Yield for each watt and the cost of a gramYield for each watt and the cost of a gram

Purpose and scope

Most persons speak about plants when they speak about grow rooms. But the result of one division shows if the room can continue to operate. The division is: all the dollars that you pay in one year, divided by all the grams that you sell in that year. If the result is less than the price that you get for one gram, you have a business. If it is not, you have a license but no business. Your room has a loss, and no information about terpenes changes the result.

The problem is that the usual metrics, g/m² and g/W, are for other tasks. They are metrics for agronomy, and growers use them to compare results with other growers. Each one ignores a part of the cost. This paper shows the three usual denominators and the function of each denominator. It also shows the benchmarks from the literature with their limits, because the limits are large.

Then the paper calculates the full list of costs for an example room of 100 m² (1,076 ft²) and shows each step of the arithmetic. After this, the paper shows labor (almost all growers think that this cost is smaller than it is), cycles each year (a number that multiplies the production and that is not easy to see), grades of quality, a tornado chart of the sensitivity, and the break-even.

If you can divide two numbers, you can do all the steps in this paper. Unit economics is the selection of the correct two numbers to divide.

Definitions

Nine terms are necessary for the remaining sections. When two growers calculate the economics of a room and get different results, they frequently use the same term for different fractions.

Evidence and limitations

Summary of the cost for each gram

g/m², g/W, and g/kWh: how to select the correct metric

The three metrics divide the same harvest by three different quantities, and each metric gives you different information. The error is not that you use the metrics. The error is that you use one metric as the primary metric of the business and ignore the information that the metric does not show.

g/m² of canopy is the number for agronomy. It compares crops, cultivars, and steering decisions for the same floor area, and most investigations give this number. It does not include time. Cycles of nine weeks and of twelve weeks can give the same g/m², but the shorter cycle gives 30% more each year. It also does not include power, labor, or the quantities of the grades.

g/W of installed light is a metric from the time when growers used it to compare lamps. Section 06 examines this metric in full. It shows the quantity of crop for each unit of lighting equipment. But the denominator is the nameplate watts. The metric ignores the hours in which the lamps operate and all the energy that the HVAC uses. Most important, it ignores the fixture generation that gives the watts.

g/kWh all-in (or its reciprocal, kWh for each kg) divides by all the kilowatt-hours that go through the meter: lights, HVAC, dehumidification, pumps, and all other loads. It is the one denominator that agrees with the electricity bill that you get from the supplier. It is also the primary metric for benchmarks of facilities. The PowerScore of the Resource Innovation Institute examines a facility with only two numbers. The numbers are kWh for each unit of flowering canopy and grams for each kWh. The PowerScore includes more than 350 growers^[1].

The difference is very large: indoor production uses approximately 18 times the energy for each gram that outdoor production uses^[2]. Thus this metric is very important for an indoor room. It is less important for a greenhouse.

Yield benchmarks from the literature and their limits

The values for the yield of cannabis in the literature have different conditions and different denominators. Some values are projections and not measurements. Before you compare your room with a benchmark, examine the sources and the size of the differences.

The baseline from police data. The best large-sample value for g/m² in the literature is also the first value in time. Police in the Netherlands weighed the plants from illegal grow rooms that they found. The model for the median room has 15 plants/m² at 510 W/m² of HPS, and it gives 33.7 g for each plant and 505 g/m²^[3].

The division of 505 g/m² by 510 W/m² is 0.99 g/W. It is very possible that this one investigation is the source of the value ‘one gram for each watt’ that growers frequently give. The investigation is twenty years before this paper, and it gives a median value for HPS rooms.

The tests with careful conditions. In the tests of Potter and Duncombe, the HPS lamps operated at 270, 400, and 600 W/m². The tests measured 0.9–1.6 g/W, and the best result in g/W was at the lowest power^[4]. More light gave more grams, but the value in g/W was lower. The tests measured how the grams for each unit of power become smaller when the power increases.

Tests with new LED fixtures show the same result from the other side. The yield of dry flower continued to increase almost in proportion to the light intensity, to a maximum of approximately 1,800 µmol·m⁻²·s⁻¹. There was no plateau^[8]. A second test at 600–1,000 µmol found that each 100 µmol more gives approximately 4.6 g for each plant (approximately 51 g/m² at approximately 10 plants/m²). The yields in this range are approximately 276–447 g/m²^[6]. The group of Bugbee did three tests with hemp at high light for cannabinoids and measured 500–750 g/m² in the three tests^[7].

