Cannabis Lab Selection

Content date: 2026-04. Cannabis lab procurement: evaluation rubric, RFP template, onboarding checklist, per-panel pricing benchmarks, turnaround SLAs, multi-lab strategy, red flags, and named labs by geography. This is procurement guidance, not legal or medical compliance advice. Verify lab accreditations directly with the lab and with the state regulator before contracting.

Canonical COA interpretation / testing science: see coa-testing.md. Operator workflow / remediation / state rules: see lab-testing-operations.md.

Summary

Lab selection is a 6-12 month procurement decision. Switching labs mid-contract is painful (chain-of-custody resets, COA history gaps, retail-shelf COA-matching rework). Pick carefully.
Multi-lab strategy is the default for serious operators — primary lab + secondary/dispute lab. Reduces lab-shopping accusations and provides a verification path.
In scope: evaluation rubric with weights, copy-paste RFP template, onboarding checklist, per-panel pricing (as of 2026-04), turnaround SLAs (as of 2026-04), multi-lab strategy, red flags to avoid, named labs with geographic footprint (as of 2026-04).
Out of scope: COA interpretation (see coa-testing.md); sample pull + failure handling (see lab-testing-operations.md); lab-technician compensation (see compensation.md §Lab Technician).
Procurement-advice disclaimer: Everything below is procurement guidance, not legal or medical compliance advice. Verify lab accreditations directly with the lab and your state regulator before contracting.
Pricing moves fast. Benchmarks below are date-tagged 2026-04. Confirm current pricing with 2-3 labs before any long-term contract.
Three questions this file answers: (1) Which lab is the right primary? (2) How do I run a defensible RFP? (3) What's a fair price and turnaround in my market right now?

Lab Evaluation Rubric

Score each candidate lab 1-5 on each criterion; apply the weight; compute total. A lab scoring below 3.5/5.0 should not be your primary. Weights below sum to 100% across 9 criteria; adjust weights for your specific market if needed (e.g., if you're a concentrate-heavy operator, bump Portal / API to 10% and drop Customer Service to 7%).

| Criterion | Weight | Scoring 1 (poor) → 5 (excellent) | Red Flag | |-----------|--------|----------------------------------|----------| | Accreditation tier | 20% | 1=not accredited · 5=ISO 17025 + state accreditation + proficiency-testing history | Missing state accreditation disqualifies | | Turnaround SLA | 15% | 1=>10 days · 5=3-5 days standard with 24-48hr rush | >7 day standard is uncompetitive | | Per-panel pricing | 15% | 1=>2 SD above market · 5=at or below market | Prices >50% above market need justification | | Dispute / retest policy | 10% | 1=no path · 5=clear written dispute process with retest credit | "We don't dispute our results" = red flag | | Data quality / proficiency history | 10% | 1=no published proficiency data · 5=recent NIST proficiency-test results published | Failed or undisclosed proficiency tests | | Customer service | 10% | 1=>48hr response · 5=<4hr response during business hours | Unresponsive account management | | Chain-of-custody rigor | 8% | 1=weak sampler protocols · 5=documented sampler training + witnessed pulls | No chain-of-custody manual = high risk | | Portal / API quality | 7% | 1=PDF-only · 5=real-time API + dashboard + CSV export | Relying on emailed PDFs scales poorly | | References from other operators | 5% | 1=no references · 5=3+ operator references with positive feedback | "No references available" is concerning |

Total: 100% across all 9 criteria.

How to Use the Rubric

Step-by-step workflow for running the rubric across candidate labs (plan ~4-6 weeks end to end):

Gather 3-5 candidate labs. Pull from §Named Labs by Geography (below) based on your state footprint. Ensure at least one is a regional specialist and at least one is a multi-state platform.
Issue the RFP (see §RFP Template). Require 20-page max response + appendices for certificates + proficiency-testing records.
Score each RFP response independently by at least two reviewers (compliance manager + VP Ops). Average the scores.
Weight and compute total. Weighted score = sum of (criterion score × weight). Highest possible total = 5.0.
Rank candidates. Top two candidates move to reference checks + finalist interviews.
Reference check top two. Call 3+ operator references per finalist. Ask: "What went wrong, and how did the lab respond?" — that question separates real references from cherry-picked ones.
Negotiate final contract with #1. Price, SLA commitments, dispute process, termination clauses.
Pilot for 90 days. Run the rubric again at 90 days with your actual experience data. Formal review decision at month 11 for renewal.

(As of 2026-04; typical procurement cycle runs 60-90 days from RFP issuance to contract signature.)

Worked Example

Illustrative scoring for a California-based multi-location operator evaluating 3 candidate labs:

| Criterion | Weight | SC Labs Score | Kaycha Score | Regional Lab Score | |-----------|--------|---------------|--------------|---------------------| | Accreditation | 20% | 5 | 5 | 3 | | Turnaround | 15% | 4 | 5 | 3 | | Pricing | 15% | 3 | 4 | 5 | | Dispute policy | 10% | 4 | 4 | 2 | | Proficiency | 10% | 5 | 4 | 3 | | Customer service | 10% | 4 | 3 | 5 | | Chain-of-custody | 8% | 5 | 4 | 3 | | Portal / API | 7% | 5 | 4 | 2 | | References | 5% | 5 | 4 | 3 | | Weighted total | 100% | 4.40 | 4.25 | 3.40 |

(Illustrative; actual scores must come from your direct evaluation. As of 2026-04.)

The operator would move SC Labs to primary (4.40) and Kaycha to secondary (4.25); the Regional Lab scoring 3.40 sits below the 3.5 threshold and would not be contracted.

