The Promise That Sold a
Million Books
Eliyahu Goldratt published The Goal in 1984, and it changed
how manufacturing thought about production flow. The core idea was
almost insultingly simple: every system has exactly one constraint at
any given time, and if you want to improve the system’s output, you must
find and exploit that constraint. Everything else — every machine, every
worker, every process step that isn’t the bottleneck — has spare
capacity by definition. Optimizing a non-bottleneck doesn’t increase
throughput. It just builds more inventory somewhere else.
For a brief, shining moment, manufacturers understood this. They
mapped their flows. They found their constraints. They focused their
improvement energy where it mattered. And then, slowly, the insight
dissolved into the same corporate mush that swallows every good idea.
Theory of Constraints became a training module. The Five Focusing Steps
became a slide. The constraint became a line item in a quarterly
review.
And the actual bottleneck — the real, physical, capacity-limiting
step in the production line — sat exactly where it always had, quietly
governing the output of the entire plant while nobody touched it.
The Five
Focusing Steps (And How We Sidestep Every One)
Goldratt gave us a clear, sequential methodology. Let’s walk through
each step and examine what actually happens in most manufacturing
organizations.
Step 1: Identify the
Constraint
What it means: Find the single resource, station, or
process step that limits the system’s throughput. The constraint is
where work piles up, where cycle time equals the longest in the line,
where demand exceeds available capacity.
What actually happens: Nobody wants to admit which
station is the bottleneck. The constraint is politically loaded. If your
department owns the constraint, you’re the problem. So instead of honest
identification, you get spreadsheets that average out utilization across
all stations, hiding the bottleneck in a fog of aggregate metrics. Or
you get “we have multiple constraints” — which is usually wrong. A
system with genuinely active constraints at multiple points is rare and
unstable; most of the time, there’s one dominant bottleneck, and
resolving it reveals the next one downstream. But saying “we have
several constraints” lets everyone off the hook.
The symptom: You cannot answer the question “what is
the slowest station on your line?” in one sentence without qualifying it
with fifteen caveats.
Step 2: Exploit the
Constraint
What it means: Get every possible unit of output
from the constraint before investing any money. Schedule it to never be
idle. Eliminate breaks that coincide with shift changes. Make sure it
never waits for material, never waits for a tool, never waits for a
setup that could have been done externally. The constraint’s minute is
the system’s minute — a minute lost at the constraint is a minute lost
forever, with no recovery possible.
What actually happens: The constraint is treated
like every other station. Operators take breaks whenever convenient. The
bottleneck machine sits idle for twenty minutes while someone finds a
forklift to bring the next pallet of raw material. Setup happens on the
clock, line stopped, because nobody prepared the tools in advance.
Maintenance is scheduled “when we have time” rather than during
non-production windows. The constraint runs at sixty to seventy percent
of its potential, and nobody notices because nobody is measuring the
constraint’s active cutting/processing time separately from its
available time.
The symptom: Nobody has ever calculated the cost of
one hour lost at the constraint. If they had, the math would be
terrifying — it’s not the machine’s hourly rate, it’s the entire line’s
throughput value per hour.
Step 3: Subordinate
Everything Else
What it means: Every non-constraint resource must
serve the constraint. Upstream stations should not produce faster than
the constraint can consume — excess production just builds inventory.
Downstream stations should be ready to process whatever comes off the
constraint immediately. The entire rhythm of the factory should be set
by the constraint, the way a drummer sets the pace for a rowing
crew.
What actually happens: Each department optimizes
locally. Upstream stations run at maximum efficiency because their KPIs
reward utilization, not synchronization. They build mountains of
work-in-process in front of the constraint and call it “being
productive.” Downstream stations complain about the constraint’s
inconsistent output and lobby for their own buffer inventory. The
factory floor becomes a warehouse for parts that are waiting — waiting
for the one station that can’t keep up, while every other station
proudly reports ninety-five percent utilization.
This is the step most organizations fail at, and it’s the most
damaging failure, because it inverts the entire logic of TOC. Local
optimization at non-constraint resources doesn’t just fail to help — it
actively hurts. Every unit produced above the constraint’s capacity is
waste. It consumed raw materials, labor, energy, and machine time to
become inventory that will sit, degrade, and eventually become obsolete.