The meta-analysis, and the limits of the data. Backer and the other authors put together the data from the literature. They found efficiencies in the reports of 0.31–1.97 g/W, which is a range of 6 times. They also found projections of yield for large areas from 3.4 to 3,590 g/m², which is a range of approximately a thousand times. The cause is that the projections use numbers from small test areas for large areas that did not have a crop^[5].

They also found that the yield for each watt decreased when the installed W/m² increased. They found that the yield for each m² increased when the time of flowering was longer. In the two results, the cause is the denominator and not the plant.

Grams for each watt: a lighting metric from the time of HPS

The value ‘one gram for each watt’ was a good approximate value when all the rooms used the same lamp. With LED, the metric shows when you got your fixtures, because a watt in the denominator gives more photons than it gave with HPS.

The efficacy of a fixture is the quantity of photons that the fixture gives for each joule of energy. The unit is µmol of photons for each joule. Double-ended HPS, the lamp type for the value ‘one gram for each watt’, gives approximately 1.72 µmol/J.

In 2020, the best LED fixtures that growers measured gave approximately 3.0 µmol/J (blue/red) and 2.78 (white/red). The practical limits are approximately 3.4–4.1^[9]. In 2014, the best LEDs gave 1.7, which is equal to HPS. In one fixture generation, the photons for each watt increased to approximately two times the previous value.

Examine the effect on g/W when there is no change in agronomy. Use the same example crop, 450 g/m² at 900 µmol·m⁻²·s⁻¹. HPS at 1.72 µmol/J supplies 900 µmol with 900 ÷ 1.72 = approximately 523 W/m². An LED at 2.8 µmol/J supplies the same photons with 900 ÷ 2.8 = approximately 321 W/m².

The photons, the plants, and the grams are the same. The HPS grower calculates 450 ÷ 523 = 0.86 g/W. The LED grower calculates 450 ÷ 321 = 1.40 g/W. The agronomy of the two growers is the same.

The change in efficacy also changes the decision about the purchase of fixtures. In the lighting tests of Bugbee, the white+red LED gave 4.6% less yield for each m² than HPS. It gave 27% more yield for each dollar of electricity^[7].

If you use g/m², the LED is worse. If you use the yield for each dollar of electricity, which is a cost metric, the LED is better. The data are the same, but the denominator is different and the decision is opposite. This one test shows the primary result of this paper.

Cost items for each gram

The cost for each gram uses a short list of items. The method is easy, but you must include all the items. Eight items are sufficient for a small indoor facility:

• Labor: the wages plus the other costs that you pay with the wages (paid leave, insurance, and tax). The item includes all personnel who touch the crop, also you, at a market rate.
• Energy: lights, HVAC, dehumidification, pumps, and controls. Include all of the energy. Use the value on the electricity bill and not the value on the fixture nameplate. Two numbers show how large this item is for indoor production. An estimate for the US shows that indoor production used 1% of all the electricity of the US approximately ten years before this paper^[10]. The emissions that models give are 2,283–5,184 kg CO₂e for each kg of flower, and they change with the climate^[11].
• Media + nutrients: substrate, salts, CO₂, and consumables for IPM.
• Rent: for the total floor area and not for the canopy area.
• Depreciation: the fit-out and the equipment, divided by the number of years in which you use them.
• Testing + compliance: the panels of laboratory tests for each batch, plus the licenses, the QA time, and the records. Data from California show a cost of approximately $136 for each pound (approximately $0.30/g) for the mandatory testing only. This cost includes the sampling and the failure rates^[12]. This item is large and you cannot ignore it.
• Packaging + consumables: bags, totes, labels, and gloves.
• Other overhead: insurance, security, administration, repairs, and software.

The example room of 100 m²: all the steps

1. 1
   Calculate the installed power

   100 m² × 350 W/m² = 35,000 W = 35 kW installed. Do a check of the light intensity: 350 W/m² × 2.6 µmol/J = 910 µmol·m⁻²·s⁻¹. This light intensity is a usual target for LED in flowering.
2. 2
   Grams for each cycle

   450 g/m² × 100 m² = 45,000 g for each cycle.
3. 3
   Cycles each year

   56 days of flowering + 7 days of turn time = 63 days of cycle length. 365 ÷ 63 = 5.8 cycles each year.
4. 4
   Grams each year

   45,000 g × 5.8 = 261,000 g = 261 kg each year.
5. 5
   Lighting energy

   35 kW × 12 h × 56 days = 23,520 kWh for each cycle of lighting.
6. 6
   All-in energy

