Lean manufacturing examples from shop floor to office

Lean is easier to understand through examples than definitions. This guide pairs common problems with the lean tool that addresses them.

The shop-floor examples are typical situations, not case studies of named companies. Their numbers are illustrative, chosen to show the size of change each tool aims at, not results from a real plant. The examples outside manufacturing come from published sources.

Key takeaways

Each lean tool answers a specific kind of waste, such as waiting, searching or rework.

The tools work in combination, because they support each other.

The same ideas apply outside factories, in hospitals, offices and software teams.

Toyota: where the examples start

The Toyota Production System is the original lean example. Toyota describes it as resting on just-in-time and jidoka.

Just-in-time means each process makes only what the next process needs, when it needs it. Jidoka means machines and people stop when something goes wrong, so problems are fixed at the source.

Most of the examples below are one of these two ideas, applied to a specific problem.

Long changeovers: SMED

Problem: a press takes a long time to change from one part to the next. So the plant runs large batches and holds a lot of stock.

Lean fix: the team films a changeover. It separates internal steps, done while the press is stopped, from external steps done while it runs. Tools and dies are staged in advance.

Result (illustrative): the changeover drops from 90 minutes to 25, because 50 minutes of the old changeover went on fetching dies, tools and paperwork that can be staged while the press runs. With 2 hours a shift set aside for changeovers, the press can now change over four times a shift instead of once, so batches can shrink to about a quarter of their old size. The SMED page walks through the method.

Stockouts and overstock: kanban

Problem: a line runs out of some parts while others pile up, because orders follow a forecast rather than actual use.

Lean fix: a two-bin kanban system. When the first bin empties, its card triggers a refill while the line works from the second bin.

Result (illustrative): each bin is sized to cover demand during the refill time. If the line uses 40 brackets an hour and a refill takes 3 hours, each bin holds 120 brackets plus a 10% buffer, so about 264 brackets in two bins replace a forecast-driven pile of 1,500. Stock is replaced as it is used. E-kanban software does the same with barcode scans when there are many parts or suppliers.

Forecast-driven stock is a large cost across the economy. US businesses held $2,818.8 billion of inventory at the end of August 2026, about 1.33 months of sales, according to the Census Bureau.

Searching for tools: 5S

Problem: operators spend time looking for tools, gauges and paperwork.

Lean fix: a 5S event sorts out what is not needed and gives every remaining item a marked place, such as a shadow board.

Result (illustrative): a week of timed searches before the event logs about 20 minutes per operator per shift. A repeat count a month after the event shows about 3. Missing items are obvious at a glance. The 5S and kaizen guide explains how to keep it that way.

Repeat defects: poka-yoke and andon

Problem: the same assembly error keeps reaching final inspection.

Lean fix: an operator pulls the andon when a part will not fit. The team finds the cause and adds a poka-yoke fixture that only accepts the part the right way round.

Result (illustrative): reversed parts found at final inspection fall from about 12 a week to none, because the fixture will not accept a reversed part. Andon calls at that station rise for a week as operators flag parts that don't fit, then fall once the supplier removes a burr. The error is prevented at the station rather than caught later. Andon software can record each call and the response time.

Unbalanced line: takt time

Problem: some stations wait while others are overloaded, and output falls short of demand.

Lean fix: the team calculates takt time and rebalances work so each station's cycle time sits just under it. The takt time calculator helps with the numbers. Say a shift has 450 available minutes and demand is 300 units, so takt is 90 seconds. Four stations run at 70, 105, 80 and 85 seconds. Moving 20 seconds of work from the 105-second station to the 70-second one gives 90, 85, 80 and 85, all within takt.

Result (illustrative): output rises from about 257 units a shift, set by the 105-second station, to the 300 the customer needs. Work flows at the pace of demand, and the next bottleneck becomes visible.

One line, several tools

An illustrative example shows how the tools combine. A bracket line misses its target of 300 units on most days.

A value stream map shows 4 days of stock between stamping and assembly, and an hourly tally shows the stamping press down for changeovers about 2 hours a shift.

The team runs SMED first, because smaller batches are the condition for everything else. Then it replaces the stock pile with a kanban loop sized to the new batch and rebalances assembly to takt. 5S at the press supports the faster changeover, because every die now has a marked place.

The order matters. Kanban before SMED would starve the line, because the press could not yet change over often enough to refill small bins.

Lean in healthcare

Virginia Mason in Seattle began applying the Toyota Production System to health care in 2002. Its system includes the idea that staff can stop the process when they see a safety problem.

That is jidoka, applied to patient safety instead of car parts.

Lean in software and startups

Software teams use kanban boards with work-in-progress limits to keep work flowing. The guide on why limiting WIP makes teams faster explains the effect, and flow metrics tools measure it.

Startups apply lean to product development with build-measure-learn, from Eric Ries's The Lean Startup. The smallest useful test replaces a full product launch.

How to judge a lean example

Published lean examples, including vendor case studies, are easy to find and hard to compare. Four questions help.

Is there a baseline, measured the same way before and after? Did the result last, checked months later rather than at the end of the event? What else changed at the same time, such as demand, staffing or a new machine? And what got worse, such as extra changeovers adding setup scrap?

An example that answers all four is worth more than a bigger number that answers none.

Finding your own example

Walk your process and look for the eight wastes, from waiting to unused talent. The eight wastes page lists them.

Pick the one that costs the customer most and match it to a tool above. The lean manufacturing guide shows how the tools fit together.

Small plants can get outside help with this step. Clients of the Manufacturing Extension Partnership, a NIST program that supports small and medium-sized manufacturers, reported $15 billion in new and retained sales and $2.6 billion in cost savings in fiscal year 2024.

Frequently asked questions