Continuous improvement in lean: examples and strategies

Continuous improvement is the part of lean that never ends. The fifth lean principle, perfection, says there is always more waste to remove.

This guide explains how lean teams improve continuously: the PDCA cycle, daily kaizen, kaizen events and the habits that keep them going.

Key takeaways

Continuous improvement in lean runs on the PDCA cycle: plan, do, check, act.

Daily kaizen handles small problems, and kaizen events handle focused, larger ones.

Standard work is the baseline. Without it, there is nothing to improve on.

Improvement lasts only when leaders review it, coach it and make time for it.

What is continuous improvement in lean?

Continuous improvement is the ongoing effort to make processes better through many small, tested changes. In lean it is often called kaizen.

The continuous improvement method page covers the definition, principles and process. This guide focuses on how lean teams practice it.

Masaaki Imai popularized the term in his 1986 book Kaizen: The Key to Japan's Competitive Success. It describes improvement by everyone, every day.

It matches the last of the five lean principles in Womack and Jones's Lean Thinking: perfection. The lean methodology guide covers all five.

Start with the problem, not the fix

Many improvement efforts begin with a favorite answer: new software, more staff, a new layout or a new rule. Lean asks the team to define the problem first. What is happening, where, how often, and what does it cost customers, staff, safety, quality or delivery?

A vague line such as 'the process is inefficient' gives the team nothing to test. A specific one does: 'customer orders wait two days for review because requests are batched and assigned by hand.' The specific version already points at the batching and the manual assignment.

The PDCA cycle

PDCA is the cycle behind lean improvement. Plan a change based on the cause of a problem. Do it on a small scale. Check the result against the plan. Act by adopting, adjusting or abandoning the change.

Each turn of the cycle is an experiment. A failed experiment still teaches the team something, which is why lean treats it as progress.

A3 reports put a PDCA cycle on one sheet for larger problems. The A3 step-by-step guide shows how.

One PDCA cycle, worked through

An illustrative example of one full turn of the cycle.

Plan: a packing team misses its 2 pm truck on about 3 days in 10, because orders released after 11 am pile up. The team predicts that releasing orders every hour, instead of in two large batches, will cut misses to 1 day in 10 or fewer.

Do: the team trials hourly release on one packing bench for two weeks. Check: misses on the trial bench fall to 1 day in 10, but pickers report walking the same aisles twice as often.

Act: the team adopts hourly release and starts a second cycle on pick routes, because the first change moved part of the problem upstream. The prediction made in the plan step is what makes the check meaningful.

Standard work is the baseline

Improvement needs a baseline. Standard work records the current agreed way to do a job.

When a change works, the team updates the standard. When results slip, the standard shows what changed. Without it, gains disappear without anyone noticing.

Daily kaizen vs kaizen events

Daily kaizen is small improvements made as part of normal work. An operator moves a bin closer, or a team changes a label that keeps causing mistakes.

Kaizen events are short, focused workshops on one process or problem. A cross-functional team maps the current state, tests changes and leaves a new standard behind.

Events make visible step changes, while daily kaizen keeps improving between events and stops gains from slipping.

The kaizen event guide gives a day-by-day agenda for running one.

Continuous improvement examples

Continuous improvement strategies

A continuous improvement strategy is the plan for how improvement happens every week, not only in projects. Five strategies follow.

Make problems visible. Boards, andon signals and simple charts show where work is stuck, so problems surface early.

Give every idea an answer. A short weekly review decides each idea: do it, test it or explain why not.

Coach rather than fix. Leaders ask questions at the gemba and help teams run their own PDCA cycles.

Link improvement to strategy. Hoshin kanri points improvement effort at a few goals that matter, so small changes add up.

Standardize each gain. A change is finished only when the standard work and training are updated.

Finding improvement opportunities

Ranking ideas by impact and effort

Collecting and tracking ideas

Most improvement ideas come from the people closest to the work. A simple idea board, reviewed weekly, is enough to start.

Each idea needs an owner, a next step and a decision. Ideas that sit unanswered teach people to stop suggesting them.

Kaizen software and continuous improvement software help when ideas come from many teams or sites, and when you want to share what worked.

A simple idea board

Four columns are enough: new, being tested, done and not doing. Each card carries six fields: the problem in one line, the idea, who raised it, the owner, the next step with a date, and the decision.

A 15-minute weekly review at the board keeps it alive. Decide every new card: do it, test it, or not doing with a reason given to the person who raised it. Check cards under test against their dates, and thank the people whose ideas reached done.

Ideas that need money or another department go to the next level's review within a week, so the team sees them move.

Making improvement last

Leaders keep improvement alive through routine. A daily huddle at a board, regular gemba walks and a weekly review of open ideas all help.

Time is the other ingredient. Teams that are fully loaded with production work have no capacity left to improve it.

The lean management guide covers the daily management system and leader standard work in more detail.

The questions leaders ask set the tone. 'Who made this mistake?' teaches people to hide problems. 'What did the process make hard?' or 'What let this error reach the next step?' teaches them to look for causes.

Why improvement programs stall

Measuring progress

Track a few results that matter to the customer, such as lead time, quality and on-time delivery. A falling idea count with a rising time to close usually means ideas are collected but not answered. OEE suits equipment-heavy processes.

Track the improvement activity too: ideas raised, ideas closed and time to close. A healthy system raises and closes ideas steadily. Illustrative levels for a team that has run an idea board for a year: about one idea per person per month, most ideas closed within 30 days, and over half put into practice. These are examples to calibrate against, not benchmarks.

Systematic improvement can pay back at a national scale. A 2011 economic evaluation for NIST by Albert Link and John Scott put the benefit-to-cost ratio of the Baldrige Performance Excellence Program at 820 to 1, comparing the benefits reported by 273 award applicants from 2007 to 2010 with the cost of running the program.

Pair outcome measures, which show whether results improved, with process measures, which show whether the team changed the conditions behind them.

Balance them too. Speed measured without quality invites rework. Utilization measured without flow keeps people busy while customers wait. Cost measured without customer value can remove a step that was quietly protecting the result.

Starting continuous improvement: nine steps

Frequently asked questions