In some cases it will, and in others it might not. The reason for this is that different labs generally don’t obtain the same results even from the same flower sample (several published reports show this variation). Depending on the exact procedures your lab uses, the alignment of the Purpl PRO results with your lab may differ. But even if the Purpl PRO provides a different absolute result than your lab, it can still be used for comparisons between flowers, establishing trends in potency, pheno- hunting, etc.

Plant Variety and Averaging

The Cannabis plant has THC and other Cannabinoids unevenly distributed throughout the plant. It is common knowledge that you should avoid smoking seeds and stems. There is the unpleasant taste, but also there's no THC there. Likewise, we know that there's lots in the buds (they get you the highest), and at least a little bit on the leaves (you can see trichomes on the sugar leaves).

When your cannabis is analyzed in the lab, the lab is taking a certain amount of cannabis, homogenizes it by grinding it up, and then dissolves a portion of your submission in solution for analysis. (See Also: How Does Lab Testing Work) The analysis tells you how much of each compound (like THC, CBD, etc) are present in the sample. These results are then representative of the entire batch. However, we know that there are reasons why the sample submitted may not be truly representative: For instance, if you only submitted your very best buds for the sample. (See also: Lab Shopping, THC Inflation) Since the plant itself is also varies a lot from bud to bud, this can also skew the test.

When you test with the Purpl PRO, you are testing that particular part of the plant. Physically, the most important part of the reading comes from the sample that is pressed against the glass in the very center (think of a circle in the middle of the glass, about the size of the holes in binder paper). This means that the reading may jump around a bit. To arrive at a number like the labs do, you must make a number of readings that average out the variety within the plant, and the sample being tested should be a homogenized samples that is representative of what you're trying to test. If you want an overall reading, you need to sample from everywhere. But this also means that you can test to compare potency data from particular parts of the plant (top colas only for instance) to see how they compare. Furthermore, because you don't have to wait for results, and because there is no per-test charge like those that Labs require.

Lab-to-Lab Variation

While labs have expensive equipment and white lab coats, they also introduce error into the process. From our data gathered testing Cannabis around the country, lab tests of the same sample at different labs generally don't match. (See also: Variation in Cannabis Testing). Lab protocols are generally considered trade secrets, and are therefore not reviewable by the public. To address this issue, Purpl publishes White Papers that quantify our accuracy and allow our customers to see how accurate the Purpl PRO is.

Prediction Error

We have invested a lot of resources in achieving accuracy in testing. Our instrument generates a "squiggly line" (or x/y plot of wavelength versus intensity) and then the predictive model looks for THC and CBD peaks. The model uses data we learned from HPLC testing to assign a percentage value to the peak based on its size. These predictions are based on what we've seen in the past, so it's always possible that some new strain appears that interferes with the prediction. We are proud of our claim of Mean Average Error of +/- 2%, but it is an average. Therefore, sometimes we're right-on the mark, sometimes we're off by an even greater margin, but overall, the error averages to 2%. This means you may encounter samples that don't get an accurate prediction. But you will encounter many, many more accurate readings.

We also regularly undertake data campaigns to update and maintain the predictive model's performance for whatever kind of Cannabis currently available in the marketplace.


In the effort to get accurate potency data, constant effort is required to avoid introducing error. But it is also unavoidable. Therefore, we must use some common-sense techniques to get the data we want.

From an end-user perspective, this means following the testing procedure in a careful and repeatable way, being conscious of good sampling, and averaging results.

Purpl Scientific as a company of course has responsibility for this as well. We are responsible for building and maintaining an accurate predictive model. This model has also been evaluated in a blind validation, with published results.