Perpetual harvest: scheduling rooms and batches

A single flower room finishes its whole crop in the same week, every cycle, then sits part-empty during cleaning and restart before the next one begins. Add rooms with staggered flip dates and the same facility harvests on a fixed interval instead of a single date each cycle. The only design question is how many rooms that interval costs you, and the answer is arithmetic, not guesswork.

Stagger flip dates, not plant counts

A perpetual harvest schedule staggers the date each flower room switches to 12/12 so that batches reach harvest on a regular interval, weekly or fortnightly, instead of the whole facility finishing at once and then sitting idle between cycles. Nothing about plant count or canopy area changes: each room still carries a full crop, just offset in time from its neighbours. What changes is the shape of the output, from one spike a cycle to a steady drip.

"Room" here means whatever unit actually gets its own independent photoperiod and climate control. A single large flower hall split into light-sealed, independently timed zones works the same way as separate rooms in separate buildings, provided each zone can run its own 12/12 clock without light bleeding into its neighbours; the room count below is really a count of independent flip-capable zones, however the walls are drawn.

The governing relationship is simple once the flower-stage length is fixed:

Rooms needed ≈ flower-stage length (weeks) ÷ target harvest interval (weeks), rounded up

Take a 9-week flower cycle and a target of harvesting every 3 weeks. Nine divided by three is exactly three: flip room A in week 0, room B in week 3, room C in week 6, and from week 9 onward one room finishes every 3 weeks like clockwork. That is the textbook answer, and industry guides on running a perpetual harvest describe the same staggered-flip logic: keep a new room's worth of plants moving into flower on a fixed interval so the pipeline never runs dry ^[1]. It is also, on its own, wrong for any real room, because it assumes a room goes straight from harvest back to a new flip with no time in between.

Turnaround time is the week the formula forgets

Between the last plant leaving a room and the next one starting flower there is stripping, cleaning, sanitising surfaces and irrigation lines, and repotting or transplanting the incoming batch. None of that is instant, and none of it can usually happen while the room is still finishing its previous batch. How long it takes depends on what triggered it: wiping down benches and flushing driplines for the next batch of the same cultivar is quick, while a full strip-and-disinfect after a pest flag or between unrelated cultivars is not, and the two get confused when a single "turnaround" figure is used for both. Treat it as an addition to the raw flower length, not a rounding error, and see the sanitation programme for what a changeover actually involves:

Rooms needed ≈ (flower-stage length + turnaround) ÷ target harvest interval, rounded up

Add a 1-week turnaround to the 9-week example above and the total time a batch occupies a room becomes 10 weeks. At a 3-week interval, 10 ÷ 3 is 3.33, which rounds up to 4 rooms, not 3. The extra room is not a safety margin. Trace room A through the schedule: it flips in week 0, flowers to week 9, and is not clean and ready for its next batch until week 10. But the 3-week cadence needs a new room flipping in week 9, and room A is not free yet. Without a fourth room, the schedule either slips a week or someone skips a cleaning step.

• Flower, 9 wk
• Turnaround, 1 wk: room occupied, not growing
• Harvest, with its week

Worked example: 9-week flower cycle, 1-week turnaround, 3-week target harvest interval (article’s own figures).

Scaling to weekly harvest

The same formula scales in either direction. A facility that wants to harvest every single week, still running a 9-week flower cycle with a 1-week turnaround, needs 10 ÷ 1 = 10 flower rooms, or equivalent zoned capacity inside larger rooms, each staggered exactly one week apart.

The interval itself is a decision to make before the formula runs, not an output of it. A fixed trim crew that can process one room's worth of flower efficiently in a week points toward a weekly or fortnightly cadence sized to that crew. A member-distribution cap that empties predictably points the other way: an association whose members draw down stock over roughly a month sees little benefit from harvesting weekly and paying for the extra rooms it costs, where a monthly cadence at 3 rooms covers the same demand. Pick the interval from the constraint that actually binds, labour capacity, sales cadence, member drawdown, or dry-room throughput downstream, then apply the room-count formula to that number rather than to whatever interval sounds impressively continuous.

• 10 wk ÷ interval, the raw result
• Rounded up to a whole room
• The four-room schedule in the worked example

Computed: rooms = (9-week flower + 1-week turnaround) ÷ interval, rounded up to a whole room. Monthly taken as 4.3 weeks.

Monthly intervals (about 4.3 weeks) need 10 ÷ 4.3 = 2.3, rounded up to 3 rooms: two rooms in flower at any moment plus the one mid-turnaround. The relationship is not linear. Halving the interval from a fortnight to a week does not add a fixed number of rooms, it roughly doubles them, because every room still carries the same 10-week occupancy regardless of how often a new one is due.

Feed the flower rooms on the same clock

A perpetual flower schedule only works if propagation and vegetative growth feed it on the same staggered interval. If flower rooms flip every 3 weeks, a fresh batch of transplant-ready clones or seedlings has to be ready every 3 weeks too, which means the veg room's own capacity, canopy area and plant count, has to be sized to that cadence, not to a single batch. Get this wrong and the symptom is not a scheduling error in the flower rooms, it is an empty flower room on flip day with nothing ready to go into it.

The knock-on sizing is straightforward once the veg length is known. A veg stage that runs 4 weeks against a 3-week flip interval means a new veg batch has to start before the previous one has finished: at any moment the veg room is holding roughly 4 ÷ 3, about 1.3 batches worth of plants, one nearly ready to move and a second already growing behind it. Veg canopy area has to be sized for that overlap, not for a single batch passing through, or the room that looked adequate on a single-batch walkthrough runs out of bench space the first time two batches genuinely overlap.

