This follow-up to my recent post “Facts and Alt-Facts: How 2 labs Report on Potency” serves primarily as a response to those readers and interested parties that took the time to comment either on how the two charts aligned with their expectations or how my methods were allegedly deceitful.
I’ll put aside the “attack the messenger” tone uniting the latter group of posts, and focus today on the reported “Total Cannabinoid” trends reported by the top-volume Friendly lab and the top-volume PRAGUE lab for the period spanning May, 2015 through Dec. 2016.
The LCB sent guidance to the labs on May 11, 2015 that the number they were to report in the TOTAL field while recording potency for flower was to be THCmax (totally-decarboxylated THC).
During the period between June, 2014 and the first half of May, 2015 different metrics were being reported for TOTAL by different labs.
While Analytical 360 reported TOTAL using THCmax from the beginning, CONfidence ANALytics reported a total of multiple Cannabinoids, a number of which were not among those required for measurement or reporting by the LCB.
This led to CONfidence having larger (and, for a number of months, MUCH larger) TOTAL numbers than many of their competitors.
An earlier post on HI-Blog shares a short animated chart that demonstrates the inflation of TOTAL values reported by each lab compared to the calculated standard of THCmax. The big brown circle labeled lab N corresponds to CONfidence and the initially large blue circle labeled Lab J corresponds to Analytical 360.
Each of these labs has been testing and reporting their results since mid 2014 (near the beginning of this market).
Here is an annotated chart of their reported TOTAL Cannabinoids since May of 2015. This chart corresponds primarily to THCmax (although I believe the LCB began to recently require that TOTAL reflect the summation of both THCmax and CBDmax (the totally decarboxylated measure of CBD + .877*CBD)).
I’ll limit my comments to one overall and one for each labelled phase of the chart.
The basic trend displayed by Analytical 360 during this apples-to-apparently-apples comparison is one of gradually decelerating growth. The trend displayed by CONfidence is best described as cyclical. The low points in this almost-annual cycle correspond to the primary outdoor harvest season seen in October and November.
The OVERALL thing of interest is that Analytical 360 does not share the cyclical trend so clearly displayed by CONfidence.
It is reasonable to assume that CONfidence simply captures a higher share of the OUTDOOR harvests than does Analytical 360 (and that, in the fall, they represent a higher proportion of CONfidence’s testing than they do of Analytical 360’s). Analytical 360’s physical presence in Yakima suggests that they’d capture at least some of Eastern Washington’s outdoor grows.
Overall, the difference between the two seasonal patterns – while striking – is explainable.
Note that it took CONfidence about 6 weeks to come into compliance with the LCB’s May 11 guidance. All but one of the other “inflating” labs snapped into compliance within a week. That gained CONfidence a significant competitive advantage in a growing market rewarding higher cannabinoid levels.
The three months following the LCB guidance on reporting TOTAL as THCmax are the window from which I drew the data used to measure the “Friendlyness” of the labs. While the Cannabis Transparency Project had made some noise about their preliminary lab work earlier in 2015 (April and May), I did not publicize that I was performing a comprehensive comparative assessment of the labs until I published my first lab report in November.
During period B, PEAK Analytics opened for business and rapidly began to set new records for reported average potency. I also published the multi-article series that detailed differences in reporting across the labs. CONfidence hit a low-point at the peak of the fall harvest season (Nov., 2015) and then began showing a rapid increase in potency. Analytical 360 displayed fairly constant slow growth in average potency. That growth appears to have flattened a bit following period B.
During period C, characterized by media attention on the labs due to both my Friendlyness posts and to the Pesticide failures publicized in early 2016, CONfidence continued to grow their reported potency, while Analytical 360 reported fairly steady levels.
During period D, the Liquor and Cannabis Board initiated a workgroup looking into Laboratory Quality Assurance issues. I expressed concerns twice during that time to the LCB Examiner’s office. Once to express concern and caution regarding the appointment of two Friendly labs (CONfidence and Integrity) to the Lab QA Workgroup and once regarding the apparent re-emergence of inflationary potency reporting (I specifically called out PEAK as an obvious outlier).
In the 3 months following this increase in regulatory attention, CONfidence’s reported TOTAL average fell 2.5 percentage points — reaching a floor THREE MONTHS PRIOR to the peak of the fall harvest. I am not sure what might account for this. It certainly seems consistent with an entity capable of and willing to engineer numbers so as to “appear below suspicion” — particularly when compared to the relatively constant output reported by Analytical 360.
