / remediate / destroy / re-source), remediation method selection + state labeling rules + economic-threshold decision tree, recall protocols, COA-to-product reconciliation.
Out of scope: COA interpretation, testing panels, potency calculations, annotated COA examples, state testing matrices — stays in coa-testing.md (canonical). Lab selection / procurement / RFP / named labs — see lab-selection.md. Retail vault + cash handling SOPs (Phase 27 Loss Prevention). Labor-cost KPIs (Phase 26).
Procurement-advice disclaimer: Everything below is procurement guidance, not legal or medical compliance advice. Treat every state rule as "verify with state regulator" before acting.
Regulatory status changes fast. CO "Decontaminated" labeling became effective 2025-07-01; MA decontamination labeling was earlier; other states have rolling or pending rules. Verify state-specific rules at the point of decision. (All state rules referenced here are as of 2026-04.)
Three questions this file answers for the operator: (1) What's my realistic sample-to-shelf timeline? (2) What do I do when a batch fails? (3) When does remediation make economic sense vs destroy vs re-source?

Sample-to-Sellable Timeline

The question every buyer asks in their first week on the job: "A batch just finished curing. When can I put it on the shelf?" The answer depends on panel type, lab workload, and whether your operation is vertically integrated. This section gives a realistic end-to-end timeline so you can plan receiving slots, retail floor-sets, and promotional calendars without over-committing.

Typical Timeline (Standard Panel, Flower)

| Day | Step | Owner | Gate | |-----|------|-------|------| | D0 | Harvest / batch creation | Cultivator | Metrc tag applied | | D1-D2 | Batch conditioning (trim, cure, moisture-balance) | Cultivator | Moisture + aesthetic check | | D3 | Sample pull (ISO 17025 chain-of-custody by lab sampler) | Lab sampler | Operator countersigns chain-of-custody form | | D3 | Lab intake / accessioning | Lab | Seed-to-sale manifest entered | | D4-D9 | Analysis (potency / pesticides / heavy metals / microbials / residual solvents / mycotoxins / water activity) | Lab | Per-panel SOPs | | D10 | COA released; uploaded to operator portal + state system | Lab | Final report + CoA PDF | | D10 | Operator reviews COA; files in batch record | Ops / QA | No-fail confirmed | | D11 | Release to distribution / dispensary shelf | Ops | Dispensary receiving |

(As of 2026-04; industry synthesis per lab-selection.md §Turnaround SLA Benchmarks. Verify against your specific lab's written SLA commitments.)

The practical reality: a standard full-compliance panel on flower adds 7-11 days between harvest-complete and sellable-on-shelf. Rush panels can compress that to 3-4 days at +50-100% cost. Any timeline tighter than 3 days is either a lab cutting corners or a rush queue that will miss SLA under load.

Accelerated Timeline (Rush Panel)

| Day | Step | Owner | Premium | |-----|------|-------|---------| | D0 | Harvest / batch creation | Cultivator | — | | D1 | Expedited conditioning + sample pull | Cultivator + Lab sampler | Sampling rush fee ~$200-500 (as of 2026-04) | | D1-D2 | Rush analysis (24-48hr SLA) | Lab | +50-100% per-panel pricing (as of 2026-04) | | D3 | COA release | Lab | — | | D3-D4 | Operator review + release to shelf | Ops | — |

(As of 2026-04; rush availability concentrated in CA, CO, MA, FL, NV, OR, NY. Smaller markets rarely support <72hr turnaround. Confirm rush availability BEFORE planning a promotion or new-product drop.)

When to pay for rush:

| Situation | Pay for Rush? | |-----------|---------------| | New-product launch with committed retail date | Yes — late product costs more than rush fee | | Routine harvest, no promotional calendar pressure | No — standard SLA preserves margin | | Batch pulled from quarantine after failed retest | Yes — inventory-hold days are expensive | | Expiring trim/shake that must move to concentrate | Hybrid — bundle with other rush work to amortize | | Customer-specific order (wholesale) with delivery date | Yes — wholesale SLA failure has reputational cost |

(As of 2026-04; rush economics vary by market and lab — model against your specific lab's rate card.)

Chain-of-Custody (ASCII Diagram)

[CULTIVATOR VAULT]
      │
      │ (1) Batch ready; operator notifies lab
      ▼
[LAB SAMPLER arrives on-site]
      │
      │ (2) Sampler pulls representative samples per ISO 17025 SOP
      │     Operator witnesses; both parties sign chain-of-custody form
      ▼
[LAB COURIER / SAMPLER TRANSPORT]
      │
      │ (3) Sample bagged + sealed + tagged; sampler transports
      │     Operator retains sampler's copy of chain-of-custody
      ▼
[LAB ACCESSIONING]
      │
      │ (4) Lab intake logs sample; unique lab ID assigned
      │     Seed-to-sale system updated (Metrc / BioTrack)
      ▼
[LAB BENCH ANALYSIS]
      │
      │ (5) Panels run per accredited SOPs; data QC'd
      ▼
[COA ISSUED]
      │
      │ (6) COA posted to operator portal + state system
      │     Operator notified via email/portal
      ▼
[OPERATOR RELEASE DECISION]
      │
      ├─── PASS → release to distribution / dispensary shelf
      └─── FAIL → failure decision tree (§Failure Decision Trees)

(Pattern as of 2026-04; verify your state's seed-to-sale integration requirements. Most accredited labs follow this exact flow.)

Why the Timeline Varies

Panel mix matters. Potency-only can clear in 2-5 days (as of 2026-04); full compliance runs 5-10 days. Residual-solvents-only on concentrates can clear in 3-5 days.
Lab workload cycles. Harvest season (September-November in outdoor markets) adds 1-3 days to standard SLAs; concentrate-heavy states see bumps around Q4 promotional cycles.
Sampler scheduling. In-demand markets have 2-5 business-day lead times for sampler visits. Plan batch-completion dates backward from this.
Dispute and retest paths extend the timeline. If a panel comes back borderline, the retest queue adds 3-7 more business days.
State portal lag. In some states (NY, NJ as of 2026-04) the state seed-to-sale system has a 24-48hr lag between lab upload and operator visibility. Don't confuse portal-lag with a genuine COA problem.

Timeline-Risk Buffer Recommendations

As of 2026-04, operator-playbook guidance for building timeline buffers:

| Scenario | Recommended Buffer | |----------|-------------------| | New-product launch (firm retail date) | +7 business days beyond standard SLA | | Routine harvest (no promotional commitment) | +2 business days beyond standard SLA | | Harvest-season peak (Sep-Nov outdoor) | +3-5 business days beyond standard SLA | | State-portal-lag states (NY, NJ) | +2 business days beyond COA receipt | | Cross-state wholesale commitments | +7 business days; confirm receiving-state rules before ship date |

(As of 2026-04; customize based on your lab's historical SLA performance.)

Inventory-Aging Impact

A batch in quarantine is inventory capital sitting idle. For a 50-lb batch worth $45,000 wholesale, each day in quarantine is ~$12 in carry cost (3% annualized cost of capital, as of 2026-04). A 14-day quarantine adds ~$170 carry cost to the batch — small compared to panel cost but meaningful for high-throughput operators. Factor this into your "pay for rush" calculation.

Sampling Process (Operator Side)

Canonical for lab-science sampling: see coa-testing.md §The Sampling Process. The section below covers what the OPERATOR does to prepare for, witness, and close out a sampling event — the pre- and post-sample workflow that coa-testing.md does not cover.

Preparing the Batch

Batch-size cap. Most states cap a test batch at a specific quantity. Typical caps as of 2026-04 (verify with state regulator):

CA: 50 lb flower per batch
CO: 15 lb flower per batch (DOR rule)
MA: 10 lb flower per batch (CCC rule)
NV, OR, WA: 10-15 lb flower per batch (verify current rule)
NY, NJ: emerging rules; verify with state regulator

Exceeding the batch cap means the batch gets split before sampling — each sub-batch is its own sample event and its own COA. Build 10-15% cost/time buffer into harvest planning for batches that will split.