Criterion Deep-Dives

Accreditation tier (20%). ISO 17025 is table stakes — no cannabis lab should be in your RFP without it. The differentiator is state accreditation (each state has a specific list of approved labs as of 2026-04) + proficiency-testing history. A lab with 5+ years of public NIST proficiency results shows operational maturity. A lab that's ISO-accredited but has no published PT record is a "check the box" lab; be cautious.

Turnaround SLA (15%). Three-to-five business days is the market standard for full compliance as of 2026-04. Rush 24-48 hours at +50-100% premium is common. Labs claiming <3 day standard turnaround either have unused capacity (risky — may become a bottleneck at peak) or are cutting corners. Verify with their current operator references.

Per-panel pricing (15%). Compare against the §Per-Panel Price Benchmarks table. Prices below the low end of the range trigger a red flag — either they're undercutting to win business (unsustainable) or they're running a lower-rigor panel. Prices above the high end should come with a clear justification (premium service, faster rush, better data).

Dispute / retest policy (10%). A clear written dispute process is rare but critical. Retest credit ("free retest if your primary COA is challenged") is a differentiator — some labs offer it, most don't. The absence of any dispute process is a red flag.

Data quality / proficiency history (10%). NIST proficiency-testing programs publish results. Labs that participate and perform well will volunteer the data. Labs that avoid PT programs or refuse to share results are operating in a data-opacity zone.

Customer service (10%). Response time during business hours is the most concrete signal. A lab that responds to a routine inquiry in <4 hours during business hours is investing in account management; a lab that takes 48+ hours is short-staffed.

Chain-of-custody rigor (8%). Look at sampler training, sampler turnover, chain-of-custody form structure, and sample-isolation protocols. A lab that can hand you a blank chain-of-custody form and a sampler-training manual on demand is operationally mature. A lab that fumbles the request is not.

Portal / API quality (7%). PDF-only delivery is a drag on scale. Real-time API access (SC Labs, Confident Cannabis, Kaycha have good APIs as of 2026-04) enables automated COA-to-product reconciliation. CSV export for historical data is a minimum; REST API for real-time is the target.

References (5%). Three+ operator references is a signal of maturity. Ask the references about specific events ("Tell me about a failed panel; how did the lab respond?") rather than general "how are they doing?" questions. Weak references tell you what's wrong; strong references tell you what's right.

When to Re-Score Your Primary Lab

Quarterly during the first year of the relationship.
After any significant event (missed SLA >2x, disputed result, customer-service escalation).
Annually in steady state, with formal review decision for renewal at month 11.
After any regulator enforcement action against the lab (public or internal knowledge).

(As of 2026-04; annual cadence is a minimum.)

RFP Template

Copy-paste, customize organization name + state, and send to 3-5 labs. Response format standardization is the point — comparing RFP responses is how you score the rubric.

REQUEST FOR PROPOSAL
Cannabis Testing Laboratory Services
[OPERATOR NAME] — [STATE] — [DATE]

1. INTRODUCTION
   1.1 Organization: [OPERATOR NAME]
   1.2 License type(s): [retail / cultivator / manufacturer / MSO]
   1.3 State(s): [list]
   1.4 Expected monthly test volume: [X full compliance panels + Y potency-only + Z rush]
   1.5 Contract term requested: [12-24 months with annual review]
   1.6 Procurement contact: [name, email, phone]
   1.7 About us: [2-3 sentence company description]

2. SCOPE OF SERVICES
   2.1 Required panels: [full compliance / potency only / pesticide only / heavy metals / microbial / residual solvents / mycotoxin / water activity / terpene profile]
   2.2 Sample pull services: [lab-sampler pull / operator drop-off / hybrid]
   2.3 Geographic coverage: [list states/cities where samples originate]
   2.4 Rush availability required: [Yes/No; 24-48hr target]
   2.5 Retest handling: [detail expected retest volume and process]
   2.6 Supplemental services (if desired): [terpene profile, water activity, extended heavy-metals]

3. ACCREDITATION REQUIREMENTS
   3.1 ISO 17025 accreditation (required; provide current certificate)
   3.2 State accreditation(s) (required for all states listed in 2.3; provide state-specific certificates)
   3.3 Proficiency-testing history (provide last 2 years of PT results)
   3.4 Regulatory warning-letter or enforcement-action history (provide if applicable)
   3.5 Sampler credentials (sampler training program detail; sampler roster)
   3.6 Liability insurance (provide certificate with coverage amounts)

4. PRICING
   4.1 Per-panel pricing (provide current rate card for all panels in 2.1)
   4.2 Volume discount tiers (at 50+, 100+, 200+ samples/month)
   4.3 Rush surcharge structure (24hr, 48hr, 72hr premium rates)
   4.4 Re-test credit policy (free retest of same batch? cost-share? dispute credit?)
   4.5 Invoicing terms (net 30 / net 45 / net 60 / other)
   4.6 Annual price-escalation cap (protect against mid-contract rate increases)

5. SLAs
   5.1 Standard turnaround (provide with 95th-percentile commitment, not average)
   5.2 Rush turnaround (24-48hr commitment; identify rush availability by panel type)
   5.3 Sampler-visit scheduling lead time
   5.4 COA delivery mechanism (portal / email / API) and timing
   5.5 Missed-SLA remedy (credits, dispute path, contract-exit trigger)

6. DISPUTE AND RETEST POLICY
   6.1 Dispute process (written, with timeline and acknowledged-within-window)
   6.2 Retest credit structure (free retest of same batch? cost-share?)
   6.3 Escalation path if dispute unresolved (named contact)
   6.4 Sample retention (duration of lab-retained material for dispute purposes)

