DMAIC, and the appropriate use of it.

There's a loop the manufacturing world certifies people to run — Define, Measure, Analyze, Improve, Control. If you grow food, you already walk it every season. We didn't build OAT on DMAIC; we built it from the grower's need, and it landed on the same loop from the other side.

You already do this — it just has a name

Walk a season. You notice a problem or set a goal — you define the need. That tells you what to keep an eye on, so you measure it. You read the numbers and decide whether things are in range or it's time to act — you analyze. You're always after a better way for next time — you improve. And when you find one, you make it stick so it happens again without you having to remember — you control it.

That's DMAIC. A grower runs all five. The factory just put it on a wall chart and certified a specialist to run it on an assembly line. Good practice converges — and when a discipline built for machine shops and a craft built for living plants arrive at the same loop from opposite directions, that's a sign the loop is real, not a sign anyone copied anyone.

So we're not here to sell you Six Sigma, and we won't defend it in a greenhouse — a greenhouse isn't an assembly line, and we'll get to where that matters. We're here to show that the bones under what we do are sound and recognized, and then show you the parts we do differently, on purpose.

DMAIC in a minute

DMAIC (say “duh-MAY-ick”) comes out of Six Sigma, the quality discipline that grew up at Motorola and GE in the 1980s and ’90s. Five steps, run as a loop, to fix and steady a process that isn't performing:

• Define — the problem, and what “good” means.
• Measure — how it actually behaves now, with instruments you've checked.
• Analyze — what's really driving the result.
• Improve — change the things that matter.
• Control — make the fix stick so it doesn't drift back.

It earned its reputation. We took it seriously enough to learn it, which is the only honest way to know what's worth keeping.

The five moves — and where this site teaches the how

DMAIC tells you what to do. The rest of this site is the how — the practical, grower-shaped, appropriate how for each move. Read this as a map.

Define — the need comes first. You can't pick a sensor until you know what you're trying to do. This is the rebuke to every catalog that sells you a pile of gadgets before you've named a problem: need first, instrument second. That's the front door of this whole site — what are you trying to do? — not a product aisle.

Measure — and trust your gauge. Here's the one we'd underline twice. A $12 sensor that quietly drifts is worse than no sensor, because it tells you everything's fine while it isn't. The how — what to measure, what to buy, how to wire it, and how to know it's telling you the truth — lives in the Sensing & Data stack.

Have — keep it. (The move DMAIC skips.) This is the one the factory loop has no step for, and it's the heart of how we work. DMAIC measures; it doesn't keep. For a grower, the reading you keep — stamped with when and where, in a form you own and can take with you — is the asset. It's what lets this season learn from last, and what lets the AI you'll have in five years make sense of a crop you measured today. Lose the notebook and the whole loop forgets, so we treat the data as the king it is.

Analyze — make sense of it, now and across seasons. Two speeds. Right now: is this reading in range, or is it time for an alert or an action? And over time: what's the pattern across weeks and seasons? The fast one needs today's numbers; the slow one needs the ones you kept. The how — dashboards, trends, the calculators, and AI where it actually earns its place — lives on the Use side of the site, with the tools.

Improve — get better, this season and for years. We always want a better way: a better recipe, a better routine, a better call. Some of that is now (the vent should crack a little earlier than you thought). Some is later, when enough kept data finally shows you something a single season never could. Improvement here isn't a one-time project — it's the reason you keep the data in the first place.

Control — make it stick. By hand, by machine, or both. This is the move people get wrong about us. Control doesn't mean a wall of automation. It means the better way repeats — by a person following an updated routine, or by a machine doing it for you, whichever fits. When you learn to feed a little earlier and you write it into the recipe, that's control, and it's as low-tech as a sticky note. When you let a controller crack the vent while you sleep, that's the same move with a machine doing the hands — and the industrial-discipline version is there for the grower who runs every dial. A grow recipe is exactly this: it becomes a written routine a person follows, or a device that follows it for you, or both. You pick the one that fits the job.

Where ours goes its own way

Two differences — and they're the whole reason a greenhouse isn't an assembly line.

It never closes. DMAIC is a project: it starts, it ends, the specialist goes home. Growing doesn't end; it cycles. So a grower runs two loops, not one — a fast loop every day (measure → is it in range? → act) and a slow loop every season (look across the kept data → find a better way → fold it into the recipe → sharpen next season's plan). Both turn on the same shaft, and the shaft is the data you kept. That shaft is the step the factory loop is missing — which, not by accident, is the same reason factory gains tend to fade while a grower's get smarter every year.

Every step has a dial. Define can be a thought or a formal plan. Measure can be your finger or a sensor array. Have can be a notebook or a database. Analyze can be your gut or a model. Control can be a changed habit or a programmed controller. The loop runs anywhere along each of those — and the loop is the method; the tech is just how far up each step you reach. That's the truest thing on this page, and it's the whole idea behind OAT: technology isn't the hardware, it's the method — and appropriate technology is setting each dial exactly as high as the need pays for, and not one notch more. A living plant in real weather will never pay back parts-per-million precision, so we don't chase it. Good enough, on purpose, is the right answer here — not a compromise.

Control, and where the liability lives

One caution, because we'd rather say it than hide it. Control is benign right up until two things happen at once: the action gets automated, and someone else is asked to guarantee it. A grower automating their own greenhouse, on their own setpoints, owning the result, with honest failure warnings — that's fine, and it's theirs. What we won't do is stand in the middle of that chain as the one who warrants your crop. We inform, we show, we hand you a reference you can own and run; you keep your hand on the setpoints and the consequences. That's not timidity — it's the same appropriate-technology measuring stick, aimed at our own scope.

Where it lands

We didn't import a methodology. We looked at a respected one the way we tell you to size up any tool you're about to buy: see the whole thing honestly, take the part that meets the need, leave the rest, and don't pay for precision you can't use. The parts worth keeping were already living in the way this site is built — sense it, keep it, use it — and the part the factory loop is missing, keeping the data, is the part we built the whole house around.

A method is a tool like any other. The appropriate move was never to swallow DMAIC whole or wave it off — it was to know it well enough to take the good of it. That's what this page is: appropriate technology, applied to an idea instead of a sensor. If the loop makes sense to you and you'd like a hand setting the dials, there's a quiet door for that — but everything you need to walk it yourself is already here.