Batch uniformity. Samples must represent the whole batch. If a 15-lb batch contains visually distinct material (different cultivars, different harvest dates, different moisture bins), it SHOULD be split before sampling even if it's under the cap. Non-uniform batches fail more often because the sample pull skews toward the densest or driest material.

Batch hold. Once the batch is designated for sampling:

Pull the batch from active inventory in Metrc (status: "sampling pending").
Physically isolate the batch in a sampling-pending quarantine area (tagged, separated, access-restricted).
Document the quarantine event in your batch record — date, tag ID, responsible person, physical location.
Freeze any further manipulation (no re-trimming, no re-curing, no re-packaging) until the sample pull is complete.

Pre-sample inspection. Before the sampler arrives, do your own quick QC:

Visual inspection for mold, pests, foreign material — if you see it, your sampler will too.
Moisture check — outside 8-12% moisture content is a common microbial-fail trigger.
Packaging integrity — is the batch in the sampling container you committed to?
Metrc tag match — the tag on the batch in your vault must match the tag on the record submitted to the lab.

Scheduling the Sampler

ISO 17025-accredited labs send trained samplers to the operator's facility in most states (CA, CO, MA, FL, NV explicitly require lab-sampler pulls as of 2026-04; operator-drop-off is allowed in some markets but not preferred for chain-of-custody strength).
Typical scheduling lead time: 2-5 business days (as of 2026-04; varies by market and lab workload).
Rush sampling available for +$200-$500 per sampling event in most markets (as of 2026-04; Kaycha, SC Labs, and regional labs have documented rush fees in this range).
Standing sampling day. Operators doing >5 batches/week typically negotiate a weekly standing-sampling-day in the RFP (e.g., "every Wednesday, 10am-2pm") — removes scheduling friction, gets favorable pricing.
Sampler quality matters. When negotiating with a lab in the RFP stage (see lab-selection.md), ask about sampler training, turnover, and whether they rotate samplers to prevent familiarity bias. Samplers who have worked your facility for 2+ years develop useful pattern recognition on your batches; samplers who rotate quarterly keep chain-of-custody discipline sharp.
Out-of-state sampling. For multi-state operators, confirm the lab can sample in each state you operate. Some labs with broad ISO 17025 accreditation still lack state-specific sampling-license authorization in every market.

Sampler-Visit Scheduling Calendar

A mid-size operator (20-50 lb flower / month) typically schedules:

| Event | Frequency | Pre-Scheduling Buffer | |-------|-----------|----------------------| | Routine sampling visits | 1-2x per week | 3-5 business days | | Rush sampling (paid premium) | As needed | 24-48 hours (as of 2026-04) | | Retest sampling (after fail) | Per batch | Same as rush unless standing agreement | | Split-sampling to secondary lab | Quarterly or per critical batch | Scheduled with both labs | | Audit-support sampling (regulator inspection) | Event-driven | Immediate — regulator sets the clock |

(As of 2026-04; calendar recommendations synthesize typical operator patterns.)

Witnessing the Sample Pull

What the operator watches for:

Sampler is the named individual on the lab's sampler roster (check credentials).
Sampler uses sanitized tools between samples (no cross-contamination across batches).
Sample container is sealed + signed in the operator's presence.
Chain-of-custody form captures: sample date/time, sampler name, operator witness name, batch ID, sample weight, container seal number, transport conditions.
Any deviation from SOP (batch already unsealed, sample weight out-of-range, container not pre-labeled) is documented in the chain-of-custody form, NOT explained away verbally.

Common operator errors (avoid these):

Operator pulls their own sample to "save time" — voids chain-of-custody in most states.
Operator lets sampler pull from only one corner of the batch ("grab-sample") — non-representative; fails audit.
Operator does not countersign the chain-of-custody form — COA can later be challenged as lacking attestation.
Batch is unsealed before the sampler arrives — batch now has to be quarantined longer for re-sealing and re-sampling.

Post-Sample Batch Handling

Batch stays in quarantine from sample-pull through COA return. Do NOT begin pre-packaging or labeling against the batch yet — you don't know if it passes.
Metrc status: most states allow a "testing in progress" status. Don't skip this — it's the audit trail showing you didn't sell a batch before COA return.
What NOT to do during quarantine: (1) re-sample or re-pull material, (2) re-cure or re-trim, (3) re-package into retail containers, (4) move the batch out of the designated quarantine area. All of these break the chain-of-custody that the lab just established.
Retention. Retain the unused sample material (lab-retained portion) per state rule — typically 6-12 months (as of 2026-04; verify with state regulator). If a dispute or retest is needed, you'll want this material accessible.

What the Sampler Takes With Them

A typical lab sampler leaves with:

Sample material (15-30g for flower full panel; smaller for concentrates; follow the lab's specific SOP)
Signed chain-of-custody form (original; operator retains copy)
Metrc manifest linking the sample to the batch tag
Seal log (container seal numbers recorded in both the lab's and operator's copy of the chain-of-custody form)

Sampling Event SOP (One-Page Checklist)

A one-page operator-side checklist to run during every sampling event. Print this, laminate it, put it on the wall of the sampling-quarantine area.

[ ] Batch is in sampling-pending quarantine 24+ hours before sampler arrives.
[ ] Batch tag on physical container matches Metrc record.
[ ] Batch moisture is within 8-12% (document measurement).
[ ] Visual inspection clean (no visible mold, pests, foreign material).
[ ] Operator witness assigned and on-site at scheduled sampling time.
[ ] Sampler ID/credentials verified against lab roster.
[ ] Chain-of-custody form reviewed before sample pull.
[ ] Sample pull witnessed; seal numbers documented.
[ ] Operator counter-signs chain-of-custody form.
[ ] Operator copy of chain-of-custody retained in batch file.
[ ] Batch re-sealed + returned to sampling-pending quarantine.
[ ] Metrc status updated to "testing in progress."
[ ] Batch record updated with sample date, sampler name, lab destination.
[ ] Expected COA return date calculated and placed on receiving calendar.
[ ] Any observed deviations from SOP documented in batch file.

(As of 2026-04; customize for your specific lab's SOP variations.)

Sampling Frequency and Batch-Size Optimization

Operators often over-test or under-test because they haven't calibrated batch-size against sampling-event cost. Some frameworks as of 2026-04:

Full-compliance panel costs $350-550 regardless of batch size (as of 2026-04). Small batches (<3 lb) are cost-inefficient relative to panel cost. Consider: combining small same-cultivar sub-batches into one test batch if state rules allow.
Rush sampling-event fees apply per visit, not per batch. A standing weekly sampling day amortizes the per-visit fee across many batches.
Batch-split strategy. For 15+ lb batches near the state batch-size cap, consider splitting into two 7.5-lb sub-batches if it reduces fail-blast-radius (one fail = half the loss vs one fail = full loss). The trade-off is 2x panel cost.

Failure Decision Trees

When a batch fails, three questions drive the decision: (1) Is the failure real or within lab measurement uncertainty? (2) Is remediation allowed for this contaminant type + this state? (3) What is the economic threshold for remediate vs destroy vs re-source? This section walks each panel-type failure with a decision matrix.

