Large improvement projects get the attention: a new machine, a new system, a redesigned layout. But much of what makes a workplace efficient, or not, sits in small things. A tool kept in the wrong place. A label printer shared by three benches. A form filled in twice. A part handled five times on its way through. None of these is worth a project on its own, and together they can absorb a large share of a working day.
These losses persist because the people closest to them have stopped seeing them. Walking to the cupboard for tape is just part of the job. Waiting for the printer is normal. Everyone is busy, so the work looks productive. Seeing waste requires deliberately stepping back and watching, and fixing it requires a routine that turns small observations into small, tested, lasting changes.
This article explains how to observe work properly, a simple cycle for fast local improvements, how to lock in a better method so it does not drift, how to keep workplace organisation from fading, and how to tell whether the improvement routine itself is working. It is general information for owners, supervisors and teams in workshops, warehouses, kitchens and any other hands-on workplace.
Go and see: observing the work
The Japanese term gemba means the place where the work happens. A gemba walk, or more usefully a gemba stand, means going there and watching. A practical approach:
- Pick one spot and stay there for 30 to 60 minutes. Moving around turns observation into a tour.
- Watch the work, not the people. The question is how the process makes people move, wait, search and repeat, not whether anyone is working hard.
- Write down what you see: walking, waiting, searching, reaching, handling the same item again, fixing errors, workarounds.
- Do not solve problems on the spot. The urge to intervene is strong, especially for experienced people. Note it and keep watching.
- Explain why you are there. Tell people beforehand that you are looking at the process, and afterwards ask them about what you saw: they usually know the causes and have ideas.
Observing with someone from another area often helps, because a fresh pair of eyes notices what familiarity hides.
What to look for
A useful checklist when observing comes from lean thinking, which groups waste into a few familiar kinds:
- Waiting: for materials, information, approvals, machines or other people.
- Motion: walking, reaching, bending and searching.
- Transport: moving materials or items more than necessary.
- Extra processing: steps the customer does not need, such as re-entering data or over-finishing.
- Excess stock: more materials or work in progress than the next step needs.
- Defects and rework: anything done twice.
- Making too much or too early: producing ahead of need.
- Unused skills: people’s knowledge and ideas left untapped.
The lean as an operating system article covers how these fit into a wider way of running work.
A fast local improvement cycle
Point kaizen is a name for small, local, quick improvements led by the people who do the work. A seven-step loop keeps them disciplined:
- Identify a local problem or waste at a workstation or small area.
- Find the cause using local data and the team’s knowledge, for example by asking “why” a few times.
- Develop a countermeasure: small, practical and low cost, designed with the people doing the work.
- Trial it on a small scale for a short time.
- Measure and check against the baseline, and ask the people using it what they think.
- Standardise it if it works: update the method, the instructions and the visual cues.
- Share and sustain: tell others, recognise who contributed and check the new method is still being followed.
The defining features are scope and speed: one workstation or area, days or weeks rather than months, mostly low-cost changes, and the people doing the work leading. For chronic problems with unclear causes that cross several areas, a structured project is better. The fixing a recurring problem with DMAIC article covers that approach.
Once an improvement is proven, treat it as the new standard and the starting point for the next improvement: standardise, do, check, act. Improvement without standardisation drifts back.
Lock in physical improvements
For changes involving a new fixture, layout or tool, a five-step routine stops the gain disappearing:
- Measure the baseline: the current time, walking, rework and sequence.
- Design and prototype the aid: a fixture, jig, shelf or tool holder, tested before it is made properly.
- Write standard work: the new method, usually with photos, clear enough that another trained person can repeat it.
- Update the system: routings, job times, material lists or other records that planning and costing rely on.
- Re-measure and keep tracking long enough to confirm the change has held.
Each step prevents a common failure. Without a baseline, the saving cannot be proved. Without standard work, the result depends on one person. Without the system update, planning and pricing drift from reality. Without re-measurement, the improvement is only an estimate.
Keeping 5S alive
5S is a method of workplace organisation, usually described in five steps: sort (remove what is not needed), set in order (a place for everything), shine (clean and inspect), standardise (make the arrangement the normal way) and sustain (keep it going). The first three are easy and satisfying. The last two are where most workplaces fail. A big clean-up produces a tidy area for a few weeks, then clutter slowly returns.
5S fades for predictable reasons: it is treated as a one-off event, it is called housekeeping and seen as cleaning rather than as part of doing the work well, nobody owns each area, and leaders stop looking. To keep it alive:
- Give each area an owner who maintains its standard.
- Make the standard visible: outlines, labels and photos of how the area should look.
- Build in short daily routines, such as a five-minute reset at the end of each shift, rather than occasional big clean-ups.
- Use simple tests, such as whether anyone can find a given tool or item within a minute or two.
- Have leaders walk the areas regularly and ask about the standard, not just about output.
- Link it to the work: fewer searches, fewer errors, safer movement. Organisation that does not make the work better will not last.
A well-organised workplace also tells customers and new staff something about how the business works.
