The pacemaker process is the single point in a value stream that you schedule, and it sets the pace for the whole stream. You send the production schedule to just this one process; everything downstream flows to the customer without its own schedule, and everything upstream is triggered by pull.

Most plants schedule everything. Every process gets its own work order, its own priorities, its own idea of what to build next, and the result is a value stream pulling in a dozen directions with inventory piling up wherever two schedules disagree. The pacemaker concept from value stream mapping fixes that with a discipline that feels almost too simple: schedule one process, and only one. The Lean Enterprise Institute defines the pacemaker as "any process along a value stream that sets the pace for the entire stream" (Lean Enterprise Institute, Pacemaker Process). Get the pacemaker right and the rest of the stream marches to one drumbeat instead of arguing.

What is a pacemaker process?

A pacemaker process is the one process in a value stream that receives the customer schedule and sets the rhythm for everything else. It is the drummer the whole line keeps time to. You schedule the pacemaker to takt time the rate of customer demand, and because every other process is linked to it by flow or pull, scheduling that single point paces the entire stream.

The rule that makes it work has two halves. Downstream of the pacemaker, everything must flow to the customer, no supermarkets, no separate schedules, just continuous flow or a first-in-first-out lane. Upstream of the pacemaker, everything runs on pull triggered by what the pacemaker consumes, usually through supermarkets and kanban. So the pacemaker is the boundary: schedule flows into it, product flows out of it, and pull ripples back from it. That is why you only need to schedule one place.

Schedule only the pacemaker: pull upstream, flow downstreamSchedule one point; the stream keeps time to itSUPPLYWELDASSEMBLYpacemakerTESTSHIPSCHEDULEPULL (supermarkets)FIFO FLOW to customerUpstream: pull back from what the pacemaker uses. Downstream: flow forward, no separate schedule.
Only the pacemaker gets a schedule. Upstream processes are pulled by what it consumes; downstream processes flow to the customer in FIFO. One scheduling point paces the whole stream.

Why schedule only one point?

Because scheduling every process guarantees they disagree. Each process working to its own plan optimizes itself, not the flow, and the mismatch between neighbors shows up as inventory in one direction and starvation in the other. That is textbook overproduction: an upstream process building to its own schedule floods the next step with work it cannot use yet.

Scheduling a single pacemaker removes the argument. There is one instruction in the stream, and every other process either flows from it or pulls to it, so there is nothing to fall out of sync. It also makes problems visible fast. When only one point is scheduled and the rest flows, a fall-behind anywhere shows up immediately as the flow stalls, instead of being masked by a buffer somebody built ahead. And it makes the plant easier to run: schedulers manage one point instead of ten, and the floor has one clear pace to hold. This is the same logic as heijunka level the demand signal at one point and let the stream follow, rather than trying to level ten places at once.

How is a pacemaker different from a bottleneck?

This is the distinction people get wrong most often, and LEI calls it out directly: the pacemaker "should not be confused with a bottleneck process, which necessarily constrains downstream processes due to a lack of capacity" (LEI, Pacemaker Process). A bottleneck is a physical constraint, the slowest step, the one without enough capacity to keep up. A pacemaker is a scheduling choice, the point you decide to instruct. They are answers to different questions: the bottleneck is "where is the stream physically limited," the pacemaker is "where do we send the schedule."

They are not always the same place, and confusing them leads to bad decisions. You might place the pacemaker at final assembly for good flow reasons even though the bottleneck is an upstream oven. In that case the oven still governs how much the stream can produce, and you manage its capacity separately, while the pacemaker governs the rhythm and mix. If your bottleneck analysis and your pacemaker happen to land on the same process, fine, but do not assume it. Choose the pacemaker for flow, and manage the constraint for capacity.

Pacemaker versus bottleneck: two different questionsPacemaker and bottleneck answer different questionsPACEMAKERa scheduling choice"where do we sendthe schedule?"sets rhythm + mixBOTTLENECKa physical constraint"where is capacitythe limit?"caps total outputThey can be the same process, but do not assume it: choose one for flow, manage the other for capacity
The pacemaker is where you send the schedule; the bottleneck is where capacity runs out. They may coincide, but choosing the pacemaker for flow and managing the constraint for capacity are separate jobs.