7. DATA AND INTEGRATION
   7.1 Portal access (URL, user-management, staff-level permissions)
   7.2 API availability (documented REST or SOAP; version history)
   7.3 CSV / bulk-export options
   7.4 Seed-to-sale integration (Metrc / BioTrack / etc.)
   7.5 Data-ownership clause (who owns the COA data — operator or lab)
   7.6 Data-retention duration (how long lab retains operator's historical COAs)

8. CHAIN OF CUSTODY
   8.1 Sampler training requirements (curriculum, refresher cadence)
   8.2 Chain-of-custody form structure (provide blank form)
   8.3 Batch-isolation protocols while in lab custody
   8.4 Retention time for unused sample material
   8.5 Transport conditions (temperature, seal integrity, time limits)

9. EVALUATION CRITERIA
   See Lab Evaluation Rubric provided with this RFP (or inline in Section 13).

10. RESPONSE FORMAT
    10.1 Length limit: 20 pages plus appendices for certificates
    10.2 Format: PDF via email to [procurement contact]
    10.3 Required attachments: ISO 17025 cert, state accreditations, last 2 PT results, rate card, liability-insurance cert, 3+ operator references

11. TIMELINE
    11.1 RFP issued: [DATE]
    11.2 Questions accepted through: [DATE]
    11.3 Responses due: [DATE]
    11.4 Evaluation period: [DATE range]
    11.5 Finalist interviews: [DATE]
    11.6 Award: [DATE]
    11.7 Contract start: [DATE]

12. CONTACT
    12.1 Procurement lead: [NAME, EMAIL, PHONE]
    12.2 Questions to: [EMAIL]
    12.3 Questions-NOT-to: no outreach to individual employees beyond the procurement lead

13. EVALUATION RUBRIC (optional inline)
    [Paste the full 9-criterion rubric if not provided as an attachment]

14. TERMS AND CONDITIONS
    14.1 Confidentiality: responses are confidential and not shared with competitors
    14.2 No commitment: this RFP does not commit the operator to award a contract
    14.3 Costs: respondent bears its own cost of responding

How to Customize the RFP

For single-location operators. Drop §2.3 if you're only in one state; keep §3.2 focused on the one accreditation.
For MSO operators. Expand §2.3 with every state footprint; require state-specific accreditations in §3.2.
For cultivators only. Focus §2.1 on flower panels; skip concentrate-specific panels.
For manufacturers only. Emphasize residual solvents, terpene profile, and concentrate-specific panels; skip full-compliance bundles.
For retailers only. Most retailers don't contract directly with labs (product arrives with COA); this RFP is for integrated operators. If you're pure retail, your "lab question" is about verifying upstream COAs — see lab-testing-operations.md §COA-to-Product Reconciliation Workflow.

(As of 2026-04; adjust RFP structure to your specific procurement needs.)

Common RFP Mistakes

Too narrow on scope (§2). A vague scope produces wildly varying RFP responses. Be specific about panel mix and monthly volume.
No SLA in §5.5. Without a missed-SLA remedy, you have no leverage when the lab misses its commitment. Insist on credits or contract-exit triggers.
No data-ownership clause in §7.5. If you switch labs, who owns the historical COA data? Resolve this up front.
No annual price-escalation cap in §4.6. Labs that can raise prices freely mid-contract destroy your cost predictability.
Waiving the reference requirement. Lab references are the single strongest signal. Don't let a candidate lab skip this — if they resist, that's an answer.

Onboarding Checklist

A 12-week onboarding runway gets the lab relationship from contract signature to steady-state operations. Compress only if your state entry date is hard; otherwise follow this cadence (as of 2026-04).

Weeks 1-2: Contract and MSA

[ ] Executed MSA with dispute process, indemnification, and data-ownership terms
[ ] Executed NDA if proprietary batches/cultivars discussed
[ ] Account setup: billing contact, technical contact, escalation contact
[ ] Rate card confirmed in writing
[ ] SLA commitments confirmed in writing (standard + rush)
[ ] Dispute-and-retest policy confirmed in writing with named escalation path
[ ] Data-ownership and retention terms confirmed in writing

Weeks 3-4: Chain-of-Custody + Sampling Protocol

[ ] Lab sampler introduction and credentials verified (as of 2026-04)
[ ] Chain-of-custody form template reviewed
[ ] Sampler-visit scheduling protocol documented (standard + rush lead times)
[ ] Sample-isolation protocol at operator facility confirmed
[ ] Batch retention policy confirmed (duration lab keeps unused sample material)
[ ] Transport conditions confirmed (seal integrity, temperature, time limits)
[ ] Sampler training records filed
[ ] Backup-sampler identified for primary-sampler absences

Weeks 5-8: First Production Batches

[ ] First full-panel batch submitted (pilot run)
[ ] COA received + reviewed for format and data completeness
[ ] Portal access verified for ops/compliance staff (including access-control and role-based permissions)
[ ] API integration (if applicable) tested against Metrc / BioTrack / POS system (as of 2026-04)
[ ] Mock dispute flow exercised (to confirm process works)
[ ] Seed-to-sale system integration verified
[ ] Retest/dispute path exercised on a borderline result (controlled test)
[ ] Batch record workflow documented and tested

Weeks 9-12: SLA Audit and Rate Review

[ ] 95th-percentile turnaround audit against committed SLA
[ ] Rush turnaround audit
[ ] Invoice accuracy audit
[ ] Account management responsiveness audit (time-to-first-response)
[ ] Rate card review (any unexpected charges?)
[ ] Contract renewal / extension decision trigger set for month 11
[ ] Multi-lab strategy baseline established (volume allocation primary vs secondary)

Month 4-11: Steady State Operations

[ ] Monthly KPI dashboard review (SLA performance, invoice accuracy, dispute count)
[ ] Quarterly rubric re-score
[ ] Semi-annual business review with lab account management
[ ] Month 11 formal renewal decision with go/no-go

(As of 2026-04; adjust cadence based on operator scale and lab complexity.)

