Planning cultivation from member allowances, not demand guesses

Cultivate for "members times the legal maximum times 30 days" and the arithmetic itself will mislead you. A 120-member Malta-style association working from a 50 g monthly cap ^[1] gets 6,000 g as its target, and almost no month will see the membership actually draw anything close to that. Grow to the ceiling and the surplus sits in storage as a spoilage and security problem the association cannot legally sell down or transfer away. Grow to a guess instead, and a bad month means rationing members at the counter, precisely the outcome the harm-reduction model exists to prevent.

Three numbers, not one: the ceiling, the forecast and the yield

A cultivation plan sized from a single estimate, done once at licensing and left alone, is already wrong by the second cycle. Run three lines instead, as a standing dashboard the cultivation lead reviews every cycle, not a one-off spreadsheet:

• Legal ceiling. The worst case the law allows: every member drawing the full monthly amount, every month. This number barely moves. It only changes when membership size or the cap itself changes.
• Forecast demand. What the membership is actually likely to draw, built from the association's own distribution log: who came in, how much they took, how that has trended over the last few cycles. This is the number that should drive the grow plan.
• Planned grow-cycle yield. What the cultivation team has actually committed the current canopy plan to produce, given the rooms, cultivars and schedule already in motion.

The plan is healthy when the yield line tracks the forecast line. The ceiling line is there to bound the worst case, not to size the canopy: a plan built to the ceiling is a plan built for a scenario that essentially never happens.

Why the forecast sits well below the ceiling

Start with the cap arithmetic itself, before any forecasting. A member may draw 7 g a day, but only 50 g a month ^[1], and 7 multiplied by 30 comes to 210 g: nobody can legally draw the daily maximum for a full month, so the monthly figure is already the binding number, not the daily one. Multiply that monthly ceiling by the membership and you have the worst case the law allows, not a demand estimate.

Actual demand depends on how many members visit the distribution counter in a given month and how much each one draws, and both numbers vary member to member and month to month. Most members do not consistently draw near the cap; a handful might, most draw well under it, and some months a member draws nothing at all. None of that shows up in the ceiling arithmetic, only in the association's own dispensing record.

Building the forecast line without letting one odd month steer it

The forecast line is not a ratio you fix once and reuse. It is a rolling calculation the cultivation lead recomputes every time a new month of dispensing data lands, and it needs rules, or every unusual month will yank next cycle's canopy plan around with it.

Start with a minimum data bar: don't trust the forecast line until you have at least two completed grow cycles of the association's own distribution records behind it. One cycle cannot tell you whether a high or low month was a trend or a one-off; two gives you a second point to check the first against. An association without that history yet should size the current cycle against the legal ceiling as a stress-test cap only, while it accumulates real numbers, exactly as the check below routes it.

Once there is history to work from, three rules keep the line honest:

• Use a trailing average, not the latest month alone. A rolling average of the last two to three cycles smooths out a single unusual month, a member who batch-bought before a holiday, a strain launch that briefly spiked interest, without erasing a genuine trend.
• Weight recent cycles more while membership is still growing. New members rarely draw near the cap from month one; most ramp up over their first few cycles as they learn the product range and settle into a routine. A membership that grew by 20% last quarter will keep pulling the average up for several months after the growth itself has levelled off, so lean on the two or three most recent cycles rather than an all-time average.
• Treat a genuine step change as a reset, not an input to smooth over. If the association just opened the waitlist and admitted a new cohort, started distributing resin for the first time, or changed its opening hours, don't fold the cycles before that change into the same average as the cycles after it. Recompute the ratio from the post-change cycles alone, even if that means trusting fewer data points for a while.

A practical outlier rule: if one month's total draw sits more than roughly a quarter away from the trailing average in either direction, treat it as a flag to investigate before the next canopy decision, not as fresh evidence to fold straight in. Check whether it was a stockout that suppressed demand, a strain nobody wanted sitting unmoved, or a real shift in the membership's behaviour, then decide how much weight the new figure deserves once you know which it was.

Grams stop being one unit once resin is on the shelf

Since May 2025, Malta's daily and monthly caps cover flower and processed product together, not flower alone, and since January 2026 the conversion between them runs on three THC bands rather than one flat ratio ^[2]. A "gram distributed" is not one flat unit once a second product form exists on the shelf:

A member who draws 10 g of resin tested at 70% THC spends 30 g of their 50 g monthly allowance, not 10 g. The same 10 g at 45% THC spends 20 g; at 25% THC it spends 10 g, gram for gram. Get this wrong in a quota ledger and a member who looks well under their monthly limit on a raw-gram count can already be over it once their resin draws are converted.

