Choosing a laboratory is one of the few vendor decisions in cannabis production that can invalidate everything upstream of it. Your cultivation can be excellent and your processing tight, and a weak testing partner will still leave you with results you cannot defend, delays you cannot plan around, or a compliance problem you did not know you had.
Knowing how to choose a cannabis testing lab comes down to a short list of things that are genuinely diagnostic — and a longer list of things that sound important and are not.
1. Accreditation, Specifically ISO/IEC 17025
A state licence permits a laboratory to operate. It is not an assessment of technical competence. ISO/IEC 17025 accreditation is — an independent accreditation body evaluates the laboratory’s methods, equipment, personnel competence, quality systems and its ability to produce valid results, then re-assesses periodically.
Ask for the certificate and read the scope. Accreditation is granted for specific tests on specific matrices, not as a blanket endorsement. A laboratory can be accredited for cannabinoid analysis in flower and not for pesticide analysis in edibles. If the analysis you need is outside the accredited scope, the accreditation does not cover your work.
2. Method Validation for Your Product Type
Analytical methods do not transfer cleanly between matrices. Extracting cannabinoids from dried flower is a different problem from extracting them from a chocolate, a gummy or a beverage, and each of those brings interferences that can suppress or inflate a result.
Ask directly: has this method been validated on my product type? A laboratory that mainly runs flower may be genuinely excellent at flower and unprepared for a high-fat infused product. This is the single most common reason an operator gets results that do not match reality — covered further in why edibles are tested differently.
3. Instrumentation Suited to the Panel
You do not need to audit the instrument list, but the equipment should match the work. Pesticide screening at regulatory detection limits requires tandem mass spectrometry. Heavy metals at parts-per-billion require ICP-MS. Potency should be run by HPLC rather than a method involving heat, which alters the acidic cannabinoids mid-analysis.
A laboratory that is vague about how a given analysis is performed is telling you something.
4. Proficiency Testing Performance
Proficiency testing programmes send laboratories blind samples with known values and compare the reported results. It is the closest thing the sector has to an objective scorecard.
Ask whether the laboratory participates and how it has performed. A laboratory that participates and will discuss its results — including the ones that prompted corrective action — is demonstrating exactly the posture you want. One that has never heard of the question is a different proposition.
5. Turnaround Time You Can Actually Plan Around
Consistency matters more than headline speed. A laboratory that reliably returns a full panel in six days is more useful than one that quotes three days and delivers somewhere between three and eleven.
Ask about average turnaround by panel, what happens at peak harvest, and how retests are handled. Also be sceptical of turnaround well below the norm for the full panel. Culture-based microbial methods take days because organisms grow at their own pace. There is a floor.
6. Reporting Quality
Ask to see a sample certificate of analysis before committing. A good report states the methods used, the action limits applied, limits of detection and quantitation, the batch identifier, sample and analysis dates, and an authorised signature.
“Not detected” without a stated detection limit is close to meaningless — it could reflect a highly sensitive method or an insensitive one. A laboratory that publishes its limits is telling you how strong its statements are.
7. Willingness to Explain a Result
This one is underrated. When a batch fails, the question that matters is why. A laboratory that will walk you through the data, explain what the finding suggests and help you narrow down the cause is worth more than one that emails a PDF and moves on.
That relationship is what turns testing from a cost centre into production feedback. It is the difference between knowing a batch failed microbials and knowing that your water activity is running high in one drying room.
8. Independence
The laboratory should have no financial interest in the outcome. Common ownership between a laboratory and the operations it tests, revenue concentrated in a handful of clients, or informal arrangements around retesting all create pressure in the wrong direction.
Independence is the entire premise of third-party testing. Where it erodes, results stop being evidence — the mechanics of that are covered in lab ethics and data integrity and in how third-party testing protects patients.
9. Price, in Context
Price belongs on the list, last. Testing is a real cost and it is reasonable to manage it — but the cheapest quote is frequently the most expensive outcome once you account for a batch that has to be retested, a shipment delayed by inconsistent turnaround, or a result that does not hold up.
Compare like with like. A quote covering fewer analytes, a narrower pesticide list or higher detection limits is not the same service at a lower price. We break the economics down in what cannabis testing costs and what drives the price.
Warning Signs
A few patterns are worth treating as disqualifying.
- Any suggestion of a favourable result. Marketing that hints at higher potency numbers or a friendlier pass rate is describing something other than measurement.
- Reluctance to share accreditation scope or a sample COA. Both are routine documents.
- Vagueness about methods. A competent laboratory can state what instrument runs which analysis without hesitation.
- Turnaround promises that ignore biology. Especially on microbial panels.
- No documented process for handling a disputed result. Disputes happen; a laboratory should have a defined path for them.
Questions Worth Asking Before You Sign
- What is your ISO/IEC 17025 scope, and does it cover every analysis I need on my matrices?
- Have your methods been validated on my specific product types?
- How do you participate in proficiency testing, and how have you performed?
- What is your average turnaround per panel, and what happens during peak season?
- Who performs data review, and is it independent of the analyst?
- What are your detection and quantitation limits for the analytes that matter to me?
- What is your process if I dispute a result?
- How long do you retain samples and raw data?
Any laboratory worth working with will answer all eight without friction. The answers, more than the price sheet, are what tell you what you are buying.
If you are weighing options for an Oklahoma operation, our testing services page sets out the panels we run, and you can request a quote with your product types and volumes for a direct comparison.