Onboarding RACI

A RACI (Responsible / Accountable / Consulted / Informed) mapping for onboarding activities:

| Activity | Responsible | Accountable | Consulted | Informed | |----------|-------------|-------------|-----------|----------| | Contract negotiation | Procurement lead | VP Ops or COO | Legal, Finance | GM, Compliance | | MSA execution | Legal | VP Ops | Procurement | GM, Finance | | Account setup | Compliance manager | Procurement | Lab account manager | Ops team | | Sampler-schedule setup | Compliance manager | Compliance manager | Lab sampler lead | Cultivation, Ops | | Portal/API integration | IT / Engineering | IT Director | Compliance, Ops | Retail, GM | | First-batch pilot | Inventory manager | Compliance manager | Lab account manager | GM, VP Ops | | SLA audit at week 12 | Compliance manager | VP Ops | Procurement | GM, Finance | | Rate-card review | Procurement lead | VP Ops | Compliance, Finance | GM | | Month-11 renewal decision | VP Ops or COO | CEO / COO | Compliance, Finance, Legal | GM |

(As of 2026-04; adjust to your org structure. Small operators collapse roles — "Compliance Manager" and "Procurement Lead" may be the same person.)

Onboarding Cost Model

Expect $15,000-$50,000 in total onboarding cost depending on operator scale (as of 2026-04):

Legal review of MSA: $2,000-$10,000 (depends on counsel rates and contract complexity).
Compliance labor during onboarding weeks 1-12: $5,000-$25,000 (100-250 compliance-manager hours).
IT / integration labor: $3,000-$10,000 (API setup, portal user-management, seed-to-sale integration).
Pilot batch testing: $2,000-$5,000 (3-5 full-panel batches as pilot).
Mock dispute drill: $500-$1,500 (lab participation in test dispute).

Amortize across contract term; a 24-month contract at $25,000 onboarding cost = ~$1,000/month onboarding amortization on top of ongoing per-sample costs.

Per-Panel Price Benchmarks

(As of 2026-04; confirm against 2-3 active labs before long-term contract. Prices are per sample, not per lb. Pricing varies by market — CA and CO tend to be ~10% below the midpoint of the range; NY, NJ, and newer markets tend to be at or above the midpoint.)

| Panel | Price Range (per sample) | Notes | |-------|--------------------------|-------| | Potency only (cannabinoid profile) | $80-$150 | Fast turnaround; most commonly used for raw flower pre-manufacture | | Full compliance panel | $350-$550 | Includes potency + pesticides + heavy metals + microbial + residual solvents + mycotoxins + water activity | | Pesticide-only | $150-$225 | Targeted follow-up on failed panel | | Heavy metals | $120-$175 | Lead / cadmium / arsenic / mercury standard; some labs add extended | | Microbial | $100-$150 | Typically includes total yeast + mold + E. coli + Salmonella; some states require more | | Mycotoxin | $75-$125 | Usually bundled with full compliance | | Residual solvents (concentrates only) | $100-$175 | Butane, propane, ethanol, etc. | | Terpene profile (optional) | $75-$150 | Not typically compliance-required | | Water activity | $50-$100 | Cheap; bundled with full compliance in most states |

Rush premium: +50% to +100% for 24-48hr turnaround (varies by lab and workload; as of 2026-04). Availability: CA, CO, MA, FL, NV, OR, NY typically yes; smaller markets variable.

Volume discounts: Labs typically offer 10-20% volume discount at 50+ samples/month; 15-30% at 200+ samples/month (as of 2026-04). Negotiate in the RFP (§4.2).

Sampling-event fees: Separate from per-panel pricing. Typical $0-200 per sampling visit included in base rate (large operators); $200-500 rush sampling fee for same-day or next-day visits (as of 2026-04).

Price drift. 2024-2025 pricing was flat to slightly down as lab capacity caught up with demand (~5% decline per industry surveys). 2026-04 pricing is stable; next refresh recommended 2026-07.

Pricing Red Flags

>30% below market without a clear volume-discount or promotional explanation. Either a loss-leader or corner-cutting.
>50% above market without a clear premium-service justification (faster turnaround, broader panels, better data).
Unbundled "surprise" fees at invoice time (sampling-visit fee not in the rate card, retest surcharge not disclosed, per-sample setup fee).
Escalation without cap. Labs that raise prices mid-contract without a pre-agreed cap.
Locked-in multi-year contracts with no exit clause. A 3-year contract with no termination right is a trap when the lab underperforms.

Wholesale vs Retail Pricing

Most labs don't differentiate between wholesale and retail operators in pricing — it's volume-based. The exception is vertically-integrated MSOs who negotiate portfolio-wide pricing; they typically pay 15-25% below individual-operator rates (as of 2026-04).

Turnaround SLA Benchmarks

(As of 2026-04; confirm with current lab contracts. SLA "commitment" means the 95th-percentile commitment — ask labs for this, not the average.)

| Panel Type | Standard (business days) | Rush (hours) | Rush Premium | |-----------|--------------------------|--------------|--------------| | Potency only | 2-5 days | 24 hours | +50-75% | | Microbial | 4-7 days | 48 hours | +75-100% | | Full compliance | 5-10 days | 48-72 hours | +75-100% | | Pesticide/heavy-metal | 5-7 days | 48 hours | +50-100% | | Residual solvents | 3-5 days | 24 hours | +50% | | Mycotoxin (bundled) | — | — | — | | Water activity (bundled) | — | — | — |

(As of 2026-04; Kaycha Labs claims 48-hour delivery as standard with sample-automation — verify current operational commitments.)