   23,520 × 2.2 = 51,744 kWh for each cycle. With 5.8 cycles, the result is approximately 300,000 kWh each year. Do a check: 300,000 ÷ 261 kg = approximately 1,150 kWh for each kg. This value is low for indoor production. Many rooms operate at 2 to 4 times this value^[2].
7. 7
   Calculate the cost of energy

   300,000 kWh × $0.20 = $60,000 each year.
8. 8
   Add the other cost items

   Labor $200,000 · rent $60,000 · testing + compliance $46,000 · depreciation $43,000 ($300,000 ÷ 7) · other overhead $40,000 · media + nutrients $26,000 · packaging $20,000. With energy: $495,000 each year.
9. 9
   Divide

   $495,000 ÷ 261,000 g = $1.90 for each gram of product. The cost for each gram is the primary metric of the room. All the other parts of this paper are methods to change it.

Use each denominator from Section 04 for the same room. The table shows the information that each denominator gives:

Labor costs

When you speak to a new grower about the cost of indoor production, the grower speaks about electricity. In the example room, the electricity bill is $0.23 for each gram. The cost of labor is $0.77 for each gram. It is the largest item, and it is three times larger than the next item. Many estimates do not include labor, or they give it a value of zero, because the grower thinks ‘I will do the work myself’.

Start with the division: $200,000 of payroll divided by 261 kg is $766 for each kg. At a cost of $25/hour with all the other costs, the result is approximately 31 hours of work for each kg of product. Most of these hours are for one task: hand trimming. The usual rate of a hand trimmer is approximately 0.45–1.4 kg (1–3 lb) of dried flower in an 8-hour shift.

The wage is $15–20/hour, or $100–200 for each shift as a piece rate^[13]. Divide the hours of the shift by the rate: the result is approximately 6–18 hours for each kg for trimming only. Use 10 hours. At $25/hour with all the other costs, the cost is $250 for each kg, or $0.25 for each gram, for trimming only. The cost of trimming is more than the cost of electricity.

The sum of the tasks is approximately 18 h/kg, but the payroll gives approximately 31 h/kg. The 13 missing hours are for work that does not touch a bud. Examples are the work with mother plants and vegetative plants, meetings, cleaning, records, sick leave, and time between tasks with no work. The difference is the utilization. Thus a model that calculates the number of personnel from a list of tasks always gives a labor cost that is less than the payroll. Calculate the labor cost from the payroll, and use the task minutes to find the items that you must change.

• Measure before the purchase. A trim machine with a rate of 9–18 kg/h (20–40 lb/h)^[13] is much faster than 0.9 kg for each shift (2 lb for each shift). Before you pay for the machine, examine the grade of the product that it makes. Then examine if your buyer accepts this grade (Sections 11 and 14).
• Make the peaks of work smaller. In the weeks of harvest, you must have 3 times the personnel of week 3 of flowering. If the rooms start flowering at different times (Section 10), the problem of the number of personnel becomes a problem of the times of the tasks.
• Include the cost of your hours. If the cost of your hours is $0, each room that has a loss shows a profit in the model.

Crop cycles each year

All that you make in one year is the grams for each cycle × the cycles each year. Growers monitor the first term and ignore the second term. The turn time is the number of days from the harvest of one crop to the start of flowering of the next crop. The turn time changes the cycles each year, and thus it changes all the grams that you make in one year.

The effect of the turn time is large because it multiplies the grams of each cycle. With a turn time of 7 days, the room operates 5.8 cycles each year and sells 261,000 g. If the turn time becomes 21 days, the room operates 4.7 cycles and sells 213,300 g. Causes can be a slow cleaning, clones that are not available, or a part that you get one week after the correct date.

Two more weeks in each turn time give 47,700 g less each year. At a blended price of $2.20, the revenue is more than $100,000 less, and the cost does not decrease. No feed schedule has an effect of this size.

In the meta-analysis of Backer, a longer time of flowering increased the yield for each m²^[5]. You must calculate the effect of a longer flowering time correctly. One more week of flowering must give more grams than the same week gives in a new cycle. At 45,000 g for each cycle, a cycle length of 63 days gives approximately 714 g for each day. A cycle length of 70 days must give approximately 50,000 g for each cycle, which is 11% more, to give the same result. Do this division before you make the ripening longer, and not after the change.