Where should the pacemaker sit in the value stream?

Usually near the customer end. LEI notes the pacemaker "usually is near the customer end of the value stream, often the final assembly cell." The reason is the flow rule: everything downstream of the pacemaker has to move to the customer in continuous flow or FIFO with no supermarket, so you want as little of the stream downstream of the pacemaker as possible. Put the pacemaker late and you have only a short, simple flow to manage after it.

There is an important exception. LEI adds that "if products flow from an upstream process to the end of the stream in a FIFO sequence, the pacemaker may be at this upstream process." In other words, if everything from some upstream point onward already flows in strict FIFO to shipping, that upstream point can be the pacemaker, because the flow rule is already satisfied downstream of it. The test is always the same: can everything downstream of your candidate pacemaker flow to the customer without its own schedule? If yes, it can be the pacemaker. If no, move the pacemaker downstream until it can.

How do you set up a pacemaker process? A sequence

  1. Map the value stream first. You cannot pick a pacemaker without seeing the whole stream. Draw the value stream map so every process, inventory, and information flow is on one page.
  2. Find where continuous flow ends. Identify the most downstream point from which product can flow to the customer without a separate schedule. That point, or just downstream of it, is your pacemaker candidate.
  3. Confirm the downstream flow rule. Check that everything after the candidate can run as continuous flow or a FIFO lane to shipping. If a supermarket or separate schedule is unavoidable downstream, move the pacemaker later.
  4. Set the pacemaker to takt. Schedule it to takt time so its rhythm matches customer demand, and level the mix at this point with a heijunka box rather than releasing in big batches.
  5. Put the upstream on pull. Connect every upstream process to the pacemaker with supermarkets and kanban so they build only what the pacemaker withdraws, never to their own forecast.
  6. Release and check at a fixed pitch. Use paced withdrawal to release work to the pacemaker and pull finished goods in small, frequent increments, so a fall-behind shows up in minutes.

Run this and you have converted a stream of arguing schedules into one instructed point with flow after it and pull before it. The plant now has a single, honest heartbeat.

How do pitch and paced withdrawal keep the pacemaker honest?

Setting the pacemaker to takt tells it the average rate, but average rate is easy to fake by racing ahead early and coasting late. Pitch and paced withdrawal are what make the pace real. Pitch is a small, consistent increment of time, often takt time multiplied by the pack-out quantity, and paced withdrawal means you release work and pull finished goods from the pacemaker once every pitch, in that increment, all shift. Because you check the pacemaker against plan every pitch, a fall-behind is caught within one pitch interval rather than at end of shift. The pacemaker keeps the rhythm; pitch keeps the pacemaker honest.

FactWhat LEI statesPrimary source
Definition"Any process along a value stream that sets the pace for the entire stream."LEI, Pacemaker Process
Not a bottleneck"Should not be confused with a bottleneck process, which necessarily constrains downstream processes due to a lack of capacity."LEI, Pacemaker Process
Location"Usually near the customer end of the value stream, often the final assembly cell."LEI, Pacemaker Process
Upstream exceptionIf products flow to the end in FIFO, the pacemaker may sit at that upstream process.LEI, Pacemaker Process
The pacemaker is defined by scheduling and flow, not by capacity. That is why LEI is careful to separate it from the bottleneck.

How does live floor data help you run the pacemaker?

A pacemaker only paces the stream if you can see it fall behind. If the pacemaker's actual output against plan is reconstructed at end of shift from a paper count, the whole point, catching a fall-behind fast, is lost. Plants that capture the pacemaker's output live, pitch by pitch, can see the gap open the moment it happens and react while the shift is still running. That is the practical value of a live factory visibility layer over your existing lines, no rip-and-replace: the pacemaker's heartbeat becomes a signal anyone can watch instead of a number someone tallies later. It also keeps the upstream pull honest and stops a batch production habit from creeping back in as build-ahead. See how digitizing the floor first plays out in the CLS case study. Pick one point to schedule, make everything flow to it or pull from it, and put a live signal on its pace. That is the pacemaker working as designed.