Check that the improvement routine works
An improvement routine needs its own measures. Counting ideas alone is misleading; people can submit many ideas that go nowhere. A better small dashboard:
| Measure | What it shows |
|---|---|
| Ideas raised per person | Whether people feel ideas are welcome |
| Share reviewed within a set time | Whether the business responds quickly |
| Share trialled | Whether ideas move from talk to action |
| Share standardised | Whether trials become the new normal |
| Verified benefit | Whether the changes improve safety, quality, time or cost |
| Spread of contributors | Whether participation is broad or comes from a few people |
A high number of ideas with few trials usually means a bottleneck in approval. Low participation in one area or shift may mean those people do not believe their ideas will be acted on. Avoid targets on idea counts, which invite trivial submissions. The turning workplace initiative into useful improvements article covers supporting individual initiative.
Respect makes improvement last
Improvement routines fail quickly if people experience them as surveillance or as a prelude to job cuts. Be clear about why you are observing, credit the people whose ideas work, and say what time freed by improvements will be used for, such as taking on more work, reducing overtime or improving quality. People improve work they feel part of, and quietly resist changes done to them.
Boundaries still apply
Small and fast does not mean uncontrolled. Changes that affect safety guarding, product specifications, validated settings, approved work instructions, drawings or customer requirements must go through the business’s normal change controls. A quick improvement to how tape is stored needs no approval; a change to a machine’s safety interlock or a product’s specification always does.
A worked example
This is an illustration. A 20-person spare parts distributor packs about 300 orders a day across four packing benches. The owner feels the packing area is always busy but often behind.
Gemba. The warehouse supervisor stands near the benches for 45 minutes. She notes that packers walk to a shared cupboard for tape and labels many times an hour, queue at a single label printer, and regularly repack orders after choosing the wrong box size. Afterwards, she talks with the packers, who confirm the problems and suggest fixes.
Baseline. A simple timing over a week shows average packing time of about 4.5 minutes per order.
Point kaizen. The packers and supervisor trial three small changes over two weeks: a tape and label station at each bench, costing about $300 in total; a printed box-size guide matched to common part dimensions; and a second label printer moved between two benches. Average packing time falls to about 3.6 minutes per order. Across 300 orders a day, that saves about 270 minutes, or roughly 4.5 hours of packing time each day.
Standard work. A one-page photo card at each bench shows the layout and the box-size guide. The standard packing time in the system is updated so order planning reflects the new pace.
5S. Shadow boards and labelled shelves are set up at each bench, each bench has an owner, and a five-minute reset closes each shift. The owner walks the area for ten minutes each week and asks each packer about their bench standard.
The routine. An improvement board lists ideas, the date raised and their status. Over the next three months, 18 ideas are raised, 12 are trialled and 9 become standard. Two packers who had never suggested anything before contribute ideas once they see earlier ones acted on within a week.
How this applies to a small Australian business
- Stand and watch the work regularly, at one spot, without solving on the spot.
- Ask the people doing the work about causes and fixes.
- Use a seven-step loop for small, local improvements.
- Standardise what works, including system records.
- Give every area an owner and a visible standard.
- Build short daily routines instead of occasional big clean-ups.
- Track the improvement routine from idea to verified benefit.
- Keep change controls for safety, quality and specifications.
- Say what freed time will be used for.
Signals worth watching
- Busy people and late work at the same time.
- Frequent walking to fetch tools or materials.
- Queues at shared equipment.
- Tidy areas that slowly return to clutter.
- Many ideas raised and few trialled.
- Improvements that depend on one enthusiastic person.
- People unsure what happens to time their ideas save.
Common mistakes
- Observing on the move and missing what really happens.
- Fixing problems on the spot before understanding them.
- Treating 5S as cleaning rather than part of the work.
- Running big clean-up events without daily routines.
- Counting ideas instead of tracking them to results.
- Skipping standardisation, so improvements drift back.
- Using improvement as cover for cuts, which kills participation.
Frequently asked questions
How often should managers do gemba walks? Regularly and briefly, such as weekly, works better than occasional long visits. The habit matters more than the length.
Doesn’t watching people make them uncomfortable? It can. Explain beforehand that you are studying the process, not judging people, and share what you saw with them afterwards.
Do we need Japanese terminology? No. Use whatever words your team prefers. The practices matter, not the names.
How do we find time for improvement when we are busy? Small improvements usually free time quickly. Start with one area and one problem, and protect a short regular slot for the routine.
Does this work in offices too? Yes. Watching how invoices, quotes or orders move through an office often reveals the same waiting, searching and re-entry as a workshop.
Should we reward ideas? Recognition works better than payment for small ideas. Acting quickly on ideas is the strongest encouragement of all.
Questions to ask
- When did someone last stand and watch this work for half an hour?
- What do people walk, wait or search for most often?
- Which improvements have we standardised, and which have drifted?
- Who owns the standard for each area?
- How quickly do we respond to an idea?
- Do our improvements show up in our system records and planning?
Bringing it together
Small, frequent improvements depend on seeing the work clearly and acting quickly. Stand and watch at one spot, involve the people who do the work, and use a short loop to trial and standardise local fixes. Lock physical improvements in with a baseline, standard work, system updates and re-measurement. Keep workplace organisation alive through owners, visible standards, daily routines and leaders who keep looking. Then measure the routine itself, from idea to verified benefit, and treat people with respect, so improvement becomes part of how the business works every day.
Source: KEVOS notes, drawing on earlier KEVOS manufacturing handbooks on gemba observation, the point kaizen cycle, the five-step shop-floor improvement method, 5S sustainment and measuring participation in continuous improvement. Examples and figures in this article are illustrations. This article is general information.
