The fastest capacity you already own is your changeover

A plant manager asks for capital to add a new line. The business case is throughput: we're running flat out, we need more capacity. It's a reasonable request. It's also, very often, the most expensive way to solve the problem — because a large share of the capacity they're missing is already on the floor, sitting inside every changeover.

A changeover is the interval between the last good part of one run and the first good part of the next. During it, the line is staffed, energised, and producing nothing. Most plants treat that as a fixed cost of doing business. It is not fixed. It is one of the most improvable — and most ignored — assets in the building.

What a changeover actually costs

The numbers on structured changeover reduction have been stable for decades, because the method — SMED, Single-Minute Exchange of Die, developed by Shigeo Shingo — is well understood and repeatedly proven.

80%

Changeover-time reduction is routine with SMED — most implementations land between 50% and 80%

50%

Reduction is common from a single 1–3 day SMED event, before any tooling is even modified

4 hr → 3 min

Shingo's landmark result on a 1,000-ton stamping press — a ~99% reduction

< 10 min

The "single-minute" target: getting setup into single-digit minutes

Shingo's real insight — it isn't about turning wrenches faster

Everyone remembers the headline: "single-minute" changeovers. The genuinely useful insight is quieter. Most changeover time is not mechanical work at all — it is preparation. Hunting for tools. Waiting for the next die to arrive from storage. Finding the setup sheet. Making adjustments by trial and error until the first part passes.

Shingo's move was to split every changeover step into two kinds:

Internal setup — work that can only happen with the machine stopped.
External setup — work that can happen while the machine is still running the previous job.

Then: relentlessly convert internal into external, and streamline what's left. The wrench turns were never the bottleneck. The organisation around the wrench turns was.

See it happen

Here is a 70-minute changeover where every step is done with the line stopped. Press the button and watch what happens when the external steps move off-line — and the internal ones get streamlined.

Machine-stopped time

70 min

The changeover, step by step

Internal (machine stopped)

External (can run while producing)

Moved off-line — now done while the machine is still producing

Before: all 70 minutes happen with the line stopped — paid-for, zero-output time. Press Apply SMED.

The SMED sequence

1. Observe & separate	Film one real changeover. Sort every step into internal (stopped) or external (running). This alone usually reveals that a third or more of the "changeover" never needed the machine stopped.
2. Convert	Move internal work to external: pre-stage tooling and the next die at the machine, pre-heat, kit and stage all parts and paperwork before the line stops.
3. Streamline	Attack what genuinely must be internal: quick-release fasteners instead of bolts, standardised settings instead of trial-and-error adjustment, shadow boards so nothing is hunted for.
4. Standardise & sustain	Make the faster changeover the documented standard — not a one-off hero effort. The gain only compounds if every future changeover inherits it.

What it's worth on your floor

Changeover time recovered is production time gained — on equipment you already own, with no capital and no commissioning. Enter your own numbers.

Changeover ROI calculator

A rough estimate of the capacity hiding in your changeovers. Defaults assume a 60% reduction on a 70-minute changeover.

Changeovers per week

Minutes saved each

Contribution / hour

Capacity recovered / year

728 hrs

Value / year

$1.1M

A SMED event costs a few days of a team's time. A new line costs seven figures and a year. For a high-mix plant, changeover is almost always the cheapest capacity you can buy.

Is your changeover a SMED candidate?

A quick read. Nothing tracked, nothing saved.

Eight honest signals

Check what is true today.

A typical changeover takes longer than a coffee break.

The time varies a lot depending on who is running it.

Tooling and the next die are hunted for, not staged in advance.

Getting the first good part involves trial-and-error adjustment.

No one has ever filmed a changeover to study it.

The operators who do it aren't the ones asked how to improve it.

"That's just how long it takes" is the accepted answer.

Setup and changeover loss isn't measured as its own number.

What to watch next

First week

The camera does the convincing

The single highest-leverage step costs nothing: film a changeover and watch it back with the team. The waste that was invisible in real time becomes undeniable on playback.

First event

30–50% comes off before any capital

Separating internal from external and staging the external work usually halves the stopped time in a single structured event — with zero tooling spend.

3–6 months

The gain becomes structural

Quick-release tooling, standardised settings, and a documented standard push the reduction toward 60–90% and make it stick across shifts — so it compounds instead of decaying.

Before you write the capital request, film a changeover. There is a real chance the capacity you are about to spend a million dollars on is already sitting on your floor — in the gap between the last good part and the next one, waiting for someone to treat it as the asset it is.

Chasing capacity you might already own? Happy to compare notes.

Get in touch →

Where these numbers come from

Shigeo Shingo, A Revolution in Manufacturing: The SMED System — the internal/external principle and the 1,000-ton press case
Lean Production, SMED (Single-Minute Exchange of Die) — method and the six-step approach
Kaizen Institute, SMED: reduce changeover, boost efficiency — typical 50–80% reduction ranges
Hero image: site asset (wid-plant.jpg)