Flow & Pull Systems

Push System

Make to a schedule. Hope the next station can absorb it. Usually it cannot.

What is a Push System?

A push system is a production approach where each operation produces to a forecast, schedule, or material requirements plan, regardless of whether the next operation is ready to consume it. Material is pushed forward through the value stream on the timing of the schedule, not on the rhythm of downstream demand. Push is the default mode of most non-lean shops and the cause of most WIP problems.

Push systems are the default in most manufacturing shops, which is the main reason their lead times are long, their WIP is high, and their schedule changes constantly. Push is not a deliberate choice. It is what happens when nobody installs a pull mechanism. MRP, ERP, and traditional scheduling all push by default. Recognizing push as a system, with its own predictable failure modes, is the first step toward replacing it with something better.

"Push is what you have when nobody installed anything else. The default is not the same as the right answer."

How a push system works

The mechanism starts with a forecast or order book. That forecast gets exploded backward through the bill of materials and the routing for each part, producing a schedule that tells each operation what to make and when. The schedule goes to the floor as printouts, ERP screens, or work-order packets. Each operation runs its slice of the schedule independently. The turner makes turning batches. The mill makes milling batches. The assembly station makes assembly batches. Each operation pushes its output forward to the next station, regardless of whether that station is ready.

The system is locally efficient and globally inefficient. Each station looks productive because it is busy. The schedule has been calibrated to keep utilization high. But because no station gets a signal from downstream about whether to slow down or stop, mismatches between station speeds turn into WIP piles. The faster station fills the buffer; the slower station drowns. The expediter then spends most of the day walking the floor, finding bottlenecks, and rearranging priorities.

The other failure mode is the schedule itself. Push is only as good as the forecast it runs on, and forecasts are wrong. Demand shifts, orders change, customers cancel. When the schedule has to be rerun, every operation gets a new printout and the floor reorganizes around the new priorities. The reorganization itself takes time, generates errors, and creates more expedites. In a busy push shop, the schedule gets rerun weekly, sometimes daily. The floor never settles into a rhythm because the rhythm keeps changing.

Push has fundamental limits. Tighter scheduling, better forecasts, and more aggressive expediting all hit diminishing returns. The leverage in lean is to stop pushing and start pulling: replace the schedule with signals from downstream consumption, and let each operation produce only when the next operation has consumed.

Where a push system fits on the shop floor

Picture a 40-person fab shop running steel parts for industrial OEMs. Production is driven entirely by MRP, which generates work orders weekly. The shop runs about $400,000 in WIP between operations, has a 30-day lead time to customers, and uses an expediter full-time to chase hot orders. The owner is convinced the answer is a better ERP. The system they have is six years old and slow.

A push diagnosis would name the problem differently. The ERP is not the issue; the model is. The shop runs push because it has always run push. A new ERP that runs MRP faster will still run MRP, which still pushes. The leverage move is to convert the highest-volume parts (the runners in runner-repeater-stranger classification) to a pull mechanism with kanban signals and supermarkets, and leave low-volume custom parts on push. The mixed approach is common in SMB shops: pull for the steady work, push for the one-offs. The conversion is cheaper than the new ERP and addresses the actual flow problem, not just the dashboard layer above it.

Common mistakes with push systems

Not naming it. Most shops do not recognize they are running push. Naming the system is the prerequisite for evaluating alternatives.
Trying to fix push with more push. Better forecasts, tighter schedules, more expediting all hit diminishing returns. The mechanism itself is the limit.
Hybrid push and pull on the same parts. A shop that keeps the push schedule alongside pull signals usually reverts to push under pressure. Pick one signal per part.
Blaming the ERP. Push lives in the model, not the software. A faster ERP runs push faster, which is not what the shop needs.
Treating all parts the same. Push and pull can coexist if they cover different part categories. High-volume runners should pull; one-off custom work can stay on push or sequenced pull.

Push system and related Lean tools

A push system is the opposite of a pull system and is the underlying model behind most batch-and-queue production. It is what just-in-time was designed to replace: in JIT, downstream consumption triggers upstream supply through kanban signals, while push produces to a schedule regardless of consumption. Most lean conversions start by identifying which parts to leave on push (low-volume, custom, irregular) and which parts to convert to pull (steady, repeated, high-volume).