You paid to create a problem.
The symptom: You have more inventory than you
should, and the inventory is concentrated in specific locations —
upstream of a station that can’t process it fast enough.
Step 4: Elevate the
Constraint
What it means: If you’ve exploited and subordinated
and the constraint is still limiting throughput, spend money. Add a
second machine. Add a shift. Outsource the bottleneck operation. Upgrade
the technology. This is where capital investment happens — but only
after you’ve squeezed every bit of free capacity through steps 2 and
3.
What actually happens: This step gets skipped to
immediately. Manufacturers love buying equipment. A new CNC machine, a
faster line, a bigger oven — capital expenditure is the default answer
to every throughput problem because it requires no thinking, no process
change, no organizational discipline. The constraint gets “solved” by
throwing money at it without first exploiting it or subordinating to it,
which means the new equipment is oversized, underutilized from day one,
and the real bottleneck has already migrated somewhere else.
Or the opposite: the constraint never gets elevated because capital
approval requires an ROI calculation that nobody can produce, because
nobody has been measuring the constraint’s performance, because Step 1
was never properly completed.
The symptom: You bought a million-dollar machine and
throughput didn’t improve. Or you need a machine and can’t justify it
because you have no data.
Step
5: Repeat — Don’t Allow Inertia to Become the Constraint
What it means: Once you’ve elevated the current
constraint, it’s no longer the constraint. The bottleneck has moved —
somewhere else in the line, maybe upstream, maybe downstream, maybe in
the supply chain or the engineering department or the sales pipeline. Go
back to Step 1 and find the new constraint. Do not become
complacent.
What actually happens: The organization declares
victory. “We fixed the bottleneck.” A banner goes up. A case study gets
written for the next industry conference. And the new constraint — the
one that was revealed the moment the old one was resolved — sits
unaddressed for the next three years. The factory’s throughput plateaus
immediately after the improvement and nobody understands why, because
they’re still monitoring the old constraint’s metrics, which look great,
while the new constraint quietly limits everything.
The symptom: You improved a station’s output by
forty percent and total throughput went up two percent.
The Drum-Buffer-Rope
Illusion
Goldratt’s operational methodology for TOC is called
Drum-Buffer-Rope, and it’s elegant. The constraint is the drum — it sets
the pace for the entire line. A buffer of inventory is maintained in
front of the constraint to protect it from disruption — not a massive
buffer, just enough to absorb variability in upstream supply. The rope
is the signaling mechanism that releases raw material into the line at
the pace of the drum — no faster — preventing overproduction.
In practice, Drum-Buffer-Rope becomes
“Drum-Buffer-Buffer-Buffer-Buffer.” The buffer management discipline
collapses because nobody can agree on how much buffer is enough.
Operators add safety stock “just in case.” Supervisors add more because
they don’t trust the upstream process. Soon there’s a week’s worth of
inventory in front of a constraint that could be protected by four hours
of buffer, and nobody can find anything in the pile.
The rope — the release mechanism — is the first thing to go. Material
handlers release work orders based on what’s available, not based on the
constraint’s signal. The drum beats at its own pace, but nobody is
listening. Upstream stations churn out parts. The buffer grows.
Work-in-process clogs the aisles. And the constraint, which should be
the most pampered resource in the building, waits for a specific
component that’s buried under six pallets of parts that were produced
too early.
Throughput
Accounting: The Metric Nobody Uses
Goldratt proposed Throughput Accounting as an alternative to
traditional cost accounting, and it’s one of his most misunderstood
contributions. The core idea: measure three things — Throughput (the
rate at which the system generates money through sales), Inventory (all
the money invested in things the system intends to sell), and Operating
Expense (all the money spent to turn Inventory into Throughput).
The beauty of this framework is its clarity. Any decision can be
evaluated with a simple question: does it increase Throughput, decrease
Inventory, or decrease Operating Expense? If the answer is yes to any of
these without a corresponding negative elsewhere, the decision is good.
Local efficiency doesn’t matter. Machine utilization doesn’t matter.
Standard costing doesn’t matter. What matters is: are we making money,
and are we making it faster?
