← Insights

Inspections

Safety inspections worth scoring

July 2026 · 6 min read

Walk any factory, warehouse or site with a manager and you will get an inspection, of sorts. They notice a blocked fire exit, a coiled hose against a wall, a forklift moving faster than it should. They mention it, maybe someone jots it on a pad, and everyone moves on. That is a safety walk, and it is genuinely useful. It is also, on its own, close to worthless as a management tool - because a walk is one person's impression on one day, and next month a different person will walk a different route and notice different things.

An inspection is a different animal. It asks the same questions, in the same order, every time. It scores what it finds against a written scale, rather than a shrug or a nod. And because the result is recorded, it survives the person who did the walking - it becomes evidence you can point to, and a trend you can plot, rather than a memory that fades by the next shift change.

Why scoring beats a tick

Most inspection sheets that exist are pass or fail: a column of boxes, ticked or not. It feels rigorous. It is not, for one simple reason - pass or fail hides direction. An area that passes 18 of 20 points this month and 18 of 20 next month looks identical on paper, even if the two failing points are completely different, and even if the near-misses that scraped a pass are getting worse underneath the line.

Score each point 0-4 against written guidance instead, and the picture changes completely. An area at 78 percent this month, 85 percent next, is visibly improving - and everyone can see it, including the team that put the work in. An area sliding from 85 to 78 is visibly worth a conversation before it becomes an incident. Neither of those signals exists in a pass or fail sheet.

The written guidance matters as much as the scale. A bare 0-4 number is only as useful as its anchor: without a description of what a 2 looks like versus a 4, scores drift upward over time (everyone likes giving a good mark), and they drift apart between inspectors (one person's 3 is another's 4). Written criteria for each score keep a 3 meaning the same thing in March and in September, and the same thing whoever is holding the clipboard.

N/A that does not punish the small area

Not every point applies to every area. An office does not have a machine guarding question to answer; a small storeroom does not have a loading bay to check. The honest way to handle that is a genuine N/A option, and the important part is what happens to it in the maths.

An N/A answer has to come out of both the section score and the overall percentage, not just sit there as a zero in disguise. If it does not, a small area with fewer relevant points gets punished for the checklist's generality rather than its own condition, and the inspection stops being fair to compare across areas of different size and function. Done properly, an area with three genuine N/A answers and full marks on the rest scores the same as an area with no N/A answers and full marks on the rest - because it should.

Findings that own themselves

The inspection score tells you the state of the area. It does not, by itself, fix anything. That happens in the findings log, and the findings log is where most inspection programmes quietly fail - not because nobody wrote the finding down, but because nobody closed it. A finding worth logging needs four things:

• A risk level. Not every finding is equally urgent; a trailing cable is not a missing fire extinguisher, and the log should sort accordingly.
• A named owner. A finding with no owner belongs to everyone, which in practice means it belongs to no one.
• A due date. Without one, "soon" becomes the default answer, indefinitely.
• A status that flags itself. Open, closed, and - critically - overdue, calculated automatically from the due date rather than relying on someone remembering to chase it.

An inspection that generates a long, tidy list of findings nobody closes is not a safety programme. It is theatre with better formatting.

The score tells you where you stand. The findings log is the only part of the process that actually changes anything.

The six areas a general inspection should cover

A workplace inspection that tries to check everything ends up checking nothing well. A focused set of areas, asked consistently, works better than an exhaustive list asked once:

• Housekeeping and access. Clear walkways, tidy storage, unobstructed exits and escape routes.
• Fire and emergency. Extinguishers in date and accessible, signage visible, assembly points known.
• Electrical and plant. Guarding in place, cables and leads in good condition, equipment within test dates.
• Hazardous substances and storage. Correct labelling, safe storage segregation, safety data sheets available where they should be.
• PPE and welfare. The right equipment available, worn and in usable condition; welfare facilities maintained.
• Behaviour and documentation. Safe practice actually being followed, not just permitted, and the paperwork behind it up to date.

Those six areas cover most of what a general workplace inspection needs, whether the site is a factory floor, a warehouse, a construction site or an office. The specific points under each heading will vary a little by workplace, but the structure travels well.

Habits that keep the inspection honest

• Do not announce the exact hour. A scheduled-to-the-minute inspection measures how fast a team can tidy up in the ten minutes before the inspector arrives, not the normal state of the workplace. Announcing the week is fine; announcing the hour defeats the point.
• Rotate who inspects. The person who works in an area every day stops seeing its problems - familiarity is generous, and it is generous in a way that quietly inflates scores over time. A rotation, or an occasional cross-check by a second inspector, keeps the scoring honest.
• Act on the low scores first. A 0 or a 1 against a written scale is not a mystery; it is a specific, described gap. Work the findings list from the bottom up, and the next inspection should show it.

Where this sits alongside the other scored audits

If a layered process audit checks whether standard work is actually being followed, and a 5S audit checks whether workplace organisation is holding, a safety inspection checks the physical and behavioural condition of the workplace itself. All three share the same underlying discipline: the same questions every time, a written scale instead of a gut feel, and a trend rather than a single snapshot. Run them together and each one reinforces the others; run any one of them as a one-off exercise and it tells you almost nothing beyond the day it happened.