Decision Matrix: Pesticide Panel Failure

Pesticide failures are the hardest to remediate. The underlying contaminant is a chemical molecule bound to the plant material — irradiation, ozone, UV-C, and even CO2 re-extraction don't remove pesticides (and re-extraction concentrates some pesticides rather than removing them). Most pesticide failures end in destroy or re-source.

| Factor | Retest | Remediate | Destroy | Re-Source | |--------|--------|-----------|---------|-----------| | Failure is marginal (within lab uncertainty, within 10% of action level) | PRIMARY OPTION | — | — | — | | Pesticide is on state prohibited list | — | NOT ALLOWED | PRIMARY | Consider | | Pesticide is on action-level list + within 2x action level | — | Possible via activated-carbon wash (limited evidence; verify with state regulator) | Consider | Consider | | Pesticide is on action-level list + >2x action level | — | — | PRIMARY | Consider | | Batch was sourced from third-party cultivator | — | — | Document | PRIMARY (supplier cost) | | Operator is vertically integrated (own cultivation) | — | Limited options | PRIMARY for clean-slate | Not applicable | | Batch is already packaged for retail | — | — | PRIMARY (can't re-sample packaged batch) | — |

(As of 2026-04; pesticide action levels set by state regulator. Verify with state DCC/MED/CCB equivalents.)

The retest question. Labs have measurement uncertainty (typically ±10-15% for pesticides per ISO 17025 SOPs). If a failure is within uncertainty of the action level, a retest at the SAME lab costs $150-225 (as of 2026-04) and clarifies the number. If the retest comes back clearly failing, you have a real failure. If the retest clears the action level, you have a borderline case — document both results in the batch record.

The "send it to the other lab" question. Labs can and do produce different numbers on the same batch (measurement variance + sample heterogeneity). Sending a retest sample to your secondary lab costs the same $150-225 but gives you independent verification. Do this when the primary-lab retest is still borderline.

Supplier cost recovery. If the batch was sourced from a third-party cultivator and fails pesticides, most wholesale contracts (as of 2026-04) have a "failed-COA return" clause — you return the batch to the cultivator, they bear the loss. Confirm your wholesale MSA has this clause before you buy.

Decision Matrix: Microbial Panel Failure

Microbial failures are the MOST remediable. X-ray irradiation, gamma irradiation, ozone, UV-C, and CO2 re-extraction all address microbial contaminants, with trade-offs on terpene preservation and cost. But remediation requires labeling in most states, which affects retail price.

| Factor | Retest | Remediate | Destroy | Re-Source | |--------|--------|-----------|---------|-----------| | Failure is marginal (within lab uncertainty) | PRIMARY OPTION | Consider | — | — | | Total Yeast + Mold (TYM) >10x action level | — | PRIMARY (X-ray) | Consider if batch <$10K wholesale | — | | E. coli or Salmonella detected | Retest FIRST (false positives) | Limited — heat-kill bacteria but safety risk lingers | PRIMARY | Consider | | Aspergillus species (state-specific action level) | Retest FIRST | PRIMARY (X-ray or gamma) | Consider | — | | State disallows labeled-remediated sale | — | Not an option | PRIMARY or re-extract to concentrate | Consider | | State allows labeled-remediated sale (CO post-2025-07-01, MA, others verify) | — | PRIMARY with "Decontaminated" label in CO | — | — | | Premium flagship SKU (terpene preservation critical) | — | Skip (quality hit) | Consider | — |

(As of 2026-04; microbial action levels set by state regulator. Verify with state DCC/MED/CCB equivalents.)

The terpene-preservation trade-off. Every irradiation method costs 10-30% of terpene profile (as of 2026-04; source: cannatechtoday.com industry synthesis). For a flagship $60/eighth SKU, that's a measurable quality hit — consumers can smell the difference. For a $20/eighth value SKU, the trade-off is less visible and remediation is more economic.

State allowance matters. In CO (as of 2025-07-01), labeled-remediated flower can be sold with the "Decontaminated" label. In MA, decontamination disclosure is required but the exact label phrase should be verified with CCC. In other states, some disallow labeled-remediated sale entirely, forcing a destroy-or-re-extract decision.

Decision Matrix: Heavy Metals Failure

Heavy metals CANNOT be remediated. Lead, cadmium, arsenic, and mercury are elements — they don't transform, irradiate away, or extract-out cleanly. A heavy-metals fail is almost always a destroy-or-re-source outcome.

| Factor | Retest | Remediate | Destroy | Re-Source | |--------|--------|-----------|---------|-----------| | Failure is marginal (within lab uncertainty) | PRIMARY OPTION | — | — | — | | Failure is >2x action level | — | NOT POSSIBLE | PRIMARY | Consider if recurring | | Heavy metals from grow medium (soil, water) | — | NOT POSSIBLE | Destroy; fix source | Re-source grow inputs | | Heavy metals from packaging (leaching) | — | — | Destroy; fix packaging | Re-source packaging | | Operator is vertically integrated | — | — | PRIMARY | Re-source inputs (long-term) |

(As of 2026-04; heavy-metal action levels set by state regulator. Verify with state DCC/MED/CCB equivalents. Heavy metals from soil contamination require agronomic remediation — not batch remediation.)

The "upstream" question. A heavy-metals fail is almost always an input-chain problem — contaminated soil, contaminated water, or contaminated packaging. Fixing the batch doesn't fix the underlying input. If you see a second heavy-metals fail within 90 days on similar cultivars, stop the clock: run heavy-metals on your water source, your soil, your nutrients, and your packaging adhesive. Fix the source, then you can resume normal testing. This is a rare Rule-4 architectural problem for the cultivation side, not a batch-management problem.

Decision Matrix: Mycotoxin Failure

Mycotoxins are secondary metabolites of mold (aflatoxin, ochratoxin). They survive irradiation. A mycotoxin fail is effectively equivalent to a heavy-metals fail — destroy or re-source.

| Factor | Retest | Remediate | Destroy | Re-Source | |--------|--------|-----------|---------|-----------| | Failure is marginal (within lab uncertainty) | PRIMARY OPTION | — | — | — | | Mycotoxin detected above state action level | — | NOT POSSIBLE (irradiation kills mold but leaves toxins behind) | PRIMARY | Consider | | Recurring mycotoxin fails in similar batches | — | — | Destroy + cure process review | Re-source cure inputs |

(As of 2026-04; mycotoxin action levels set by state regulator. Verify with state DCC/MED/CCB equivalents.)

Decision Matrix: Residual Solvents Failure (Concentrates Only)

Residual solvents are leftover extraction solvents (butane, propane, ethanol, hexane, etc.). They fail because of under-purging during concentrate production. Re-purging is the standard remediation.

| Factor | Retest | Remediate | Destroy | Re-Source | |--------|--------|-----------|---------|-----------| | Failure is marginal (within lab uncertainty) | PRIMARY OPTION | — | — | — | | Failure is >2x action level | — | PRIMARY via additional vacuum-purge cycle | — | — | | Recurring residual-solvents fails | — | — | Destroy; fix extraction SOP | — | | Concentrate type is live-resin/live-rosin (low-solvent path) | — | PRIMARY via re-purge | Consider | — |

(As of 2026-04; residual-solvent action levels set by state regulator.)

Decision Matrix: Potency Misreport

Potency "failures" are usually over-label (THC% comes back below label claim) or under-label variance (THC% comes back above label claim). State rules around labeling tolerance vary.

| Factor | Retest | Remediate | Destroy | Re-Source | |--------|--------|-----------|---------|-----------| | Variance within state labeling tolerance (typically ±10%) | No action | — | — | — | | Variance outside tolerance; THC% below label | Retest at secondary lab | Relabel (lower THC claim) | Consider if unsellable | — | | Variance outside tolerance; THC% above label | Retest at secondary lab | Relabel (higher THC claim) | — | — | | Large systematic variance suggests lab issue | — | — | — | Reconsider primary lab (multi-lab strategy) |

(As of 2026-04; state labeling tolerance typically ±10-15%. Verify with state regulator.)

