A laboratory report may include a recovery percentage alongside the measured results. These numbers answer different questions. A concentration describes the reported amount of an analyte—the substance being measured. A recovery percentage describes performance in a particular analytical check.
This guide explains how to read that distinction without turning a quality-control result into a claim about a sample’s overall quality. The environmental and tissue-analysis sources below illustrate analytical principles; they do not establish EU cannabinoid-testing mandates.
What is a matrix spike?
A matrix spike is a portion of a sample to which the laboratory adds a known amount of a target analyte. In EPA’s environmental-testing description, that addition happens before preparation, cleanup and measurement. The laboratory compares the spiked portion with an unspiked portion to examine method performance in that sample matrix. See the EPA’s project quality-assurance and quality-control guidance.
The useful question is not simply “Did the instrument detect the addition?” It is “How did the analytical procedure perform when the added material encountered this sample?” For background on why the surrounding material matters, read laboratory methods and sample matrices.
What does recovery mean?
For a matrix spike, the usual calculation subtracts the unspiked result from the spiked result, divides the difference by the amount added, and multiplies by 100. The quantities must be expressed on a compatible basis. The EPA’s tissue-analysis QC guidance presents this relationship.
For an arithmetic illustration, suppose the unspiked result is 10 units, the spiked result is 28 units, and the addition corresponds to 20 units. Recovery is (28 − 10) ÷ 20 × 100 = 90%.
That 90% describes recovery of the addition in the check. It does not mean the original sample is 90% pure, contains 90% analyte, or meets 90% of a safety standard. The calculation is not a product-composition statement.
How is a laboratory control sample different?
A laboratory control sample, or LCS, checks performance in a clean reference matrix rather than the submitted sample’s matrix. Comparing its result with matrix-spike recovery helps the laboratory investigate an unexpected finding. The EPA’s QC explanation identifies inconsistent preparation, matrix interference and instrumentation problems among possible causes of poor performance.
An acceptable LCS does not erase a matrix-spike problem. Nor does a matrix-spike failure, by itself, establish precisely what went wrong. Ask the laboratory for its interpretation of the combined evidence.
Is there a universal acceptable percentage?
No single recovery interval should be applied to every report. The applicable method and documented quality requirements matter. EPA’s project QC guidance distinguishes method-specific criteria from more general guidance and places acceptance criteria within project planning.
A recovery number without its applicable limits is therefore incomplete information. Ask which limits were used, where they came from, and whether the reported value met them. Do not import a range from another laboratory, analyte or testing context merely because both reports use the word “recovery.”
What can high or low recovery tell you?
Relative to 100%, low recovery means the measured increase was smaller than the known addition; high recovery means it was larger. Neither observation is a complete diagnosis. High apparent recovery is not evidence that the sample somehow created extra material, and low recovery is not automatically a percentage by which every reported result is wrong.
A further limitation is representativeness. Newly added analyte may behave differently from analyte already incorporated into the sample. The IUPAC/ISO/AOAC International/Eurachem recovery guidelines explain that differences in binding, matrix and concentration can limit how confidently spike behaviour represents native analyte. A good spike result is useful evidence, not proof that every relevant analytical effect has disappeared.
Should you correct the result yourself?
No. Dividing a reported concentration by a recovery fraction may look straightforward, but deciding whether that correction is appropriate belongs to the validated analytical procedure. The 2026 Eurachem/CITAC uncertainty guide explains that recovery-correction decisions are normally established before method implementation and documented. It also notes that corrections based on few recovery tests carry greater uncertainty than those based on more tests.
Ask whether results are already recovery-corrected and how recovery is reflected in the laboratory’s uncertainty evaluation. Avoid producing an unofficial replacement result.
Questions worth asking about a report
- Was this spike prepared from my sample or another sample associated with the batch?
- Which analyte, addition level and preparation stage does it represent?
- What were the applicable acceptance limits and the LCS result?
- Did the laboratory investigate, repeat analysis or qualify any affected results?
- Which report notes explain the consequences for interpretation?
Read those answers alongside other controls, including laboratory blanks and report notes. Recovery supports an assessment of analytical performance; it does not independently guarantee that a material is safe, legal or suitable for any particular use.