The Promise of the Level
Playing Field
Heijunka is the lean manufacturing practice of production leveling —
smoothing out the volume and mix of production so that every day, every
shift, every hour looks roughly the same. Instead of building 800 units
on Monday, 200 on Tuesday, and 950 on Wednesday, you build approximately
650 each day. Instead of running all Product A in the morning and all
Product B in the afternoon, you interleave them in small batches:
A-B-A-C-A-B-A-C.
The logic is elegant. When you level production, you level demand on
your suppliers, your equipment, your workforce, and your quality
systems. Everything becomes predictable. Inventory buffers shrink.
Changeovers happen in small, manageable increments. Problems surface
immediately because there is no surge volume to hide behind. The factory
achieves a steady heartbeat — and a steady heartbeat is what complex
systems need to stay healthy.
Toyota discovered this decades ago. The insight was simple but
profound: uneven production (mura) creates strain, strain creates waste
(muda), and waste creates overburden (muri). Level the flow, and you
eliminate all three at once. Heijunka became one of the foundational
pillars of the Toyota Production System, sitting alongside just-in-time
production and jidoka. It was not optional. It was the rhythm section of
the entire orchestra.
So why does almost nobody do it correctly?
What Actually
Happens: The Leveling That Wasn’t
Here is what typically happens when an organization adopts Heijunka.
The lean team creates a beautiful heijunka box — a grid showing
scheduled production quantities for each product across each time slot.
Color-coded cards slot into compartments. The visual is stunning.
Management tours the production line and nods approvingly. Photographs
are taken. The heijunka box appears in a presentation to the board.
Then Monday arrives.
A key customer calls with a rush order for Product A — 2,000 units by
Friday. The leveled schedule says you should be building 500 units of
Product A, B, C, and D in rotation this week. The customer doesn’t care
about your heijunka box. The sales VP doesn’t care about your heijunka
box. The plant manager, facing the choice between satisfying a
million-dollar account and maintaining production smoothing, makes the
only rational decision he can: he throws out the schedule.
By Wednesday, the line is running nothing but Product A. By Thursday,
the changeover to Product C that was supposed to happen Tuesday still
hasn’t happened. Operators who were supposed to do a 12-minute SMED
changeover four times this week haven’t done a single one. The materials
team ordered components based on the leveled schedule, so now there’s a
surplus of B and C parts and a shortage of A parts. Expedites go out.
Premium freight costs pile up. Quality suffers because the rushed
production means skipped checks.
Friday arrives. The rush order ships. The heijunka box is still on
the wall, still beautiful, still full of cards that bear no resemblance
to what actually happened on the shop floor this week. Nobody updates
it. Nobody mentions it. It has become what every abandoned lean tool
becomes: wall art.
The
Structural Problem: Why Leveling Fails in the Real World
The failure of Heijunka is not a failure of commitment. It’s not that
organizations don’t try hard enough. It’s that the conditions required
for true production leveling are extraordinarily difficult to achieve,
and most companies try to implement the tool before building the
foundation it rests on.
Condition 1: Demand Stability. Heijunka assumes that
demand is reasonably predictable. But most companies operate in
environments where demand swings by 30-50% month to month, where
customers change orders with no notice, where sales promotions create
artificial demand spikes, and where seasonality distorts every plan. You
cannot level production when the ground beneath you is shaking. Yet
companies try anyway, creating a leveled schedule on Monday that is
obsolete by Tuesday.
Condition 2: Short Changeover Times. To produce a
leveled mix of products, you need to switch between them frequently.
That requires changeovers measured in minutes, not hours. If your
changeover from Product A to Product B takes four hours, you cannot
economically interleave them in small batches. You’ll build a day’s
worth of A, then spend four hours changing over, then build a day’s
worth of B — which is batch production with a leveling label, not actual
leveling.
Condition 3: Supplier Capability. Heijunka doesn’t
just level your production — it levels the signals you send to your
supply chain. If you’ve been ordering 10,000 of Component X every two
weeks and suddenly start ordering 2,500 twice a week, your supplier
needs to be able to handle that. Most can’t. Most suppliers have set up
their own production, inventory, and logistics around your lumpy demand
pattern. When you try to smooth it, you create chaos in their system —
and they push back.