SLA Performance Measurement

How to track lab SLA performance in your own data (monthly):

Median turnaround (days): central tendency per panel type.
95th-percentile turnaround: the commitment number — this is what the contract guaranteed.
Missed-SLA rate: % of samples that exceeded committed turnaround.
Trailing-90-day trend: is performance improving or degrading?

A lab in steady state should have <5% missed-SLA rate on its committed SLA (as of 2026-04). Above 10% is grounds for a performance conversation; above 20% is grounds for contract re-evaluation.

Rush Availability by Market

Rush availability varies by market density (as of 2026-04):

| Market | Rush 24-48hr Availability | |--------|----------------------------| | CA | High — multiple labs offer | | CO | High | | MA | Medium-high — Kaycha, regional labs | | FL | Medium — Kaycha, ACS | | NV | Medium | | OR | Medium — Kaycha, SC Labs | | NY | Medium | | AZ | Medium-low | | IL | Medium-low | | NJ | Variable — emerging market | | MI | Medium-low | | Smaller markets | Variable — confirm with specific lab |

(As of 2026-04; rush availability correlates with local lab density and market maturity.)

Multi-Lab Strategy

Core principle: Serious operators run 2+ labs. A single-lab relationship over many years creates regulatory exposure ("lab-shopping" enforcement in CA, CO, WA has targeted operators with single-lab relationships when results shifted post-switch) and concentration risk (SLA miss during peak season, accreditation suspension).

Why Multi-Lab

Dispute verification path. If your primary lab fails a batch, a secondary lab can confirm or disconfirm before you commit to destroy/remediate.
SLA insurance. If primary lab has a sample-volume surge (September-November harvest season, Q4 promotional cycles), secondary handles overflow.
Regulatory / enforcement protection. Using a single lab for years can trigger "lab-shopping" enforcement scrutiny in CA, WA, CO (see §Red Flags + §Historical Enforcement Context). Splitting volume across 2 labs with documented criteria is defensible.
Benchmarking / calibration. Periodic split-testing reveals systematic lab variance. Without a second lab, you have no baseline for "is this lab's potency reading off?" (as of 2026-04).
Business continuity. Accreditation suspension, state enforcement action, lab closure, acquisition disruption — any of these can idle your primary lab overnight. Multi-lab relationships absorb the shock.

Primary / Secondary Structure

Primary: ~80% of routine testing. Full contract, preferred pricing, API integration. Tight operational integration (portal, sampler schedule, dispute process).
Secondary: ~20%. Dispute verification + overflow + split-test audits. Simpler contract, rate-card pricing, portal access. Standing agreement so activation is fast during dispute / overflow / crisis.

For high-volume operators (500+ samples/month), consider a three-lab structure: Primary (60%), Secondary (25%), Tertiary (15%) for further resilience.

Split-Testing Protocol

When to split:

Quarterly audit (4-6 batches per quarter) to calibrate lab variance.
Any batch >$20K wholesale value — bets big enough to warrant second opinion.
Any batch that generated a dispute at primary; second lab becomes dispute-resolution.
After any regulator notice, audit, or enforcement action against primary lab.

How to structure the split:

Same sample pull event (lab samplers from both labs on-site same day).
Identical sample weights and containers (chain-of-custody issued separately to each lab).
Same panel mix tested at both.
Results reviewed side-by-side within 7 business days (as of 2026-04; may be longer during peak).

How to interpret variance:

Potency variance <5% between labs: normal measurement variance.
Potency variance 5-10%: worth investigating — sample heterogeneity, lab SOP differences.
Potency variance >10%: systemic issue — one lab may be systematically off; reconsider primary lab choice.

What triggers a lab-performance review:

Repeated variance >10% across multiple split tests.
Systematic direction (e.g., primary consistently reads higher THC than secondary across 4+ batches).
Dispute patterns (multiple disputes resolved in favor of secondary).

(As of 2026-04; calibration cadence varies by operator scale and risk tolerance.)

Cost-Benefit of Multi-Lab

Additional cost: Split-testing 4-6 batches per quarter at $350-550 each = $1,400-$3,300 / quarter (as of 2026-04).
Benefit: Dispute defense ($10K-$50K on a single contested batch), regulatory protection (avoiding a lab-shopping enforcement action worth $50K-$500K in fines + license exposure), business continuity (avoiding a 2-week primary-lab outage).
Break-even: Multi-lab strategy pays for itself with ONE prevented dispute or enforcement action over a 2-3 year contract.

Allocation Planning

How to structure the 80/20 or 60/25/15 allocation in practice:

By SKU category. Route flower to Primary, concentrates to Secondary (if Secondary has stronger residual-solvents program) — clear separation by category.
By batch value. High-value batches (>$20K wholesale) split-tested; standard batches go only to Primary.
By random sampling. Every Nth batch routed to Secondary regardless of category — produces cleanest calibration data.
By calendar. Week 1-3 of the month → Primary; Week 4 → Secondary — produces rhythm with predictable cost.

The right allocation depends on your operator type. Concentrate-heavy operators often benefit from residual-solvents specialization at the Secondary lab; flower-heavy operators often allocate randomly to keep calibration clean.

(As of 2026-04; refine allocation based on 90-day performance data.)