Generic Workflow After Any Failure

Regardless of the specific panel type, the first-hour operator workflow is identical:

Confirm the fail is real. Read the COA side-by-side with the state action levels table (see coa-testing.md §State-by-State Testing Requirements Matrix). A misread COA trigger is embarrassing and avoidable.
Quarantine the batch physically and in Metrc. Update status to "failed-testing" or equivalent in your state's seed-to-sale system.
Notify compliance manager and GM. No unilateral remediate-vs-destroy decisions at the inventory-manager level.
Capture the timeline. Note: COA receipt date + time, internal notification time, quarantine time, Metrc status change time. This becomes the audit trail if the failure escalates to regulator notice.
Trigger the decision tree appropriate to the panel type (above sections).

When to Bring in a Lawyer

Most batch failures are routine operational events. Some are not. Call your cannabis counsel when:

A failure is linked to a state regulator inspection or consumer complaint (as of 2026-04; regulator-triggered failures can escalate fast).
The failure involves a labeling claim (e.g., "organic" or "solventless" claims that now appear contradicted by lab results).
The failure is part of a pattern across multiple batches that might trigger regulator scrutiny.
A wholesale customer is disputing the failure and demanding remediation liability coverage.
The failure involves a contamination source that crosses supplier boundaries (did the input come from a vendor who sold the same input to other operators?).

Legal involvement costs $500-2,500 per failure event for a 1-2 hour consultation (as of 2026-04; mid-market cannabis counsel rates); cheaper than a botched escalation.

Operator-Level vs Enterprise-Level Failure Response

For single-store operators, failure response is a GM + inventory manager conversation. For MSO enterprise operators (5+ locations), failures escalate differently:

| Operator Type | Failure Response Path (as of 2026-04) | |---------------|------------------------------------------| | Single-location | Inventory manager → compliance manager → GM, decision within 24hr | | 2-5 location regional | Inventory manager → compliance manager → regional ops → legal, decision within 48hr | | MSO enterprise | Local inventory → regional ops → corporate compliance → corporate legal → corporate recall committee, decision within 72hr (regulatory minimums may compress this) |

The compressed-timeline pressure on MSO response comes from the regulator-notification clock; some states require 24hr notification for adverse events. Treat the escalation path as pre-defined, not invented-on-the-spot.

Remediation Methods

All remediation methods carry terpene-degradation and cost trade-offs. Read §Remediation Decision Tree below for the economic-threshold framework. Figures below synthesize industry sources (as of 2026-04; cannatechtoday.com, xrpure.com, airrosshield.com); confirm current pricing directly with active vendors.

X-Ray Irradiation

Contaminant coverage: Microbial (bacteria, yeast, mold) at DNA level. NOT effective for pesticides, heavy metals, or mycotoxins.
Efficacy: Best penetration of any method; effective on dense bud cores. Industry synthesis as of 2026-04: DNA-level inactivation of microbes across the full bud interior.
Cost: ~$3-8/lb (as of 2026-04; verify with active vendors like XRpure).
Terpene impact: 10-30% terpene degradation typical (as of 2026-04; source: cannatechtoday.com). Quality-grade operators may reject irradiated product for flagship SKUs.
Labeling implication: In CO (effective 2025-07-01), MA, and others, requires the state-specific label phrase. In CO the exact phrase is "Decontaminated" on every container of treated Regulated Marijuana. See §State Remediation Rules.
Typical turnaround: 24-72hr for a batch through a service-bureau X-ray unit (as of 2026-04).
Vendors (as of 2026-04): XRpure (primary US installer), regional decontamination service bureaus. (Verify currently operating, 2026-04.)
When it's the right choice: premium or value flower with deep-core microbial contamination, state allows labeled sale, batch wholesale value >$10K.

Gamma Irradiation

Contaminant coverage: Microbial (bacteria, yeast, mold). NOT effective for pesticides, heavy metals, or mycotoxins. (Same mechanism as X-ray at higher-energy band.)
Efficacy: Effective at DNA-level inactivation; isotope-based source (Cobalt-60 or similar) rather than electronic X-ray tube.
Cost: ~$2-6/lb (as of 2026-04; gamma-service bureaus are fewer than X-ray but cheaper per lb where available).
Terpene impact: 10-30% terpene degradation (as of 2026-04; similar to X-ray).
Labeling implication: Same as X-ray in states that require "Decontaminated"-style labeling.
Typical turnaround: 48-120hr (as of 2026-04; depends on gamma-facility batch scheduling).
Vendors: Regional gamma-irradiation service bureaus (historically food and medical-device applications; cannabis-specific vendors are growing as of 2026-04).
When it's the right choice: value flower, microbial fail, cost-sensitive remediation, state allows labeled sale. Declining vs X-ray for most operators due to logistics (isotope licensing, transport restrictions).
Note: Gamma is isotope-based and declining vs electronic X-ray. Most new deployments as of 2026-04 are X-ray, not gamma.

Ozone Treatment

Contaminant coverage: Surface bacteria/mold. NOT effective for deep-bud-core microbes, pesticides, heavy metals, or mycotoxins.
Efficacy: Weak bud penetration; effective on surface contaminants only. Fails to eliminate deep mold spores in dense indoor flower (as of 2026-04; synthesis of industry reports).
Cost: ~$1-4/lb (as of 2026-04; cheap because it's a gas treatment with minimal capital equipment).
Terpene impact: Mild (as of 2026-04; less than irradiation but not zero).
Labeling implication: Check state remediation rules. Some states treat ozone as remediation requiring disclosure; others treat it as a routine cure process.
Typical turnaround: 4-12hr (as of 2026-04; fast; can be done in-house with capital investment).
Vendors: Airross Shield and similar ozone-treatment cabinet vendors (as of 2026-04; verify current availability).
When it's the right choice: borderline microbial fails, surface contamination (light mildew), loose-cure flower where ozone can penetrate. Do NOT rely on ozone for dense bud cores or deep mold.

UV-C Treatment

Contaminant coverage: Surface microbial. NOT effective for deep-bud-core microbes, pesticides, heavy metals, or mycotoxins.
Efficacy: Low penetration; surface-only. UV-C light doesn't penetrate plant material deeply.
Cost: Low — typical range $0.50-$2.00/lb (as of 2026-04; mostly equipment amortization).
Terpene impact: Minimal (as of 2026-04).
Labeling implication: Rarely triggered as "remediation" by state regulators; generally treated as a cure-process supplement.
Typical turnaround: 2-8hr (as of 2026-04).
Vendors: UV-C cabinet manufacturers (food-safety crossover). As of 2026-04, several cannabis-specific vendors exist; verify.
When it's the right choice: surface mildew prevention, trimming-room sanitation, low-value trim/shake, quick sanitation-pass before packaging. Do NOT rely on UV-C alone for any documented microbial fail above action level.

CO2 / Supercritical Re-Extraction

Contaminant coverage: Concentrate feedstock path — extracts clean cannabinoids, leaves contaminants behind in the post-extraction biomass. Effectively "remediates" the contaminated material by redirecting it out of the flower market into the concentrate market.
Efficacy: Very effective for microbial-fail flower converted to concentrate (as of 2026-04). Less effective for pesticide-fail flower because some pesticides ARE lipid-soluble and co-extract into the concentrate.
Cost: ~$10-30/lb feedstock (as of 2026-04; includes extraction cost + concentrate processing).
Terpene impact: Drives to concentrate; not a flower-preservation method. Terpenes survive in the concentrate but the flower is gone.
Labeling implication: Check state rules on remediated-feedstock concentrate. Some states require labeling disclosure on the resulting concentrate; others treat it as normal extract.
Typical turnaround: 3-10 business days including extraction + concentrate post-processing + retest (as of 2026-04).
Vendors: Most in-house concentrate producers can do this; contract extractors also common.
When it's the right choice: (a) failed batch has concentrate-grade value per lb of derived concentrate; (b) state disallows labeled-remediated flower sale; (c) operator has own extraction capability or a trusted contract extractor; (d) pesticide-fail batch where the pesticide is NOT lipid-soluble (verify with lab).

