A Filosofia Just In Time E Suas Ferramentas Associadas

Production doesn't run on hope

When I first started dealing with lean manufacturing systems, I treated Just in Time like it was some kind of miracle approach to inventory. That lasted about six months before a supplier delay took down three assembly lines and I learned the hard way that JIT is more about discipline than magic. The philosophy itself isn't complicated. You produce or order goods only when they're actually needed, not based on forecasts that usually turn out wrong. The goal is to eliminate waste, especially the waste of capital sitting in warehouses collecting dust. Associated tools include kanban systems, single-minute exchange of die (SMED), total productive maintenance (TPM), poka-yoke for mistake-proofing, heijunka for production leveling, and value stream mapping.

Most people skip straight to kanban cards and think they've implemented JIT. They haven't. I watched a factory in São Paulo try this around 2019. They put up colored cards at every workstation and stopped ordering raw materials on schedule. Within eight weeks, they were running out of steel twice a month because nobody had bothered to renegotiate lead times with their suppliers. The cards were just decoration by then.

The philosophy just in time e suas ferramentas associadas na prática

Here's what actually happens when you implement this properly. You start with value stream mapping. You walk the floor with people who do the work, not managers looking at spreadsheets. You track the actual flow of material from receipt to shipment and you measure cycle times, changeover times, and waiting times. Most operations I've seen have a ratio of value-added time to total lead time somewhere between 5 and 15 percent. The rest is movement, waiting, and inspection. SMED comes next. You need to cut your changeover times dramatically before you can do small batch production. A typical stamping press changeover might take two hours. After SMED work, that can drop to fifteen or twenty minutes. This is where the methodology gets real. You film the changeover. You separate internal steps that require the machine stopped from external steps that can be done while it runs. You convert as many internal steps as possible to external. You speed up the remaining internal steps with standardized work, preheated molds, and quick-clamp fixtures instead of loose bolts.

Pok a-yoke is usually misunderstood too. People think it means putting sensors everywhere. It doesn't. It means designing the process so that mistakes become physically impossible or immediately obvious. A fixture that only allows a part to be loaded one way. A color-coded hose connector that can only attach to the matching port. A weight sensor that triggers an alarm if the wrong number of components goes into a package. Simple things that prevent defects from reaching the next station. I ran into a specific problem with one client where the JIT system worked perfectly until the monsoon season hit. Their supplier was four kilometers away but located in a flood-prone area. Every year between June and September, the road got washed out and they had no inventory buffer because they'd eliminated it as part of the JIT implementation. Production stopped for two to three days at a time. The workaround wasn't to add inventory back, which would have defeated the purpose. We restructured the supply chain by adding a second supplier on higher ground, forty kilometers away, and set up a cross-dock facility between them. Lead times went from twelve hours to eight but supply became reliable. That's the kind of trade-off people don't tell you about in the textbooks.

Heijunka, production leveling, is probably the most undervalued tool in the JIT toolkit. If you produce in batches based on order sequence, you create waves of demand that ripple through your entire operation. One day you're running at 120 percent capacity, the next at 60. Leveling means producing a mix of products every day in consistent volumes. It sounds counterintuitive because you'd think meeting customer demand exactly as it comes is more efficient. It's not. The variation costs more than the temporary overproduction that leveling sometimes requires. Here's something beginners consistently miss about TPM. They treat it as a maintenance program. It's not. Total productive maintenance is about maximizing equipment effectiveness, which includes availability, performance, and quality. OEE, overall equipment effectiveness, is the metric that matters. You calculate it by multiplying those three factors. A machine running at 90 percent availability, 80 percent performance speed, and 98 percent quality rate has an OEE of 70.6 percent. The global benchmark for world-class manufacturing is around 85 percent. Most factories I audit sit between 40 and 60 percent. The gap isn't usually big breakdowns, it's micro-stops and reduced speed that operators have learned to ignore.

Another thing that doesn't get enough attention is the relationship between JIT and takt time. Takt time is the rate at which you need to complete a product to meet customer demand. It's calculated by dividing available production time by customer demand. If your customers need 480 units per day and you have 480 minutes of available time, your takt time is one minute per unit. Everything in your process has to align to this rhythm. If one station takes 90 seconds and the next takes 75, you have imbalance. If a machine has a 30-second cycle time but you need output every 60 seconds, you might run it at half speed or duplicate it. The whole system is designed around this pulse. One practical implementation sequence that tends to work: value stream map your current state, identify the biggest sources of waste, implement SMED on your bottleneck equipment first, set up heijunka boxes to level your production mix, introduce kanban at the point where material flow is most unstable, then expand outward. Don't try to do everything at once. I've seen companies roll out kanban across twenty production lines in a single week. They ended up with kanban cards everywhere and chaos everywhere. It took them six months to untangle it.

There are also scenarios where JIT doesn't make sense and you should just accept that. High-variability demand environments, like custom fabrication shops where each order is fundamentally different, struggle with pure JIT. Single-source suppliers for critical components create unacceptable risk. Geopolitical instability in your supply region, tariffs that change quarterly, shipping routes that get blocked unpredictably. These aren't weaknesses in JIT, they're boundary conditions. The methodology works within certain parameters and outside those parameters it creates fragility instead of efficiency. The tools themselves are straightforward to learn. What's harder is the cultural shift. JIT requires people to speak up when something is wrong instead of working around it. It requires suppliers to become partners instead of vendors you squeeze for price. It requires management to trust that reducing inventory won't collapse the operation, which means actually doing the improvement work first rather than just pulling stock and hoping. The order matters. Do the improvements, then remove the buffer. Remove the buffer first and you just have a broken system with less cushion to absorb the failures.