When to Switch Primary Labs

A primary-lab switch is a 6-9 month transition (chain-of-custody history resets, batch-record gaps, retail-shelf COA-matching rework). Triggers:

Primary lab score in rubric drops below 3.5/5.0 on 2+ quarterly re-scores.
Regulator enforcement action against Primary (warning letter, accreditation suspension).
Missed-SLA rate exceeds 20% for 3+ consecutive months.
Dispute patterns (>3 disputes in 6 months resolved in favor of Secondary).
Pricing becomes non-competitive (rate card >25% above market benchmarks for 2+ quarters as of 2026-04).
Ownership / acquisition changes produce instability or staff turnover.

Transition plan:

Month 1-2: Dual-operate Primary and prospective-Primary (promoted from Secondary, or new lab from RFP). Batch routing shifts gradually.
Month 3: New Primary takes 50% of volume; old Primary drops to 30%; Secondary stays at 20%.
Month 4-6: New Primary ramps to 80%; old Primary winds down to 0%; new Secondary fills gap.
Month 7+: Steady state with new Primary and new Secondary.

(As of 2026-04; don't rush a primary switch — transition errors cost more than tolerating an underperforming primary for another quarter.)

Integrations and Data Portability

Before you sign a multi-lab relationship, confirm data portability:

Can you export all historical COAs from Primary when (not if) the relationship ends?
Do both labs support your seed-to-sale integration (Metrc, BioTrack, etc.) (as of 2026-04)?
Can you consolidate COA reporting across multiple labs in your batch-record system?
Are there unified KPI dashboards, or do you track lab performance in spreadsheets?

API-first labs (SC Labs, Kaycha, Confident Cannabis as of 2026-04) make multi-lab operationally easier. PDF-only labs create integration friction at scale.

Red Flags

Watch for these during lab demos, RFP response review, or the first 90 days of operation. Any one of these warrants a re-evaluation.

THC Inflation Offers

Some labs implicitly (or explicitly) offer higher potency readings. This is the single biggest red flag in cannabis testing. Specific patterns:

"Our potency comes in 2-4% higher than [competitor lab]."
Unexplained potency jumps when you switch labs.
Pressure to re-test at the new lab until a "good" number comes back.

Consequence: Regulatory enforcement (CA DCC, WA WSLCB, CO MED all have documented lab-shopping and potency-inflation enforcement actions). Brand/reputational damage. Licensing exposure. (As of 2026-04; verify current enforcement posture with state regulator.)

Vague Chain-of-Custody Answers

"Our sampler follows our protocol" — without a written manual.
Inability to produce a blank chain-of-custody form on request.
Sampler credentials not documented.
Sampler turnover >50% year-over-year without a structured training program.

Regulatory Warning-Letter or Enforcement History

State regulator warning letter in the last 24 months.
Published enforcement action.
Accreditation suspension or revocation.
Lab investigated for data falsification (public or whispered).

Unclear Dispute Process

"We stand by our results" with no written dispute procedure.
No retest credit structure.
No escalation path.
No named contact for dispute resolution.

Pricing Significantly Below Market

Prices >30% below market benchmark without an obvious volume-discount explanation (as of 2026-04).
"Free potency panel" (often paired with mandatory upsell to full panel).
Loss-leader pricing (unsustainable).

Portal / Data Opacity

PDF-only deliverable.
No API.
No historical data export.
COA data owned by the lab, not the operator.

Accreditation Gaps

Missing state-specific accreditation for one of your operating states.
Expired ISO 17025 certificate.
No NIST proficiency-testing program participation.

Operational Immaturity

No dedicated account manager.
Customer-service response times >48 hours on routine inquiries.
Inability to provide 3+ operator references.
Recent management turnover or acquisition without a stability signal.

Red-Flag Combination Patterns

Any ONE red flag above is grounds for a hard conversation. Two or more compounds to a hard no. Common compound patterns (as of 2026-04):

Pricing below market + no published proficiency testing + vague chain-of-custody. Loss-leader + data-opacity + process-weakness = operational risk.
THC inflation offer + no written dispute process + PDF-only deliverable. Three opaqueness signals stacked.
Recent enforcement action + new leadership + introductory pricing. A lab trying to reset its reputation while cutting prices to win business. High risk.
Missing state accreditation + aggressive sales outreach + 3-year no-exit contract. A lab trying to lock operators in before getting accredited.

When you see a compound pattern, stop evaluating and move on. One red flag might be explainable; two is a warning; three is a disqualification.

How to Exit a Bad Lab Relationship

If you've discovered red flags post-contract, execute an exit cleanly:

Document the pattern. Log every missed SLA, every dispute, every customer-service gap with timestamps and evidence.
Trigger the missed-SLA remedy. Invoice credits, contract-exit triggers per §RFP Template §5.5.
Give formal notice per contract. Most contracts have a 30-60 day termination-for-cause window.
Transition to secondary or to new primary per §When to Switch Primary Labs workflow.
Retrieve your data. Export all historical COAs, batch records, dispute files before terminating portal access.
Close out outstanding invoices. Don't let a bad relationship end in a payment dispute.
Add the lab to your internal "avoid" list for future RFPs.

(As of 2026-04; exit typically takes 60-90 days from decision to clean break.)

Early-Warning Behavioral Patterns

Account manager pushes you to sign a multi-year contract with no exit clause.
RFP response is light on data (no proficiency-testing results, no references, no SLA commitments).
Sampler arrives without the chain-of-custody form pre-populated with your batch information.
Invoice line items don't match the rate card agreed in the contract.
COA delivery is consistently late and "we'll credit you next month" is the pattern.

(As of 2026-04; patterns observed across multiple operator interviews.)

Named Labs by Geography

Every lab below is named as of 2026-04. Verify current accreditations, state footprints, and operational status directly with the lab before including in an RFP. This is procurement guidance, not an endorsement.