Condition 4: Workforce Flexibility. A leveled
schedule with frequent product changes requires multi-skilled operators
who can handle different products, different tools, and different
quality requirements in rapid succession. If your workforce is
specialized — John only runs Press 3, Maria only assembles Product C —
then leveling the mix means either idling people or retraining them.
Most companies choose neither. They keep running what the specialist is
good at, schedule be damned.
Condition 5: Management Discipline. This is the
hardest condition of all. Heijunka requires management to accept that
production will sometimes be lower than it could be, because the
schedule prioritizes smoothness over maximum output. When the monthly
P&L is tight and the CEO is demanding why Line 4 ran at 78% capacity
when it could have run at 92%, the plant manager folds. The schedule is
abandoned in the name of short-term throughput, and the leveling
dies.
The Pattern: How
the Tool Becomes the Theater
Here is how Heijunka typically degenerates in practice:
Month 1-3: Honeymoon. The lean team implements the
heijunka box. Production follows the schedule reasonably well because
everyone is watching. Metrics look good — output is smooth, inventory is
declining, changeovers are happening as planned. The lean team presents
results at the monthly review. Applause all around.
Month 4-6: Pressure. Real demand starts pushing
against the leveled schedule. A customer escalation happens. A
competitor launches a new product and the sales team wants to respond. A
supplier misses a delivery. Each disruption pushes the schedule further
off track. The heijunka box is still maintained, but the gap between
planned and actual is growing. Someone starts tracking “schedule
attainment” — and it’s at 63%.
Month 7-12: Erosion. The heijunka box is updated
less frequently. Operators start running whatever the expediters tell
them to run. The materials team reverts to building weekly batches
because that’s what suppliers can actually deliver. The lean team,
sensing the loss of momentum, launches a “Heijunka Revitalization”
initiative. It produces a new, improved heijunka box. The old one is
quietly removed from the wall.
Month 13+: Memory. Heijunka becomes something the
organization “tried.” It’s listed in the lean implementation log as
“implemented” with a completion date. Nobody mentions that it’s not
actually functioning. New hires are told that the company uses
production leveling. They look at the wall, see no heijunka box, and
learn not to ask.
The Hidden Cost:
Worse Than No System at All
The most damaging outcome of a failed Heijunka implementation is not
the wasted effort or the abandoned tools. It is the
cynicism it creates.
When operators watch a leveling system get implemented, celebrated,
and then abandoned — all within the span of a year — they learn
something. They learn that management’s flavor-of-the-month initiatives
are not to be taken seriously. They learn that the next lean tool rolled
out will follow the same pattern. They learn to wait it out.
This learned indifference is more costly than any production
imbalance. When you later try to implement SMED, or standardized work,
or visual management, you’re not fighting process resistance. You’re
fighting organizational memory. The operators have seen this movie
before, and they know how it ends.
There is also a direct operational cost. A partially functioning
Heijunka system can be worse than no system at all, because it creates
the illusion of control. The heijunka box on the wall says you’re
building a balanced mix. The actual production data says you’re building
whatever the loudest customer demanded. When a quality problem emerges —
say, a spike in defects on Product B — the investigation starts from the
assumption that production was running as scheduled. But it wasn’t. The
actual production pattern was chaotic, and the root cause is buried in
the gap between plan and reality.
When Heijunka Actually Works
It’s worth understanding the rare cases where Heijunka does work,
because they illuminate what’s required.
Toyota, obviously. But Toyota spent decades building
the foundation: a supplier network designed for frequent deliveries, a
workforce trained for flexibility, equipment engineered for rapid
changeover, and — critically — a sales and distribution system that
smoothed demand at the dealer level before it reached the factory.
Heijunka at Toyota is not a scheduling tool bolted onto a conventional
operation. It is an organizing principle that shapes every element of
the value chain.
Companies with mature pull systems. Organizations
that have already implemented effective kanban systems, achieved
reliable changeovers, and stabilized their supply chains can layer
Heijunka on top successfully. In these environments, production leveling
is the capstone, not the foundation.
Make-to-stock environments with managed product
portfolios. Companies that produce a bounded set of standard
products, hold reasonable finished-goods inventory, and have the
organizational authority to manage customer expectations (rather than
being managed by them) can make Heijunka work. The key is that demand is
shaped, not merely accepted.