Traditional cost accounting, by contrast, rewards overproduction. If
a machine has a standard cost of $200 per hour and you run it at full
capacity making parts that nobody needs, the accounting system shows
“efficiency” at 100%. The parts go into inventory, where they sit, tying
up capital and floor space. But the variance report looks great. The
department manager gets a bonus. And the constraint — the one resource
that actually determined how much money the plant made — sits at 70%
utilization, starved of material, waiting for the right components that
were pushed to the back of the queue by parts produced for efficiency’s
sake.
Almost no manufacturer uses Throughput Accounting. The CFO has never
heard of it. The ERP system can’t produce it. The accounting department
would need to restructure its entire reporting methodology. So the
insight dies on contact with the spreadsheet, and the factory continues
to optimize for a number that has nothing to do with making money.
What Real TOC
Implementation Looks Like
A genuine Theory of Constraints implementation is uncomfortable. It
requires you to stop optimizing most of your factory and concentrate all
your attention on one point. Here’s what it actually requires:
Single-thread focus. The entire continuous
improvement program — the Six Sigma black belts, the kaizen events, the
5S audits — gets redirected toward the constraint. If your top engineer
is working on a non-constraint station, that engineer is wasting time.
Not because the work isn’t valuable in isolation, but because improving
a non-constraint resource does not increase system throughput. It is
mathematically irrelevant.
Brutal honesty about capacity. You have to publish
the constraint’s performance data — actual active processing time versus
available time — and the number will be ugly. Sixty percent, maybe
seventy. That gap is your opportunity, but it’s also an admission that
you’ve been wasting thirty to forty percent of your plant’s most
critical resource. Managers who built their careers on “running a tight
ship” do not want this number on a dashboard.
Buffer management as a discipline. Someone — a real
human being, not a software system — must own the constraint’s buffer.
That person checks it hourly, tracks what disrupts it, and escalates
threats. When a buffer level drops below a trigger point, it’s treated
like a quality defect: investigated, root-caused, corrected. This is not
glamorous work. It is the most important operational job in the
plant.
A release discipline that overrides everything.
Material is released to the line only at the pace the constraint can
consume it. This means upstream stations will be idle sometimes.
Operators will stand around. Machine utilization at non-bottleneck
stations will drop. This is correct. This is how the system is supposed
to work. And it is the single hardest thing to explain to a plant
manager who grew up being told that idle time is failure.
Continuous re-identification. Every time the
constraint moves — and it will move, the moment you successfully exploit
and elevate the current one — the entire organization must pivot. The
old constraint becomes a regular resource. The improvement team packs up
and moves to the new bottleneck. The buffer relocates. The release point
shifts. This agility is what separates a genuine TOC implementation from
a one-time optimization project.
The Real Reason TOC Fails
Theory of Constraints fails for the same reason every management
methodology fails: it requires you to do one thing and stop doing
seventeen other things.
Most organizations cannot tolerate this. They want simultaneous
improvement everywhere. They want every department to have a project,
every station to have a metric, every shift to have an initiative. The
idea that ninety percent of the factory should be left alone while all
resources concentrate on one bottleneck is anathema to the way most
managers think about their jobs.
But TOC isn’t a suggestion. It’s a mathematical reality. The system’s
output is determined by its constraint. You can either work with that
reality or fight against it. Working with it means focus, discipline,
and the willingness to have most of your factory appear “underutilized”
by traditional metrics. Fighting against it means local optimization
everywhere, inventory piling up in all the wrong places, and throughput
that never improves no matter how many improvement projects you
launch.
Goldratt understood something that most managers still don’t: the
goal of a manufacturing system is not to be busy. The goal is to make
money — now and in the future. Every activity that doesn’t contribute to
that goal, no matter how efficient it looks, is waste.
The question isn’t whether your factory has a constraint. It does.
The question is whether you know where it is, whether you’ve done
everything possible to exploit it, and whether you have the discipline
to subordinate everything else to it.
For most manufacturers, the honest answer to all three is no.
About the Author: Peter Stasko is a Quality
Architect with over 25 years of experience in manufacturing quality
management, process optimization, and continuous improvement. He has
implemented lean and TOC principles across automotive, electronics, and
heavy industry sectors, helping organizations identify and break through
the constraints that limit their performance.