Beta Radiation

Contaminant coverage: Microbial (bacteria, yeast, mold).
Efficacy: Varies by dose; generally less penetration than X-ray or gamma but sufficient for surface-plus-shallow contamination (as of 2026-04).
Cost: ~$2-5/lb (as of 2026-04).
Terpene impact: Moderate — typically 10-20% terpene degradation (as of 2026-04).
Labeling implication: Same as X-ray/gamma in states requiring "Decontaminated"-style labeling.
Typical turnaround: 48-72hr (as of 2026-04).
Vendors: Smaller vendor pool than X-ray; regional specialists as of 2026-04.
When it's the right choice: mid-cost alternative to X-ray; appropriate for looser-cure flower where penetration requirements are moderate.

Method Selection Decision Matrix

| Your Situation | First-Choice Method | Avoid | |----------------|--------------------|--------| | Premium flower, microbial fail | Try ozone or UV-C first (minimal terpene loss), then X-ray if needed | Gamma (isotope-based; declining infrastructure) | | Value flower, microbial fail | X-ray | Ozone alone (spores survive) | | Dense indoor flower, deep mold | X-ray (penetration); consider CO2 re-extraction path | UV-C (surface-only) | | Concentrate feedstock contaminated | CO2 / re-extraction | Irradiation (unnecessary step) | | Pesticide or heavy-metal fail | NONE — irradiation does not affect chemical contaminants | Any irradiation (waste of money) | | Borderline TYM (just above action level) | Ozone + retest | X-ray (over-treatment; terpene hit) | | Surface mildew on a clean-grown batch | UV-C + retest | X-ray (over-treatment) | | State disallows labeled-remediated sale | CO2 re-extract to concentrate | X-ray, gamma, ozone (flower path blocked) |

(As of 2026-04; validate with current state regulator rules before committing to a method.)

Method Selection — Cost Comparison Table

Cost per lb of treated flower (service-bureau pricing as of 2026-04) compared across methods. Pricing excludes retest fees ($350-550 full panel).

| Method | Cost/lb Low | Cost/lb High | Best For | |--------|-------------|--------------|----------| | UV-C | $0.50 | $2.00 | Surface mildew; sanitation-pass | | Ozone | $1 | $4 | Borderline microbial; surface | | Beta radiation | $2 | $5 | Mid-cost microbial remediation | | Gamma irradiation | $2 | $6 | Value microbial remediation | | X-ray irradiation | $3 | $8 | Deep-bud-core microbial | | CO2 re-extraction | $10 | $30 | Contaminated → concentrate path |

(As of 2026-04; validate with active vendor quotes. Pricing moved ~5-10% year-over-year across methods.)

Common Operator Mistakes in Method Selection

Defaulting to X-ray for every microbial fail. X-ray is the most aggressive option; use it when you need penetration, not reflexively. Ozone or UV-C may suffice for borderline surface contamination.
Trying ozone on dense indoor bud cores. Ozone doesn't penetrate dense flower. If the microbial fail came from core rot, ozone retest is wasted spend.
Skipping retest after remediation. Remediation without retest = still-failed batch. Budget the $350-550 retest cost into the remediation decision (as of 2026-04).
Irradiating pesticide-fail batches. Irradiation does nothing for chemical contaminants. This is a costly mistake operators make when in panic mode after a pesticide fail.
Not confirming state-specific labeling rule before remediating. Remediating a batch and discovering later that your state doesn't allow labeled-remediated sale is a full write-off plus the remediation cost. Verify FIRST.

State Remediation Rules

This table summarizes state remediation and labeling rules as of 2026-04. Every row below requires regulator verification at the point of decision. The only row verified against primary regulator source is CO (sbg.colorado.gov/med). Everything else is hand-authored from the same data that seeds the states.remediation_rules JSON column — verify against the state regulator before acting.

| State | Remediation Allowed | Label Required | Exact Phrase | Effective Date | Source | Last Verified | |-------|---------------------|----------------|--------------|----------------|--------|---------------| | CO | Yes | Yes | Decontaminated | 2025-07-01 | sbg.colorado.gov/med | 2026-04-17 | | MA | Yes | Yes | Decontamination disclosure required; verify exact phrase with CCC | — | masscannabiscontrol.com | 2026-04-17 | | CA | Yes | Verify | CA DCC consumer-safety rules apply; specific label language pending verification | — | cannabis.ca.gov | 2026-04-17 | | NV | Verify | Verify | — | — | ccb.nv.gov | 2026-04-17 | | IL | Verify | Verify | — | — | idfpr.illinois.gov | 2026-04-17 | | NY | Verify | Verify | — | — | cannabis.ny.gov | 2026-04-17 | | NJ | Verify | Verify | — | — | nj.gov/cannabis | 2026-04-17 | | OR | Verify | Verify | — | — | oregon.gov/olcc | 2026-04-17 | | WA | Verify | Verify | — | — | lcb.wa.gov | 2026-04-17 | | FL | Verify | Verify | — | — | floridahealth.gov/mmu | 2026-04-17 | | MI | Verify | Verify | — | — | michigan.gov/cra | 2026-04-17 |

What the CO 2025-07-01 Rule Requires

Exact phrase: "Decontaminated" label on every container of decontaminated Regulated Marijuana (flower, shake, trim, pre-rolled, infused pre-rolled).
Effective: 2025-07-01.
Source: sbg.colorado.gov/med (Colorado Marijuana Enforcement Division). (As of 2026-04; verify current compliance notice at sbg.colorado.gov/med.)
Implication: Operators with existing pre-07-01 decontaminated inventory had a transition window; post-07-01 inventory must be relabeled before retail sale. The exact phrase matters — state audit has cited operators for "Remediated" or "Treated" labels that did not match the "Decontaminated" required phrase.
Scope: Applies to flower, shake, trim, pre-rolled, infused pre-rolled — i.e., all plant-material SKUs. Edibles and concentrates have separate rules; verify with CO MED.

MA Decontamination Disclosure Context

As of 2026-04, MA requires decontamination disclosure on treated Regulated Marijuana. The exact phrase varies by CCC guidance generation and should be verified against current CCC rules.
Source: masscannabiscontrol.com (as of 2026-04).
Implication: MA has had decontamination-disclosure-style rules longer than CO; operators migrating from MA to CO or vice versa should NOT assume the same label phrase.

CA Verify-Path Context

CA DCC has consumer-safety rules (Bus. & Prof. Code § 26000 et seq., CCR Title 4 Division 19) that touch remediated-product labeling indirectly. As of 2026-04, the specific required label phrase for remediated-sale product is pending further verification; retailers should verify with CA DCC or their CA cannabis counsel before putting remediated flower on the shelf with a CA-specific label phrase.
Source: cannabis.ca.gov (as of 2026-04).

States Marked "Verify"

The "Verify" entries above indicate one of three situations (as of 2026-04):

Rule is emerging — the state is actively drafting rules that may change quarterly.
Rule exists but is not easily accessible — regulator documentation is scattered across bulletins, FAQs, and emergency rules. Operator must call the regulator directly.
Rule depends on other factors — medical vs adult-use, wholesale vs retail, specific product category.

For every "Verify" entry, the operator's minimum-sufficient action is: (a) query the state regulator's compliance bulletin feed as of this month, (b) document the finding with date + URL, (c) act on the finding. Don't act on last-quarter's interpretation.