1. 1
   Write the cycle length

   Write the cycle length in days on the whiteboard. If you do not measure it, it can become longer. No person knows that the turn time becomes one day longer in each cycle.
2. 2
   Prepare before the harvest

   Before the day of the harvest, complete the repair list. Put the consumables for the room in position. Make sure that the personnel for cleaning are available. Do the work of the turn time in a short time. It is not a long task.
3. 3
   Prepare the vegetative plants first

   The most frequent cause of a long turn time is clones that are not available. The vegetative room must operate one full cycle before the dates of the flowering room.
4. 4
   Start the rooms at different times

   Four small rooms that start flowering one after the other give the same cycles each year as one large room. They make the trim labor equal in all weeks. A failure of one crop is 25% of the production and not 100%.

Higher prices for quality and the quantity of yield

The cost for each gram is only half of the information. Your revenue changes when the price for each gram changes, and the price is different in each group. In the US spot market at the start of 2024, the average price for indoor flower was approximately $1,378/lb ($3.04/g). The prices for greenhouse flower and outdoor flower were approximately $725/lb ($1.60/g) and $418/lb ($0.92/g)^[14]. The difference of the prices for the methods of production is 3 times, before you include the grades. In each method, the prices are different again for each grade: grade A flower, grade B flower, small buds, and trim.

Thus the model must use the blended price and not the price of the best grade. A decision to sell only the best grade changes all the model and not only one item. Compare two alternatives for the example room, which is almost at the break-even at a blended price of $1.90:

Sensitivity of the cost for each gram

Before you pay for a change, use the model to find the input with the largest effect. Start with the baseline of the example room ($1.90/g). Change one input at a time in a possible range. Keep all the other inputs the same, and calculate again. Show the results in a chart with the widest range at the top. This chart is a tornado chart:

The sequence of the bars is the primary information. Compare a change of 20% in the yield with a change of half (less or more) in all the cost of nutrients. The yield has an effect that is four times larger. The two largest bars, yield and labor, show the performance of the grower and the quality of the procedures in the room. The bars that suppliers speak about most (price of electricity, CAPEX, bottles) are the small bars.

The sizes of the changes are also important. A change of 20% in the yield can occur because of one cycle with a pest problem or one error in steering. A change of 50% in the price of electricity occurs only if you make a new contract with the electricity supplier. The large bars are also the easy bars to change, in the two directions.

How to calculate the break-even

Break-even is the yield, the price, or the number of cycles at which the profit becomes zero. When you know the break-even, you have targets with numbers. There are three divisions for the example room:

• Break-even price at 450 g/m² and 5.8 cycles: $495,000 ÷ 261,000 g = $1.90/g blended price. If the price is less than this value, you have a loss for each gram that you sell.
• Break-even yield at a blended price of $2.20: $495,000 ÷ $2.20 = 225,000 g, then ÷ (100 m² × 5.8) = approximately 388 g/m² for each cycle. This yield is the minimum. A lower yield gives a loss.
• Break-even cycles at $2.20 and 450 g/m²: 225,000 ÷ 45,000 = 5.0 cycles. Thus the cycle length must be less than 365 ÷ 5.0 = 73 days. The cycle length has a maximum value.

Use two methods to make the break-even a good tool. First, calculate it for each limit (a minimum price, a minimum yield, a maximum cycle length). Thus each person has a number that the person can change.

Second, calculate it again after each change. Costs increase slowly, prices decrease slowly, and a large profit of last year can become the break-even of this year without one large change. In mature markets, wholesale prices usually decrease^[14]. In the model, use a price range that decreases and does not increase.

Frequent errors in unit economics

A business can continue after one of these errors, but not if the error occurs many times. All of them are errors of denominators or missing cost items. None of them is an error of agronomy.

Troubleshooting

Find the symptoms first and then the causes. Do the same for a plant with a disease. But the bank statement is the sensor, and the time to get a reading is three months.

Values that you can change in unit economics

Do these tasks this week, in this sequence:

1. Make a list of your costs for the last twelve months. Include all eight items and the cost of your hours at the market rate.
2. Divide by the grams that you sell in the same twelve months. Write the $/g result in a position where all personnel can see it.
3. Write the cycle length in days on the whiteboard. Then record it for each cycle.
4. Measure the time of one full trim shift and of one full harvest day. These two tasks are your largest labor items. Do these two measurements before you make a decision about a machine.
5. Make the tornado chart with your numbers. Change the sequence of your CAPEX list to the sequence of the differences in the chart.
6. Calculate again after each period of three months. Costs increase slowly, prices decrease slowly, and the model is correct only when its numbers are new.

The benchmarks show that you must be careful with numbers from other growers. The literature has a range of 0.31–1.97 g/W^[5] and a difference of more than two hundred g/m² in the results of correct investigations^[6][7]. The number of a different person, also the $1.90 of the example room, is not your number. You can use the method in all rooms, but the results are different in each room.