← Insights

Lean

5S dies without an audit

July 2026 · 6 min read

Every factory has run at least one 5S blitz. A weekend of skips and red tags, shadow boards on the walls, floor tape gleaming, before-and-after photos in the company newsletter. And every factory veteran knows what the same area looks like six months later: the shadow board has three tools and five silhouettes, the tape is scuffed to archaeology, and the bench drawer has re-evolved into a museum of mystery fasteners.

5S does not fail loudly. It fades - and it fades for a specific, fixable reason: the fifth S never got built.

The five, briefly - and where the weight sits

Sort (remove what is not needed), Set in Order (a place for everything that remains), Shine (clean as inspection - a clean machine shows its leaks and cracks), Standardise (make the first three the documented normal, with owners), and Sustain (the discipline that keeps it all true next month).

The blitz delivers the first three in a weekend, and they are the visible, photogenic part. But 5S is not a tidiness project; it is a workplace-discipline system, and the whole system stands on the two Ss that cannot be photographed. Standardise turns a one-off clean-up into a defined condition. Sustain is the mechanism that checks the condition still holds - and the only mechanism that reliably works is a regular, scored audit.

Why audit beats enthusiasm

Enthusiasm decays; calendars do not. A monthly 5S audit works because of three properties:

• Same questions every time. Twenty-odd specific, observable questions - are red-tag items dispositioned within the rule? do tools live on their marked locations? are WIP limits respected? has the leadership walk happened? Not "is the area tidy?", which is an invitation to a shrug.
• A strict written scale. Score each question 0-4 against written criteria, so a 3 means the same thing in March as in September, and the same thing when a different auditor scores it. Without written anchors, scores drift upward and the audit flatters its way into uselessness.
• A visible trend. One score per audit, plotted on the wall of the area it describes. The trend is what makes the audit motivating rather than punitive: the team can watch their own line climb - and a dip becomes a conversation, not an ambush.

The audit is not there to catch people. It is there to catch decay - early, while it is still cheap to reverse.

Reading the results: the radar and the trend

Score five questions per S and the results split into a shape and a direction, and each answers a different management question.

The radar (by S) shows where to act. A profile strong on Sort, Set and Shine but weak on Standardise and Sustain is the classic post-blitz signature: the clean-up happened, the system did not. Weakness in Shine on a machining line often points at maintenance discipline, not housekeeping. The shape tells you which S needs the next month's attention.

The trend (audit over audit) shows whether the system works. A score bouncing around 60% is a programme living from blitz to blitz. A score climbing 52, 55, 58, 61 over four months is Sustain doing its job. Direction beats level: an area at 55% and climbing is in better health than one at 75% and sliding.

And the 0s and 1s are the work list. Anything scored 0 or 1 is this month's fix-first list for the area owner - specific, local, achievable before the next audit. That closes the loop that keeps the whole system alive: audit, fix, re-audit, improve.

Three ways teams sabotage the audit itself

• Announcing it. A scheduled-to-the-hour audit measures the team's sprint speed, not the workplace's normal condition. Fix the week, not the hour.
• Auditing your own area. Familiarity is generous. Cross-audit between areas - it keeps scores honest and spreads good ideas as auditors see each other's solutions.
• Letting leadership skip their question. If the audit asks whether the leadership walk happened (it should), and the answer is no for three consecutive months, the team learns exactly how much the programme matters. Sustain fails top-down long before it fails bottom-up.