Designing that upstream schedule, how long veg actually needs to run to produce a transplant-ready plant, and how many mother plants or seed lots that requires, is its own decision and belongs with the clone-to-flower timeline rather than here. This page's job is narrower: state the dependency, and size the flower side of it. For the physical room count translated into floor area and HVAC zoning, see sizing rooms for a perpetual harvest schedule.

What staggering buys you, and what it costs

Set against a single synchronised harvest, where the whole facility flips together and comes down together, a staggered schedule trades room count for two things trade sources consistently credit it with: a trim and harvest crew with steady weekly or fortnightly work instead of one intense push followed by weeks of comparatively little to do, and continuous revenue instead of one large batch reaching sale or distribution at a time ^[1]. The same logic extends to utilities, even where it is not usually spelled out: every room's lighting and dehumidification load peaking in the same week is harder on a facility's electrical supply and HVAC sizing than the same total load spread across a rolling schedule. A single-harvest facility of the same total annual output needs fewer rooms, but concentrates all of that labour, cash flow and utility demand into narrow windows.

Splitting one large synchronised room into several smaller staggered ones does not multiply total floor area by much: the canopy has to go somewhere either way. What it multiplies is the count of separately controlled environmental zones, each with its own dehumidifier or air handler, its own light-sealed door, its own sensor set and its own controller, instead of one larger system sized for the whole space. Smaller equipment sized per zone is usually less efficient per unit of capacity than one large system built for the combined load, so the staggered layout tends to cost more in duplicated infrastructure for the same square metreage, even when the room count difference looks modest on paper.

The cost is complexity, and it is not only counted in rooms. When rooms at different growth stages sit close together, a pest or pathogen that would have hit one synchronised crop and stopped at harvest instead has a constant supply of younger plants to move into. Compartmentalisation between rooms at different stages, separate air handling, dedicated tools, staff routes that go from younger to older plants and not back, matters more in a staggered facility precisely because something is always vulnerable somewhere in the building. One widely cited critique of perpetual models makes this point directly: sanitising a room properly, including the strong measures used between full clear-outs, is harder to justify and schedule when neighbouring rooms are always active, and older, near-harvest plants sit weaker and more exposed right as the pathogen pressure from earlier-stage rooms is at its highest ^[2]. Build the IPM programme with that constant adjacency in mind; see the IPM plan for a licensed facility for the room-segregation and traffic-flow controls that keep it manageable. Detailed labour rostering and HVAC load calculations for a staggered schedule belong to the commercial operations and climate sections; this page stops at the scheduling logic that feeds both.

Batches, records and your regulator

Rooms and regulatory batches are usually the same thing, which makes a scheduling decision a compliance decision too. Under Malta's ARUC framework, for example, plants cultivated in separate grow rooms are always separate batches, and a different cultivation method is its own batch even in the same room ^[3]. An association's licensed operational plan has to include a cultivation-cycle schedule and plant counts for the year ahead, and any change to room use, plant numbers or process needs the regulator's approval before it happens, not after ^[3]. That same directive requires the four-eye principle on every reconciliation, processing reconciled against expected yield no later than the following day, a finished-stock count at least monthly, and a quarterly reconciliation report to the regulator covering raw, intermediate and finished stock, distributed product and waste ^[3]. Add a fourth flower room to a three-room staggered schedule and you are not just pouring a slab and running HVAC ductwork, you are filing a change to a licensed plan, and adding another room's worth of daily reconciliation and monthly stock counts to the paperwork you already keep.

The pattern is not unique to Malta. The EU's 2025 rewrite of its GACP guideline for herbal starting materials, the standard EU medicines regulators treat as the compliance floor for cannabis grown as a herbal starting material, requires daily digital records of critical process parameters and unambiguous, batch-level traceability through cultivation ^[4]. Running more rooms on a staggered clock multiplies the number of parallel batch records, environmental logs and reconciliation entries a system has to keep straight without ever mixing one room's history into another's. Plan the record-keeping load at the same time as the room count, not after the schedule is already running.

Size your own schedule

The arithmetic above only needs two numbers from your own operation, not the ones in the worked examples above.

1. Get your actual average flower-stage length, cultivar by cultivar if they differ meaningfully, from your own harvest records rather than a seed-bank estimate.
2. Get your actual turnaround time from a cleaning and sanitation log, not a guess. If you have never measured it, run one cycle with a stopwatch on the process before you commit a room count to a licence filing.
3. Add the two together, divide by the harvest interval your labour, market or member-distribution schedule actually needs, and round up.
4. Check that veg and propagation capacity can produce a transplant-ready batch on the same interval; a flower-room schedule that outruns its own feeder stage is not a working schedule.

A small association aiming for fortnightly member distribution, running an 8-week flower cycle and measuring 4 days (four-sevenths of a week) of real turnaround from its own cleaning log, needs (8 + 0.57) ÷ 2 = 4.28, rounded up to 5 flower rooms staggered two weeks apart, not the 4 the raw 8 ÷ 2 calculation would suggest. The half-room the turnaround adds is exactly the kind of gap between a marketing-cycle number and a licence-filing number this page exists to close.

From there, sizing rooms for a perpetual harvest schedule turns the room count into floor area, and the batch planning worksheet carries the numbers through to a full cutting-to-cure calendar.