Our approach
Testing and Batch Verification
Last updated · 2026-09-12
A route to evidence is not the evidence itself
A bag QR code, lot lookup, certificate link or testing portal helps us find information. It does not automatically earn the testing verification. We need to inspect the underlying evidence and record what it actually covers.
What we look for
For numerical contaminant review, we look for the exact product and sample identity, lot where relevant, report date, laboratory, sample stage, tested substances, results, units, methods and appropriate detection or quantification limits. The category determines the required panel. A pass summary cannot supply numerical details it does not show.
Non-detect is not literal zero
A non-detect is interpreted against the stated analytical reporting limit. Without that limit, we cannot infer a numerical ceiling. We also keep raw-material, finished-product and prepared-product samples distinct. A result for one stage does not automatically describe another.
Partial evidence stays visible
A valid result for one substance can be useful even when the complete panel remains unverified. We describe what was reviewed and what is missing. Known detections retain their values, units, dates and sample context; we do not hide a measured result behind the words testing not verified.
Scope and independence
Historical or periodic reports have different limits from a report tied to a current retail lot. We state those limits beside the claim. Unless expressly stated otherwise, Label And Batch reviews evidence supplied or published by manufacturers and laboratories; we did not collect the sample or perform the analysis.
When a report raises a concern
We check the original report, the exact product, the sample and the stated result before changing a claim. A headline or testing summary may point to useful evidence without providing enough detail for a decision. We keep reported concerns separate from results we have verified, and explain any limits or restrictions that follow from the review.
Read the dates separately
A report’s sample and issue dates tell you when the laboratory work happened. A portal update, a page edit and a best-before date mean different things. When a lookup links an older report, we keep the original laboratory date visible and check how the report connects to the product being sold.
Check what the number describes
Read the sample description beside the result. A green bean, a roasted coffee and a prepared drink are different samples. A crop composite combines material from a stated sampling plan; it is not automatically the result for one retail bag. A concentrate diluted into water must retain the amount added and the water volume.
Match the report to your package
Start with the exact product name, flavor or roast, package size and lot code. Keep the best-before date and QR destination with that record. A clear photo of the front, full ingredient label, lot code and QR code helps us check the match. A package photo identifies the item; verification still depends on reviewing the report.
A specification is not a test result
Published specification limits describe a target. A matching laboratory report records what was measured. A certification answers only the questions covered by its scope. Ingredient, material and sourcing reviews alone do not earn a testing verification. We reserve Batch Verified for evidence tied to the named batch; product-level or model-performance evidence keeps its own label.