Long-cycle, low-mix manufacturing. In some process
industries — chemicals, pharmaceuticals, steel — production leveling is
a natural fit because the economics of batch production already enforce
a degree of rhythm. Heijunka formalizes what these operations tend to do
anyway.
The common thread: Heijunka works when the entire system is designed
to support it. It fails when it’s treated as a standalone scheduling
technique that can be layered onto an otherwise conventional
operation.
What to Do Instead:
Pragmatic Alternatives
If your organization is not ready for full Heijunka — and most aren’t
— here are pragmatic alternatives that deliver some of the benefits
without requiring the full Toyota Production System as a
prerequisite.
1. Demand Segmentation, Not Smoothing
Instead of trying to level all production, segment your demand into
three buckets: stable runner products (high volume, predictable),
repeater products (medium volume, periodic), and stranger products (low
volume, sporadic). Level the runners. Batch the repeaters on a fixed
cycle. Build the strangers to order. This gives you leveling where it’s
feasible and responsiveness where it’s necessary.
2. Pitch and Pace, Not Mix Leveling
If you can’t level the product mix, at least level the pace.
Establish a takt time for each product family and run at that pace
within each family. Even if you’re building Product A all day Monday and
Product B all day Tuesday, running each at a steady pace within its
block delivers most of the workforce stability and quality benefits of
full Heijunka.
3. Freeze Windows
Rather than trying to create a perfectly leveled perpetual schedule,
establish a freeze window — say, two weeks — during which the production
plan cannot be changed. This gives the shop floor the stability it needs
to execute, even if the plan within the window isn’t perfectly leveled.
The chaos gets pushed to the edges of the window, where it can be
managed during planning rather than during execution.
4. Make the Trade-off Explicit
The most important thing you can do is make the trade-off visible.
When a customer escalation disrupts the leveled schedule, quantify the
cost: the expedited materials, the overtime, the quality risk, the
schedule disruption to other products. Present this cost alongside the
benefit of accommodating the escalation. Let management make an informed
decision rather than a reflexive one. You may still choose to disrupt
the schedule — but at least the choice is conscious.
The
Deeper Lesson: Tools Without Foundations Are Decorations
Heijunka’s failure pattern is not unique. It’s the same pattern that
plays out with kanban systems that become push systems, with SMED
programs that never reduce changeover times, with 5S implementations
that devolve into periodic cleaning campaigns, with standardized work
documents that nobody follows.
The pattern is always the same: the tool is adopted without the
foundation. The tool is celebrated without the discipline. The tool is
abandoned without the lesson.
The lesson is this: lean tools are not independent. They are an
interdependent system. Heijunka requires SMED. SMED requires
standardized work. Standardized work requires visual management. Visual
management requires a problem-solving culture. A problem-solving culture
requires management that values learning over blame. You cannot pick the
tools you like and ignore the ones that are hard. The tools you skip
become the weak links that bring down the ones you installed.
Production leveling is one of the hardest lean tools to implement
because it requires the most organizational maturity. It asks the
factory to produce less than it could, to hold more inventory than it
needs, to invest in changeovers that seem wasteful — all for benefits
that are systemic and long-term rather than immediate and local. That is
a difficult ask in any organization. In most, it is impossible.
So the next time you see a beautiful heijunka box on a factory wall,
look at the shop floor. Look at the production logs. Look at the
expedite list. Look at the supplier delivery performance. If those tell
a different story than the box on the wall, you are looking at yet
another production leveling implementation that became a scheduling
straightjacket nobody followed — another set of peaks that were supposed
to be smoothed and instead became the chaos that everyone learned to
manage.
The tools are not wrong. The foundations are missing. And until the
foundations are built, the tools will continue to be decorations —
expensive, time-consuming, cynicism-generating decorations that make the
organization worse, not better.
About the Author
Peter Stasko is a Quality Architect with over 25 years of experience
in manufacturing quality management, lean implementation, and production
system design. He has led quality transformations across automotive,
electronics, and industrial equipment sectors, specializing in bridging
the gap between lean theory and shop-floor reality. His work focuses on
building practical, sustainable quality systems that survive contact
with real-world production pressures.
