Lean Manufacturing & Flow

Lean thinking starts by splitting everything a factory does into two buckets: work that adds value the customer would pay for, and work that does not. Cutting a stud to length, fixing a lining, fitting a window - value. Everything else - moving a panel around the shop, stacking half-built work, waiting for a part, re-doing a defect, walking to fetch a tool - is, in lean language, waste (the Japanese term is *muda*). The startling observation that launched lean is that in most operations the value-adding time is a small fraction of the total; the rest is waste that has simply never been questioned. Lean is the disciplined, continuous effort to strip that waste out so value flows with as little interruption as possible.

The classic taxonomy names the wastes so you can hunt them: over-production (making more or sooner than is needed), waiting (people or parts idle), transport (moving things around more than necessary), over-processing (doing more than the job requires - finishing to a tolerance the joint will never use), inventory (stacks of work-in-progress tying up space and cash), motion (unnecessary movement by the worker), and defects (rework and scrap). Many add an eighth: unused talent - ignoring the knowledge of the people doing the work. Read into a building factory, each one is immediately recognisable, and the figure lists them with a one-line factory reading.

The reason this matters so much off-site is that over-production and inventory - the wastes a factory most easily hides - are lethal in construction. Building modules faster than the site can erect them does not help; it just fills the yard with expensive, fragile, weather-exposed work-in-progress that ties up cash and can be damaged before it is ever used. A site that has not reached the right stage cannot take deliveries early, so a factory that optimises its own output in isolation can actively harm the project. Lean reframes the goal from 'keep every station busy' to 'make the whole system flow to the rate the project actually needs' - which is a very different, and much healthier, target. Holding that distinction is the start of thinking lean rather than merely fast.

Three linked ideas convert a pile of stations into smooth production. The first is takt time - from the German *Takt*, a beat or rhythm. Takt is the pace at which finished units must leave the line to match demand: if the project needs twenty modules a week and the line runs fifty hours, the takt is two and a half hours per module. Takt is the factory's drumbeat. Everything is then arranged so that each station's work fits *within* one takt, so a completed unit rolls off on every beat. Takt turns a vague 'go fast' into a precise, shared target that every station can be tuned against.

The second is flow - ideally one-piece flow, where a unit moves continuously from station to station with as little waiting and batching as possible, rather than whole batches lurching forward in fits and starts. Smooth flow is what keeps work-in-progress low, exposes problems quickly (a stoppage is visible at once), and shortens the time from raw material to finished unit. The enemy of flow is the bottleneck: the one station whose work takes longer than the takt, which throttles the entire line no matter how fast the others run. Finding and relieving the bottleneck - line balancing, the act of trimming or sharing work so no station exceeds takt - is a perpetual job, and the takt/line-balancing figure shows a bottleneck station poking above the takt line.

The third is pull. In a push system each stage makes as much as it can and shoves the output downstream, which breeds exactly the over-production and inventory lean hates. In a pull system, the downstream stage signals upstream - classically with a kanban card or its modern equivalent - that it is ready for the next piece, so nothing is made until it is about to be used. Pull ties the factory's rhythm to real demand instead of to each station's appetite, and it is the discipline that stops a factory from cheerfully over-producing modules a site cannot yet take. Together, takt sets the beat, flow keeps the piece moving, and pull makes sure it only moves when it is wanted - and the push-versus-pull figure contrasts the piled inventory of push with the demand-paced calm of pull.

Pull is the most counter-intuitive lean idea and the one that bites hardest in construction, so it is worth a closer look. The instinct on any line is to keep every station busy: idle workers look like waste, so a supervisor pushes each station to make all it can. Lean says this instinct is wrong, because a busy station that is over-producing is generating the most expensive waste of all - inventory that must be stored, moved, protected, and possibly scrapped if the design or demand changes. Better to let a station stand briefly idle than to build a pile nobody needs yet.

In a building factory the pull signal ultimately comes from the site. The site erects modules in a particular sequence and at a particular rate, constrained by craneage, weather and the assembly programme (Module 7). Lean off-site aligns the factory's output to that erection sequence: the factory makes the next module the site is about to crane, in the order the site needs it, delivered close to just-in-time so it is lifted into place rather than stored. This is powerful - it collapses the yard full of finished modules - but it is also fragile, and here honesty matters. True just-in-time pull demands a reliable site, reliable transport, and a reliable factory; a single delay anywhere ripples through the chain. In many real projects, and especially where logistics are uncertain, teams deliberately keep a buffer of a few finished units as insurance, trading a little inventory waste for resilience. That is a legitimate lean judgement, not a failure - lean is about seeing the trade-off clearly, not dogmatically driving inventory to zero.

This is also where lean construction as a field comes in. Lean construction takes these manufacturing ideas and adapts them to the project world - the Last Planner System for reliable short-term planning, pulling the programme from the completion date backwards, and reducing the variability that makes construction lurch. The translation is imperfect, because a construction project is a one-off with a shifting site and many independent firms, not a repeating factory line - which is precisely why off-site construction, by moving repeatable work into an actual factory, is such a natural home for lean. The deepest link is this: off-site makes construction look enough like manufacturing that lean's factory tools genuinely apply, while the site end still needs the softer, project-adapted lean-construction methods to stay in step with the factory's beat.

The last lean idea is the quietest and the most powerful: continuous improvement, or *kaizen* - the habit of making many small, relentless improvements rather than waiting for one big transformation. Because a factory repeats the same operations, every repetition is a chance to learn: a jig that could locate a part faster, a station resequenced to cut a walk, a check that could be simpler. Lean organisations capture this by listening to the people at the bench (the 'unused talent' waste), running small experiments, and standardising the better method once found - so the line gets a little better every week. Repetition is what makes this possible: you cannot improve a one-off, but you can tune something you do a thousand times. This is one of the understated advantages of off-site - the factory is a learning system in a way a series of unique site projects can never be.

But here is the point that makes lean a design subject and not just a factory-floor one: none of this flow is achievable unless the product was designed for it. Takt time assumes each unit takes roughly the same, predictable work; a line of wildly different bespoke units cannot be balanced to a beat. Flow assumes parts arrive and fit without fuss; poorly toleranced or non-standard parts cause the waiting, rework and motion that destroy flow. Pull assumes a stable, repeatable product the site can sequence. In short, standardisation and DfMA are the preconditions for lean flow. A disciplined grid, a limited kit of repeated parts, connections that go together one way, minimised part count - the whole DfMA programme of Module 3 - is exactly what gives a factory units it can pace, balance and flow. Design a bespoke, variable, hard-to-assemble product and you have pre-loaded the factory with waste no amount of shop-floor lean can remove.

So the designer has enormous leverage over the factory's efficiency long before the factory sees the drawings. Every standard part reused, every connection simplified, every unit made consistent is waste removed at source. The binding specifics - the actual takt, the line balance, the pull system and buffers for a given product and project - are the manufacturer's and the project team's to set, and depend on the system, the site and the logistics. But the *designability* of flow is yours. Lean on the floor and DfMA at the drawing board are two halves of the same discipline: making the building flow.