Operator Workflow for a Labeled-Remediated Batch in CO

Batch fails microbial panel at D10.
Operator selects X-ray remediation (Option B from Decision Tree).
Batch goes to X-ray service bureau; +$150-400 service cost for 50 lb (as of 2026-04).
Batch returned; lab re-samples per state rule; retest $350-550 full panel (as of 2026-04).
Retest clears all panels.
Operator updates all retail packaging to include "Decontaminated" label per state rule (effective 2025-07-01).
Batch released to retail shelf; typical 20-30% markdown vs non-remediated flower in same category.
Audit-record stored in batch file: failure COA, remediation service receipt, retest COA, packaging-update log, Metrc status updates.

State Rule Refresh Cadence

Recommended cadence for re-verifying state remediation rules (as of 2026-04):

| Action | Frequency | Owner | |--------|-----------|-------| | Check state regulator's compliance-bulletin feed | Monthly | Compliance manager | | Attend quarterly state regulator stakeholder meetings (virtual) | Quarterly | Compliance manager + legal | | Review labeled-remediated SKU sales data for brand-drag signal | Quarterly | Inventory manager + marketing | | Re-run "verify with state regulator" calls on all rows marked "Verify" | Annually | Compliance manager | | Re-validate primary source URLs in state-rules table | Annually | Compliance manager | | Update cross-link tables in operator SOP manuals | Annually | Compliance manager |

(As of 2026-04; annual cadence is a minimum — some states change rules quarterly and require tighter tracking.)

Operator FAQ on State Rules

Q: Do I need to relabel a batch that was remediated before 2025-07-01 in CO? A: As of 2026-04, verify with CO MED. The transition window for pre-07-01 remediated inventory closed; post-07-01 all decontaminated product requires the "Decontaminated" label.

Q: Can I sell labeled-remediated CO product in other states via wholesale? A: Only if the receiving state also allows labeled-remediated sale AND your interstate transfer complies with that state's import rules. As of 2026-04, interstate cannabis transfer is federally prohibited in all markets — this is a wholesale-within-state question only, unless you're in the OR-WA pilot interstate agreement (verify current status).

Q: Does remediation have to happen at a specific type of facility? A: State-specific. Some states require remediation at a state-licensed facility; others allow any ISO-compliant service bureau. Verify with your state regulator.

Q: What happens if I mislabel a remediated batch? A: Labeling violations range from warning letters to fines to license action. As of 2026-04, CO MED has issued warning letters for "Remediated" and "Treated" labels that don't match the state-required "Decontaminated" phrase. The cost of mislabeling is much higher than the cost of doing it right.

Remediation Decision Tree

When a batch fails, the economic decision has three commercial paths: (A) destroy the batch, (B) remediate + relabel + sell at a markdown, (C) re-extract to concentrate. This section walks the math with a worked example.

The Worked Example

Scenario: A 50-lb batch of indoor flower fails total yeast + mold at 75,000 CFU/g (action level 10,000 CFU/g). Batch wholesale value pre-failure: ~$45,000 at $900/lb (as of 2026-04; indoor premium market).

Option A — Destroy.

Destruction fee: $100-500 for regulated cannabis destruction service (as of 2026-04).
Wholesale loss: $45,000 (batch reserve write-off).
Metrc reconciliation: ~2 hours compliance labor.
Net recovery: $0, with $500 destruction cost + $200 labor cost = ($700) loss + $45,000 inventory write-off.

Option B — Remediate (X-ray) + Relabel.

X-ray cost: $3-8/lb × 50 lbs = $150-400 (as of 2026-04).
Sampling + retest: $350-550 (full panel, as of 2026-04).
"Decontaminated" label packaging rework in CO: ~$2,000 operational cost (labor + new label stock + Metrc status updates).
Expected retail markdown: 20-30% discount vs non-remediated flower (as of 2026-04; market synthesis).
Expected recovery: 70-80% of original wholesale value = $31,500-$36,000 on the batch.
Net: recover $27,500-$33,000 after remediation + retest + packaging rework costs, vs $0 destroy recovery.

Option C — CO2 Re-Extract to Concentrate.

Extraction cost: $10-30/lb × 50 lbs = $500-1,500 (as of 2026-04).
Concentrate yield: ~15-25% of flower weight = 7.5-12.5 lbs concentrate.
Concentrate wholesale value: $1,000-$2,500/lb → $7,500-$31,250 (as of 2026-04; wide range by concentrate type).
Sampling + retest on concentrate: $350-550 (as of 2026-04).
Net: recover $5,650-$29,200 after extraction + retest costs, vs $0 destroy recovery. Exits flower market for this batch.

Decision framework for this scenario:

If operator is in CO (labeled remediation allowed): Option B is typically best economically.
If operator is in a state that disallows labeled-remediated sale: Option C is the path.
If batch has pre-existing quality issues beyond the microbial fail (e.g., thin density, poor trim): Option C regardless of state.
If batch is very small (<10 lbs) and remediation fees dominate: Option A may be most efficient.

Decision Matrix

| Factor | Remediate + Relabel | Re-Extract to Concentrate | Destroy | |--------|--------------------|-----------------------------|---------| | Batch wholesale value > $10K | Favor | Favor if premium SKU | — | | Batch wholesale value < $5K | Consider (fixed costs eat margin) | Consider | PRIMARY | | State allows labeled-remediated sale | Required path | Optional | — | | State bans labeled-remediated sale | Not an option | Required path | Option | | Batch is regulator-seized | — | — | Required | | Terpene preservation matters (flagship SKU) | Skip (quality hit) | Skip (flower destroyed) | Consider | | Pesticide fail (not microbial) | Not effective | Not effective for lipid-soluble pesticides | PRIMARY | | Heavy-metals fail | Not possible | Not possible | PRIMARY | | Operator has own extraction or trusted extractor | — | Favored | — |

Economic Break-Even Points (as of 2026-04)

X-ray remediation break-even: $500-2,500 per batch (batch wholesale value below this — destroy; above this — remediate).
Ozone / UV-C break-even: Near zero on cost side; but efficacy is the constraint, not cost.
CO2 re-extraction break-even: $3,000-7,500 batch wholesale value (below this — destroy; above — re-extract).
These numbers move with service-bureau pricing. Revalidate your break-evens quarterly against your active vendor quotes (as of 2026-04; pricing moved ~5-10% year-over-year).

Don't-Overlook Second-Order Costs

Packaging rework labor. A "Decontaminated" relabel in CO is not just new labels — it's re-packaging, re-Metrc-tagging, re-training retail staff on the SKU. Budget $500-3,000 per batch depending on SKU count.
Retail-floor education. Budtenders need scripts for the "Why is this discounted?" question. Treat remediation-disclosure training as part of your quarterly training cycle.
Brand reputation drag. If 20% of your SKUs are labeled-remediated, your brand signals "quality-issues-frequent" to customers. Keep remediation below 10% of output if possible; if it's above, fix the cultivation pipeline upstream.
Wholesale-channel pushback. Some wholesale buyers refuse labeled-remediated product regardless of state legality. Confirm wholesale-channel acceptance before remediating large batches intended for wholesale.
Metrc reconciliation labor. Every remediation status change requires a Metrc transaction (batch hold, remediation event, retest event, status return to sellable). Per-batch compliance labor is 1-3 hours not including the physical remediation.
Opportunity cost of the vault slot. A batch in remediation occupies vault space that could hold a clean new batch. For high-throughput operators, this is a 10-20 day vault-slot opportunity cost worth accounting for in the decision.