This is also the second time this single conversion factor has moved: the January 2026 tiers replaced a flat 1-to-3 ratio that had only been in force since May 2025 ^[1][2]. A forecast and a ceiling built around resin distribution can move without a single plant changing, purely because the accounting rule changed under them.

Plan two grow cycles ahead of the shortfall, not one week ahead of it

A flower cycle runs for months from flip to cured, distributable stock, so a shortfall spotted at the distribution counter this month was already decided by a canopy allocation made roughly two cycles ago. Reviewing the three-line dashboard weekly is good practice for catching a trend early, but the decision it feeds, how much canopy to allocate, which rooms flip when, has to be made against the forecast for the cycle that has not started yet, not the one already growing.

Treat any gap between the forecast and planned-yield lines as a decision for the next cycle to start, not a problem for the current one to somehow absorb. A current cycle already in flower cannot add canopy; the only lever left once a shortfall shows up mid-cycle is rationing, which is the outcome this whole exercise exists to avoid.

There is no safety valve: the closed-loop rule

Associations "may not procure or transfer cannabis plants, dried cannabis, finished or semi-finished products to or from other individuals, associations, or entities" ^[3]. A shortfall cannot be bought in from a sister CHRA down the road, and a surplus cannot be sold down to one that is short. Every association is its own closed loop.

That is what makes the three-line dashboard non-optional rather than a nice-to-have spreadsheet. There is no wholesale market behind it to absorb a bad estimate in either direction, only storage and spoilage risk on one side, and a queue of members turned away on the other.

Worked example: sizing a 120-member association from ceiling to forecast

Take a 120-member Malta-style association with a stable membership over a six-month planning window. Its legal ceiling never moves: 120 x 50 g comes to 6,000 g a month, whatever the membership actually draws ^[1]. Its forecast line comes from its own distribution log, and for this worked example we assume a draw ratio that climbs from a modest quarter of the ceiling in month 1 to about 41% in month 4, as members who joined earlier finish onboarding and start drawing regularly. That growth pattern, and the exact percentages, are a stated assumption for this worked example: replace them with your own association's dispensing history before you plan a real cycle from them.

The planned-yield line moves in steps, not smoothly, because canopy allocation is only revisited when a new cycle begins, roughly every two months here, while the forecast line updates every month as fresh dispensing data lands. That mismatch in update frequency is the whole story behind month 4: the cycle now supplying month 4 was planned back when the draw ratio still looked like month 1 and 2's low-30s trend, so it was sized to 1,900 g. By the time month 4 actually arrives, forecast demand has climbed to 2,450 g, a gap of 550 g, about 22% below forecast and outside the 10–20% band a dashboard like this should treat as normal month-to-month noise.

Run the alternative for comparison: had this association simply grown to its full legal ceiling every month instead of tracking forecast demand, six months of canopy would have produced 36,000 g against a forecast of 12,100 g, very nearly three times over. The extra 23,900 g does not disappear once it is harvested; it becomes stock the association still has to count, log and guard, and cannabis products that stay undistributed have to be retested at six-month intervals ^[2], an added cost that a lean, forecast-sized plan never incurs in the first place. Over-producing is not a victimless safety margin; it is a standing storage, security and retesting bill with no legal buyer on the other side of it.

A gap of that size, caught early, is a planning fix: the cycle that should have covered month 4 needed to be sized against a forecast, not against month 1 and 2's numbers. Caught late, at the distribution counter in month 4 itself, it is a rationing problem with no legal way to top up from outside the association. The check below is the same logic as a standing rule you can run after every cycle review.

Handing this number to the room-by-room schedule

Getting the forecast and yield lines to agree is necessary but not sufficient. Turning 2,450 g of forecast demand into an actual cultivation schedule, how many flower rooms, how much canopy area, when each one flips, is a separate piece of arithmetic covered in the association cultivation calendar. This page's job stops at the target number: get the ceiling, forecast and yield lines right, reviewed every cycle, before that room-by-room schedule gets built around them.

Everything downstream assumes this number is already correct. Modelling whether new members fit inside legal and canopy headroom starts from the same three lines before a single application is approved. Choosing cultivar variety for a membership's actual preferences allocates canopy inside the yield line once it is set, not before. Uruguay's clubs run the same underlying problem with a plant-count ceiling instead of a gram ceiling; see Uruguay's membership clubs for how the arithmetic changes when the cap is plants, not grams.