California Focus

SC Labs (CA, OR, CO, MI) — Founded 2010. ISO 17025. CA DCC-compliant. Well-known for breadth of services; one of the "most comprehensive" cannabis labs in the US. Source: sclabs.com. (Verify currently serving CA, 2026-04.)
Anresco Laboratories (CA) — Oldest continuously operating US lab (since 1943). ISO 17025. CA-only footprint; premium-rigor reputation. (Verify current cannabis-testing availability, 2026-04.)
Green Leaf Lab (OR, CA) — Oregon-founded; cross-state into CA. ISO 17025. (Verify current footprint, 2026-04.)

Multi-State / National

Kaycha Labs (CA, CO, FL, MA, NV, NY, OR, TN) — ISO 17025. 500+ procedures. 48-hour turnaround claim with sample-automation infrastructure. Source: kaychalabs.com. (Verify current state availability, 2026-04.)
Confident Cannabis (multi-state) — ISO 17025 + data/marketplace platform integration. Strong API / portal integration. (Verify current state footprint, 2026-04.)

Southeast / National Hemp

ACS Laboratory (FL + national hemp/cannabis) — ISO 17025 + DEA-certified hemp. Known for hemp-testing dominance. Strong FL footprint. (Verify current footprint, 2026-04.)

West Coast

SC Labs (CA, OR) — as above.
Green Leaf Lab (OR, CA) — as above.
Kaycha Labs (OR) — as above.

Rocky Mountain / Mountain West

Kaycha Labs (CO) — as above.
SC Labs (CO) — as above.
Regional Colorado specialists (verify current listings on sbg.colorado.gov/med approved-lab list, as of 2026-04).

Midwest

SC Labs (MI) — as above.
Kaycha Labs (NY — near-Midwest; not a Midwest lab per se) — as above.
Regional Illinois specialists (verify current listings on idfpr.illinois.gov approved-lab list, as of 2026-04).

Northeast

Kaycha Labs (MA, NY) — as above.
Regional MA specialists (verify current CCC approved-lab list at masscannabiscontrol.com, as of 2026-04).
Regional NY specialists (verify current OCM approved-lab list at cannabis.ny.gov, as of 2026-04).

Historical Lab-Shopping Enforcement Context

CA — DCC warning letters 2022-2024 on potency-inflation; multiple labs cited. (As of 2026-04; verify current DCC enforcement posture at cannabis.ca.gov.)
WA — WSLCB action on potency inflation 2023. (As of 2026-04; verify at lcb.wa.gov.)
CO — MED scrutiny of lab practices in 2023-2024. (As of 2026-04; verify at sbg.colorado.gov/med.)
OR — OLCC has investigated potency-testing variance; rules on lab practices tightened 2024. (As of 2026-04; verify at oregon.gov/olcc.)

Implication for your lab selection: avoid labs with recent enforcement actions; document WHY you picked your primary lab (via the rubric) so "lab-shopping" claims don't stick.

Lab-Selection Heuristics by Operator Type

| Operator Type | Primary Lab Archetype | Secondary Lab Archetype | |---------------|------------------------|--------------------------| | Single-location retail | Regional specialist (best pricing) | Multi-state platform (dispute verification) | | 2-5 location regional | Multi-state platform (consistency) | Regional specialist (pricing pressure) | | MSO enterprise | Multi-state platform (scale) | Second multi-state + regional specialists by market | | Cultivator (vertically integrated) | Multi-state platform or regional specialist | Second lab for split-testing / calibration | | Manufacturer (concentrates) | Lab with strong residual-solvents + terpene panel | Second lab for dispute verification | | Hemp/CBD brand | ACS Laboratory or multi-state platform | Regional hemp specialist |

(As of 2026-04; adjust to your specific operator profile and state footprint.)

Evaluating New Entrants

New labs enter the market regularly. Two signals to treat with caution:

<2 years operating in the state. Operational maturity takes time; early-stage labs miss SLAs, cut sampler corners, or lack dispute processes.
Pricing 30%+ below established labs. Loss-leader pricing to win market share is unsustainable — expect rate increases within 12-18 months.

Two signals that a new entrant is worth considering:

Experienced leadership team. Ex-SC Labs, ex-Kaycha, ex-Confident Cannabis leadership indicates operational competence.
Published proficiency-testing results from day one. Signals a commitment to data transparency.

(As of 2026-04.)

Due-Diligence Questions for Any Named Lab

When evaluating a lab on this list — or any other — ask these questions in the RFP response or in finalist interviews (as of 2026-04):

Accreditation specifics. "Please provide your ISO 17025 certificate, scope-of-accreditation document, and state-specific accreditation certificates for all states where we operate."
Proficiency testing. "Please provide your last 24 months of NIST or ISO proficiency-testing results, by panel type."
Sampler training. "Please provide your sampler training curriculum, sampler roster, and sampler-turnover rate over the last 12 months."
Enforcement history. "Please disclose any regulator warning letters, enforcement actions, or accreditation-suspension events in the last 24 months, in any state."
Dispute process. "Please provide your written dispute-and-retest policy, including escalation path and retest-credit structure."
Operator references. "Please provide 3+ current operator references with contact names and permission to call."
Rate card. "Please provide your current rate card for all panels in scope, with volume-discount tiers and rush premiums."
SLA commitments. "Please provide your 95th-percentile turnaround commitment by panel type, with missed-SLA remedy."
Data ownership. "Please confirm that operator owns COA data and can export historical records in CSV and/or API format."
Account management. "Please identify the primary account manager + backup, and confirm response-time commitments during business hours."