Also during period D, CONfidence Analytics created a supposed industry standards group which was said to hold the lofty goal of ensuring and promoting good lab practices – presumably without inflation. There is another story relating to this apparently sham outfit and how CONfidence has leveraged it and some of it’s other tricks to engineer a fairly prominent role in influencing regulatory submissions pertaining to the labs (and to packaging and to labeling and to pesticides and to traceability and to whatever). Perhaps I’ll share that at another time.
During period E, CONfidence maintained a “lower than you” level of reported potency which is noticeably lower than that of Analytical 360.
During period F, the Christmas, Kwanzaa, Hanukkah powers granted the industry a wish and potency once again began to increase.
Let’s thank the potency powers that smiled on the industry in December.
Something or things is different between these two labs.
It may be Friendlyness, it may be the avoidance of punishment, it may be a differential influx of outdoor harvests 3 months early.
Or it may be nothing
.
Best guess? You know my opinion and, for now, I’ll leave it at that.
Just know that my opinion has been shaped primarily by the two years during which I’ve spent a good deal of quality time with the data the labs are spewing out. Since I first published the lab Friendlyness series, the behavior of some of the Friendlies has reinforced that opinion.
I hope you take the time to leave your thoughts regarding this work in the comments section below.
The only other thing regarding labs that I MAY post in the next few weeks is a workup of how important (if at all) the size of a customer is in determining the results each lab reports.
I have put some thought into the patterns of reporting which a lab inclined to engineer it’s results and/or desiring to keep it’s most favored customers happy and/or not wanting to appear to be inflating potencies relative to it’s competitors (or failing to fail bad lots) might be expected to produce.
Those thoughts tend to paint an unpleasant picture for such a lab’s non-favored customers.
But all that is just speculation and theorizing right now.
Really.
I can tell you put a lot of work into this, Jim. Set aside the confirmation bias and selective perception of your posts, at least you’re giving people a forum to discuss.
Confidence does not engineer its results. Not ever, and certainly not on some timeline corresponding to media and LCB attention. Why would we? We use the same method routinely, and we have for years. This “correlation” you’ve built is really engineered by you, as far as I can tell.
I understand your reasoning for wanting the “Total” field in Biotrack to mean THCmax, and the LCB made the same determination back in Spring of ’15. They gave us until July 1, 2015, to comply, and we gave our customers a 1 month+ notice so that they could prepare for the change (as some were using biotrack for label printing). The LCB is only interested in tracking THC and CBD, and they made that official.
Industry players and savvy consumers — on the other hand — we understand that marijuana licensees are not just selling THC and CBD. They sell a complex drug cocktail derived from a botanical commodity. The plant produces an oleoresin composed mostly of cannabinoids, and it is that resin which can’t be legally sourced elsewhere. It’s not just THC and CBD, but also a considerable contribution from CBG and the acid forms of all three (we call them the “big 6 cannabinoids” because they are most of what is produced, by weight). When you sum them together as a rough approximation you get “total cannabinoid” which is a much more meaningful measure of plant productivity and extraction efficiency than THCmax.
From your perspective, Jim, I think you should be bummed that the LCB decided to define “Total” as “THCmax” instead of defining it as “TotalCannabinoid=THC+THCA+CBD+CBDA+CBG+CBGA+(others add small amounts, except in distillates/ other modified)”. Their definition weakens their dataset (and by extension yours) because they no longer have the information necessary to track total cannabinoid production. Traceability could always calculate THCmax, because THC and THCA have always been in the dataset. Starting in July of ’15, the new definition they laid on us began limiting their information collection by replacing a unique field with one that could always be calculated on the fly. A poor decision in my opinion, but it’s not up to me, and it doesn’t stop the industry from still measuring and considering total cannabinoid as a primary metric of production and processing yield.
Cannabinoids and terpenoids are the targets of our industry’s methods, and it makes sense to consider their sums separately, as they have different physical properties. The best part is: within both of those groups there lies tremendous diversity…
I’m finally releasing Nick’s (hopefully) final contribution to the thread.
My thanks for your penetrating, yet Friendly, insight.
I wish you luck in your efforts to have the LCB help you to better compete with PEAK.