Three-State Decision Framework (Comparative)

Compare how the same failure plays out across three states (as of 2026-04):

| Step | CO (labeled-remediated allowed) | MA (labeled-remediated allowed) | CA (verify) | |------|----------------------------------|----------------------------------|-------------| | 1. Batch fails microbial | Quarantine; notify compliance | Quarantine; notify compliance | Quarantine; notify compliance | | 2. Remediation allowed? | Yes, with "Decontaminated" label | Yes, with CCC-specified disclosure | Verify with CA DCC | | 3. Service bureau available | Yes (XRpure active in CO) | Yes (regional vendors) | Yes (multiple) | | 4. Retest cost | $350-550 | $350-550 | $350-550 | | 5. Labeled retail path | Yes (markdown ~20-30%) | Yes (markdown varies) | Verify before committing | | 6. Concentrate re-extract path | Yes | Yes | Yes | | 7. Wholesale acceptance | Variable | Variable | Variable |

(As of 2026-04; verify state-specific rules. Every "verify" is an operator responsibility, not a static fact.)

Recall Protocols

A cannabis recall is an operator's nightmare scenario — contaminated product has already left the shelf and is in consumer hands. The playbook is: (1) stop further distribution immediately, (2) notify the regulator within the state-required window, (3) identify the at-risk product downstream, (4) execute consumer notification per regulator rule, (5) refund/replace, (6) post-mortem and upstream fix. This section walks the workflow end-to-end.

When You Must Recall

State regulator-initiated recall. Triggered when the state regulator (CA DCC, CO MED, MA CCC, etc.) identifies a batch as contaminated based on:

Post-sale testing discovering a contamination source missed during initial testing.
Adverse event reports from consumers (microbial sickness, pesticide exposure).
Cross-batch contamination discovered at a sister operator or a lab.
Regulator audit discovering a testing fraud or lab-data falsification event.

Voluntary recall. Operator-initiated when internal QC or post-hoc testing reveals:

A contamination pattern not caught in pre-release testing (e.g., late-developing mold in storage).
Packaging or labeling error that could mislead consumers (e.g., THC % overstated).
Supplier-identified contamination in a shared input (grow medium, cure input, packaging adhesive).

Timeline expectations (as of 2026-04):

CA DCC: regulator-initiated recalls require operator response within 24 hours of notice; voluntary recalls require 48-hour regulator notification.
CO MED: similar timeline; verify with MED compliance bulletin.
MA CCC: similar timeline; verify with CCC.
Other states: verify with state regulator.

Notification Workflow

Step-by-step, what the operator does once a recall is triggered:

Internal trigger. QA/compliance detects the issue (internal testing, consumer complaint, regulator notice).
Immediate quarantine. Pull all remaining inventory of affected batch(es) from all sales channels within 1 hour of trigger. Update Metrc status to "recall-hold" on all related tags.
Regulator notification. Contact state regulator via the published recall-notification channel (typically an email + phone call). Document: batch ID, failure nature, quantity distributed, quantity recovered, corrective actions.
Customer notification. Per regulator rule, notify consumers through:
- Email (if loyalty-program data available)
- SMS (if opt-in available)
- In-store signage at affected dispensaries
- Public-facing website notice
- Social media post
Retailer notification (for wholesale product). Contact all downstream dispensaries within 24 hours; provide batch tags and recall-hold instructions.
Public-health notification (if severe). Some states require public-health department notification for severe cases (E. coli, Salmonella positive, pesticide overage).
Refund/replace. Customer-service flow: full refund + optional replacement with non-recalled SKU. Document every refund.
Post-mortem. Root-cause analysis within 30 days: what broke? How do we prevent it next time?

Documentation Requirements

Audit trail an operator should retain for every recall (state-required retention typically 3-5 years as of 2026-04):

Original COA (the pre-release "pass" that turned out to be wrong — or missing)
Triggering event record (internal testing result, consumer complaint, regulator notice)
Chain-of-custody records for the affected batch
Metrc records showing status changes (testing → released → recall-hold)
Regulator notification log (email + phone call records)
Customer notifications (copies of every email, SMS, signage photo)
Retailer notifications (for wholesale)
Refund/replacement log (consumer-level)
Disposal certificates for destroyed product
Root-cause analysis document
Corrective-action plan + completion evidence

Sample Recall Scenarios

Scenario 1: Post-release microbial discovery. A batch of flower released to retail passes initial testing at the release lab, but a competing lab tests the same batch (via secondary-market purchase) two weeks later and finds elevated TYM. The operator is notified via the competing lab's cross-lab QC program.

Actions:

Quarantine remaining inventory immediately.
Notify state regulator within 24 hours.
Request retest from primary lab at operator cost.
Based on retest result: proceed with full recall or confirm the original result stands.

Scenario 2: Adverse event report. A consumer reports illness after consuming a specific SKU from a specific batch. The state department of health investigates.

Actions:

Contact state regulator immediately.
Pull all remaining inventory of the batch pending investigation.
Cooperate fully with the investigation (do NOT destroy evidence; retain lab-sample material per state retention rules).
If investigation confirms contamination: execute full recall per above workflow.

Scenario 3: Labeling error (THC overstated). Operator discovers that a batch shipped with label claim of 28% THC when actual potency COA was 22% THC.

Actions:

Notify state regulator; this is a labeling-compliance recall, not a contamination recall.
Pull remaining inventory; relabel with correct potency.
Consumer refund path: typically only required for consumers who can prove specific-batch purchase; regulator-specific.

What NOT to Do

Don't hide the problem. Cannabis recall cover-ups have triggered license revocations in every major market (as of 2026-04; verify with state regulator enforcement logs).
Don't assume a borderline result is a "safe release." If the COA is borderline and you're uncertain, retest before release — a post-release retest that fails is exponentially more expensive than a pre-release retest.
Don't destroy sample-retained material before the regulator clears the investigation. State retention rules require you to preserve the un-analyzed portion.
Don't over-communicate to consumers beyond what the regulator approves. Some state rules are very specific about recall-notice wording; off-script notices can amplify confusion.
Don't skip the root-cause analysis. A recall without a corrective-action plan repeats within 12 months (operator pattern as of 2026-04; anecdotal but widespread across published enforcement actions).
Don't delegate regulator communications to your PR firm. Regulators want to talk to your compliance manager or your legal counsel, not a PR spokesperson. PR handles consumer messaging; compliance handles regulator messaging. Separate channels.

Recall Command Center Setup

For operators >$10M annual revenue, pre-build a recall command center so that Day 0 of a recall is execution, not improvisation. Elements (as of 2026-04):

Designated recall lead (typically VP Compliance or Chief Compliance Officer). One name, one phone number, 24/7 available for 96 hours from trigger.
Pre-drafted regulator notification template. Fill-in-the-blanks with batch ID, failure nature, distribution quantity. Saves 2-4 hours on Day 0.
Pre-drafted consumer notification template (for email + SMS + website). Run through legal pre-approval; have ready-to-deploy versions by state.
Retailer notification list (for wholesale operators). Master contact sheet with every wholesale account, primary + backup contacts, SLA for acknowledgment.
Disposal vendor contract pre-negotiated. In a recall you need disposal within 24-72 hours; negotiating the disposal vendor at trigger time is too slow.
Legal retainer active. Not a "we'll call when we need you" relationship — retainer with guaranteed 4-hour response SLA on recall-grade events.
PR firm briefed on cannabis recall protocols. Not every PR firm understands that cannabis regulator messaging is different from FDA food-recall messaging.
Customer-service script library. Budtenders and call-center staff need instant-access scripts for "I bought product X last week, should I be worried?" — not improv.

Recall Cost Modeling (as of 2026-04)

A full-scale recall event costs well beyond the product write-off. Industry synthesis (as of 2026-04):

| Cost Component | Typical Range | Driver | |----------------|---------------|--------| | Product write-off | Full batch wholesale value | Inventory destruction | | Disposal fees | $500-5,000 | Regulated destruction | | Customer refunds | 100% + replacement cost | Consumer-notification compliance | | Compliance labor | $5,000-25,000 | Audit trail, regulator communications | | Legal fees | $10,000-50,000 | 1-2 weeks counsel time | | PR / communications | $15,000-75,000 | Consumer-notification execution | | Regulator fines | $0-500,000 | State-specific; CA DCC has levied >$100K on major recalls | | Long-term brand drag | Unquantified but real | Customer trust recovery takes 6-18 months |

(As of 2026-04; recall cost is highly variable. A $50K inventory loss can turn into a $200K total-cost event once labor, legal, and fines are accounted for.)