(As of 2026-04; every one of these should be answerable in writing. If a candidate lab can't answer any one, move on.)

Vendor-Selection Anti-Patterns

Operators frequently fall into these traps when selecting a lab:

"We've always used them." Inertia is not a rubric criterion. Re-score annually; evaluate new entrants.
"The sales rep is charming." Charming sales reps produce RFP responses. Operational performance produces results. Don't confuse the two.
"They gave us a good deal on the first batch." Introductory pricing is an investment-acquisition pattern, not a sustainable rate. Demand the rate-card up front.
"They're located close to us." Physical proximity matters for sampling-visit scheduling but not for COA quality. Don't over-weight geography in the rubric.
"Our competitor uses them, so it must be fine." Your competitor's lab choice was made with different priorities. Score against YOUR rubric.
"They promised us faster turnaround than the others." Promise ≠ delivery. Look at 95th-percentile actual performance from references, not brochure promises.

(As of 2026-04; anti-patterns synthesized from operator interviews.)

Vendor-Relationship Management

Post-contract, managing the lab relationship is a quarterly cadence with monthly check-ins. Elements of mature vendor-relationship management (as of 2026-04):

Monthly Check-In (30 min)

SLA performance review (95th-percentile vs commitment).
Invoice accuracy review (line-item match to contract rate card).
Dispute/retest count + resolution status.
Any escalations or incidents in the prior 30 days.
Upcoming sample-volume forecast (helps lab plan capacity).

Quarterly Business Review (60-90 min)

Rubric re-scoring.
Rate card review (any changes? any new services?).
Split-testing results from the quarter (variance analysis).
Regulatory updates affecting either party.
Renewal-decision pipeline update (for month-11 renewal decisions).

Annual Strategic Review (2-4 hours)

Year-over-year performance (SLA, price, quality signals).
Market-pricing benchmarking (are you still getting a competitive rate?).
Accreditation and enforcement updates.
Contract renewal decision with formal go/no-go.
Multi-lab allocation review.

Vendor-Health Dashboard

A simple monthly dashboard per lab to track vendor health:

| KPI | Target (as of 2026-04) | This Month | Trailing 3-Month Avg | |-----|-------------------------|------------|----------------------| | 95th-percentile turnaround (days) | Per SLA commitment | X | Y | | Missed-SLA % | <5% | X | Y | | Dispute count (monthly) | 0-2 | X | Y | | Customer-service response time | <4hr | X | Y | | Invoice accuracy | 100% | X | Y | | Portal/API uptime | >99% | X | Y |

Fill this out monthly. When two consecutive months go red, escalate to the lab's account management; when four consecutive months go red, trigger a rubric re-score.

(As of 2026-04; customize KPIs to your operator profile.)

Relationship-Health Signals Beyond KPIs

Beyond the numeric dashboard, watch for relationship-health signals (as of 2026-04):

Account manager responsiveness degradation. Responses that used to come in 2 hours now come in 12. Signals account-management de-prioritization.
Sampler turnover spike. A new sampler every 2 weeks signals internal lab instability.
Invoice surprises. "New sampling-event fee" or "equipment-use surcharge" that wasn't in the rate card.
Portal outages. Unexplained downtime >2 hours during business hours.
Retest friction. Disputes that used to resolve in 7 days now take 30. Signals process breakdown.
Reduced attention during QBR. Quarterly business review gets rescheduled 2x or the lab sends a junior rep. Signals de-prioritization.

Any of these should trigger an informal conversation with the lab's account management. Two or more should trigger a rubric re-score.

Relationship Exit Indicators

When a relationship is unfixable (as of 2026-04):

Regulator enforcement action against the lab.
Accreditation suspension or loss.
Rate-card increase >20% mid-contract with no cap.
Three consecutive quarters of sub-3.5 rubric score.
Pattern of data-integrity concerns (disputed COAs, unexplained potency patterns).

At this point, execute the exit workflow in §How to Exit a Bad Lab Relationship above.

Related References

coa-testing.md — COA interpretation, testing panels, potency math, state matrices (canonical).
lab-testing-operations.md — operator workflow, sampling, failures, remediation, recall.
compensation.md — lab-technician compensation (non-retail role).
opening-dispensary.md — where lab selection fits in the launch timeline.
legality.md — state regulatory context.
labeling.md — state-specific labeling rules including remediated-product labels.
banking.md — named-institution archetype pattern applied here to named labs.

Glossary

ISO 17025. International standard for testing and calibration laboratory competence. Table stakes for cannabis labs.
Proficiency Testing (PT). Third-party audits of lab accuracy; programs like NIST and AIHA publish results.
Chain of Custody. The documented record of who had the sample, when, and under what conditions — from sampler pull through lab analysis.
SLA. Service-Level Agreement — the lab's turnaround and performance commitments.
RFP. Request for Proposal — structured procurement document sent to candidate labs.
MSA. Master Service Agreement — the contract governing the ongoing lab relationship.
95th-percentile turnaround. The turnaround that 95% of samples meet or beat — the operator-relevant commitment (as of 2026-04).
Full-compliance panel. Bundle of all state-required panels for a given product type (potency + pesticides + heavy metals + microbial + residual solvents + mycotoxins + water activity for flower in most states as of 2026-04).
Rush. Expedited turnaround, typically 24-48 hours, at a +50-100% premium (as of 2026-04).
Retest credit. Free or discounted retest when the operator disputes a result.
Split-testing. Sending the same sample to two labs for calibration or dispute verification.

Content date: 2026-04. Quarterly refresh recommended for pricing and SLA benchmarks; annual refresh for lab footprint and enforcement history. Next scheduled review: 2026-07.