Clientele is undeniably a factor. The early downswing is most likely due to light dep greenhouses harvesting too early. Followed by massive outdoor crops being tested October and into November…by December the labs should be back to testing indoor and medical grows and the average would certainly climb
I do appreciate your efforts to analyze and make sense of the data being reported. Clearly not an easy task!
Thank-you (again), Stephanie.
Will Farley & Emily Spahn (now of Amazon), Dr. Dominic Corva (of CASP) and David Busby (of WeedTraQR) have all helped with this work – both through the original Cannabis Transparency Project and some of the transparency work I’m contributing to now through OpenTHC.
I also appreciate all of the folks that have contributed constructive feedback on this work over the past couple of years.
Thank-you, Stephanie.
I’m hoping that the lab in question will let us know by way of commenting back whether such changes in grow-type mix amongst their samples might account for their rapid and oddly-timed decline in reported TOTAL cannabinoids. Until then, the narrative that they are “fleeing a sinking ship” or “being holier than their competitors” (below suspicion) certainly fit the timing and pattern of the data they reported.
The fact that we DO NOT see a seasonal decline in reported potency by Analytical 360 during the fall harvest goes against the notion that outdoor product will produce fewer cannabinoids. The only evidence I see in support of this idea is the dip that the Friendly lab reported that hits bottom in Nov during both years displayed.
My big problem with this disparity is that, with them being a FRIENDLY lab, I tend not to give their data as much credence as I do that reported by a PRAGUE Lab such as Analytical 360. It is my assumption that Friendlyness can degrade the expression of proficiency.
Hopefully, the Friendly lab will speak on something other than the data massaging, alt-fact-producing, selectively-perceiving, bundle of confirmation bias that I supposedly am and will shed definitive light on this question for us all.
(HI-Blog web-site master’s disclosure:
I’m holding another largely ad hominin comment from the Friendly lab in moderation. I’ll release it later this weekend, given my general abhorrence of censorship.
In the meanwhile,I do not want their inappropriate words to pollute the emerging discussion amongst commenters – such as yourself – who are trying to add value to this “thread”.
Thanks again for your comment and observations, Stephanie.
I suspect that what you suggest may well be contributing to the difference reported by the Redmond-based lab.
If so, one might wonder if it is their Friendlyness that is attracting the larger Eastern-Washington outdoor grows to come out West, rather than going to one of the Spokane- or Yakima- based labs. Perhaps these seasonal farms could not get one of the cherished slots with PEAK, and have chosen to go with the next “bestest” option.
Thank you sincerely for sharing the work as well as undertaking it in the first place. I was one of those who requested the TOTAL values.
Here is what I see. Confidence makes a big effort to comply by using THCmax but as the 2015 harvest batches roll in, the real TOTAL values drop to nearly unmarketable levels. All the while indoor harvests are picking up steam and becoming more potent. When their clients consider switching to the new lab, Peak Analytics, they make a business decision. The proceeding upswing could easily be explained by ‘forgetting’ about the THCA conversion factor or even the inclusion of CBD in TOTAL. The ‘D’ period I was very outspoken against unconverted total cannabinoid and perhaps the combined raucus was a deterrent; OR perhaps, as you said, they had already tested all the batches from their favorite outdoor growers… I’m going to take a look at the Confidence THCmax values- perhaps it shows an inverse shape to this graph.
Hmm, hard to say but it seems very possible that the variance comes simply from use/nonuse of conversion factors or inclusion of CBD.
Could be, Andrew.
Completely drying a sample prior to testing also tends to increase the reported levels by about 10% (relative).
Of course, that was something the LCB also advised the labs against doing.
Thank-you
I will just add that I’m sure some producers made changes to inventory and started to grow more potent vs less potent strains to meet consumer demand. I know my company started with what we thought were good strains but had to cut some out and increase production of others based on THC. I know many have done the same. Just food for thought.
Thank-you, Mark.
One of the things I used to enjoy doing was creating dynamic simulations of processes flowing across time.
I’ve been thinking of doing an SD model of “the market” …. but quickly remembered that taking smaller bites was a better way to approach a big problem. Perhaps modelling the various factors that may be feeding into potency (and yield) might be a good place to start.
Input such as yours …. and those of many commenters and others that share their knowledge …. helps to refine my thinking on how that model might be structured.
Thanks again.