COA-to-Product Reconciliation Workflow

Every batch on your shelf should have a valid COA filed in the batch record, with Metrc tag matching, within the state-required time window (typically 6-12 months from COA date as of 2026-04; verify with state regulator). The reconciliation workflow catches three common failures: missing COA, COA-mismatch, and expired COA.

Daily COA Reconciliation Routine

For the inventory manager or compliance manager (15-30 minutes per morning in a mid-size operation):

Pull today's inbound receipt list from your ERP/inventory system. For each inbound batch:
- Confirm Metrc tag on physical package matches ERP record.
- Confirm COA PDF is on file (operator portal or lab email).
- Confirm COA batch ID matches Metrc tag.
- Confirm COA panel results are within state-required limits.
- File COA PDF in batch record with indexed metadata (tag ID, COA date, lab name, pass/fail).
Run a pre-sales-floor check on yesterday's releases:
- For each batch released to sales floor, verify COA is accessible within 30 seconds (staff-accessible URL or physical file).
- Verify retail-package label carries required COA reference (state-specific; as of 2026-04 varies by state).
Weekly sweep: for batches in storage >30 days, verify COA hasn't expired per state rule (CA flower COAs are ~90 days; others vary as of 2026-04).
Monthly audit: sample 20 batches at random; verify complete paper trail from intake → COA → retail-release → sales. Document any gaps.

Common Failures and How to Catch Them

Missing COA for a Metrc-tagged batch. The batch can't legally be sold; quarantine immediately. Escalate to compliance manager. If the batch is already on the sales floor, this is a recall scenario.
COA batch ID mismatches product label. Relabel or destroy. If already sold, this is a recall scenario for the affected units.
COA says "remediated" but product packaging doesn't carry the state-required label. Labeling violation — pull inventory, relabel, escalate to compliance. As of 2026-04, CO "Decontaminated" labeling enforcement has been active since 2025-07-01.
COA expired (state-specific; CA is ~90 days for flower as of 2026-04). Re-test required. If the product is still sellable, budget the retest cost into your inventory-aging model.
COA exists but panels are incomplete (e.g., missing microbial panel on a full-panel-required state). Contact lab for supplemental result; quarantine batch until complete.
COA lab is not on state-approved list. This is a critical compliance failure — batch must be retested at a state-approved lab before any sale.

Building the Reconciliation System

Integration-first approach. If your lab, your ERP, and your Metrc-integrated seed-to-sale platform are all integrated via API (as of 2026-04 this is available with SC Labs, Kaycha, Confident Cannabis, and major platforms like Dutchie and BioTrack), 80% of reconciliation is automated — you only review exceptions.
Manual-audit fallback. If integrations aren't in place, build a daily spreadsheet with tab-per-batch; manual COA review takes 3-5 minutes per batch.
Staff ownership. Assign one person the reconciliation role per shift; don't let it drift. Missed reconciliations turn into compliance audit findings at quarterly-review time.
Escalation path. When a reconciliation fails, the trigger must go to the compliance manager within 1 hour, not waiting for end-of-day. Delayed escalations turn into recall scenarios.

Tooling

Lab operator portals (SC Labs, Kaycha, Confident Cannabis, etc.) — the first line of defense for COA access (as of 2026-04).
Metrc seed-to-sale — ties batch tag to COA data in most states (as of 2026-04 integration varies by state).
ERP/inventory system — stores the operator-side metadata (cost, vendor, retail date).
Batch record — the single source of truth that ties all of the above together with the physical batch.

Audit-Readiness Checklist

Quarterly or at any time the state regulator notifies of an impending audit, run this checklist:

[ ] Every batch on shelf has a COA filed in its batch record.
[ ] Every COA references a lab on the state-approved list (as of the batch sample date).
[ ] Every COA batch-ID matches the Metrc tag on the physical package.
[ ] Every COA's panel-results table is complete for the state's required panels.
[ ] No COA on the sales floor is older than the state's expiration window (e.g., CA flower ~90 days).
[ ] Every remediated batch has the state-required label phrase on retail packaging.
[ ] Every recall event has a complete documentation package (trigger, actions, outcomes).
[ ] Sample retention records are current (retained samples still in archive; retention dates documented).
[ ] Chain-of-custody forms are complete for every sample event in the last 12 months.
[ ] Sampler credential records are current for every sampler who pulled samples.
[ ] Lab contract MSAs are current (expiring within 90 days flagged).
[ ] Dispute and retest records are complete for every challenged result.

(As of 2026-04; customize for your state's specific audit-readiness requirements.)

COA Reconciliation KPIs

Track these KPIs monthly to catch reconciliation drift before it becomes an audit finding:

| KPI | Target (as of 2026-04) | Red Flag | |-----|------------------------|----------| | % of batches with COA filed within 24hr of receipt | >98% | <95% | | % of batches with COA batch-ID match verified | 100% | Any mismatch | | Days between COA receipt and retail release | <2 business days | >5 business days | | % of remediated batches with correct state label | 100% | Any missing | | Audit-trail gaps (batches without complete documentation) | 0 | Any | | Dispute/retest response time (lab → operator) | <48 hours | >7 days | | Sampler-visit no-show rate | <5% | >10% |

(As of 2026-04; benchmarks synthesized from operator interviews and industry surveys.)

Training the Team

Operators who run tight reconciliation have trained their team on:

What a COA actually says. See coa-testing.md §Annotated COA Examples for training material. A receiving clerk who doesn't know what "total yeast + mold" means can't verify the COA.
Metrc manipulation basics. Every inventory manager should be able to navigate Metrc at the batch-tag level, update status fields, and pull a batch-history export for audit.
Escalation thresholds. What triggers a compliance manager call vs what's routine? Written escalation rules prevent underground "let me just release this borderline batch" decisions.
Recall-simulation drills. Run a 30-minute recall drill quarterly (not tied to any real event). Measure time-to-quarantine, time-to-regulator-notification, time-to-consumer-notification. Identify gaps before a real recall event exposes them.

(As of 2026-04; training cadence recommended at new-hire onboarding + annual refresher + post-any-failure-event refresher.)

Related References

coa-testing.md — COA interpretation, testing panels, potency math, state testing matrices, annotated COA examples (canonical — Phase 2).
lab-selection.md — lab procurement: evaluation rubric, RFP template, pricing benchmarks, named labs by geography.
compensation.md — compensation for lab techs (non-retail roles).
legality.md — state regulatory context.
labeling.md — product labeling rules including remediated-product labeling.
delivery-operations.md — operator-manual voice precedent applied here.
opening-dispensary.md — lab selection as part of the launch timeline.

Cross-Phase References

hiring-retention.md — QA / compliance roles responsible for COA reconciliation workflows.
labor-relations.md — union coverage of lab staff and QA staff varies by state.
org-structures.md — where QA / lab-coordination roles sit in the org chart.

Summary of This File's Operator Promise

This file gives you (as an operator) three tested answers: (1) a realistic timeline so you don't over-commit retail dates, (2) a failure decision tree so you don't over-destroy or under-remediate, (3) a remediation-methods toolkit so you pick the right tool for the right contaminant. All figures are as of 2026-04; verify with current vendor quotes and state regulator guidance before acting.

Content date: 2026-04. Verify all state-specific rules with current state regulator guidance before acting on a failure or remediation decision. Quarterly review recommended for pricing, SLAs, and state remediation rules; annual review for vendor footprint.