Poka-Yoke Under £50: Error-Proofing Your Cell This Week

The drill jig on Cell 7 had three bolt holes drilled in a triangle. The assembly only needed two of them. The third hole had been wrong since the cell was set up, and for two years every operator who used it drilled it correctly by ignoring hole number three. When the experienced operator went on holiday, his cover drilled all three. The part failed inspection. Nobody had ever made the jig impossible to use wrong.

That is what poka-yoke is actually for.

It Is About the System, Not the Person

Japanese manufacturing borrowed the term from Shigeo Shingo in the 1960s. The English translation is roughly “mistake-proofing” or “fool-proofing,” though Shingo himself preferred the former. The principle is simple: if a mistake is physically possible, eventually somebody will make it. Your job is to remove the possibility, not write a procedure about it.

Most quality problems in a mid-market UK cell are not caused by careless people. They are caused by systems that let careful people make errors at speed. A procedure pinned to a board three metres from the assembly point is not a control. A shaped cutout that only accepts the correct orientation of a part is a control.

The difference matters because one of them works on a Monday morning and one of them does not.

Start With a Fixture or a Guide

The cheapest poka-yoke you can build is a physical guide that makes the correct position obvious and the incorrect position awkward. A piece of aluminium angle screwed to a jig plate, shaped to accept the part in one orientation only, costs under a fiver in materials and an hour of a setter’s time.

Walk your cell and ask: where does orientation matter? Where does sequence matter? Where does the operator have to remember something rather than see something? Those are your candidates. Pick the one that causes the most rework and fix it first.

Colour and Shape Do More Work Than You Think

Shadow boards are not just for 5S audits. A shadow board that shows a red silhouette when a tool is missing tells the next operator something is wrong before they start. A set of colour-coded bins, one colour per part family, removes a decision from the operator’s hands entirely.

You can do a lot with coloured tape, coloured lids, and coloured labels before you spend a penny on anything else. The rule is that the correct choice should be visually obvious at a glance, and the incorrect choice should require effort to make.

Limit Switches and Sensors Under £50

If your process has a step that must complete before the next step begins, consider a limit switch. A basic magnetic reed switch from an industrial supplier costs around £4. Wired into a simple indicator light circuit, it tells the operator when a clamp has fully closed, a door has fully opened, or a component has seated correctly.

You do not need a PLC for this. A 12V circuit with a switch and an LED is enough to error-proof a manual step that is currently relying on the operator’s feel or memory.

For something slightly more capable, microswitch-based presence sensors start at around £8 each. A pair of them, positioned correctly on a fixture, will tell you whether a part is there and whether it is the right way round.

Go/No-Go Gauges From Bar Stock

A go/no-go gauge is one of the oldest error-proofing tools in manufacturing. You can machine a simple one from a piece of aluminium or steel bar for the cost of the material and an hour on the lathe. The gauge physically cannot pass over a part that is out of tolerance. The operator does not need to read a number or make a judgement.

If you have a dimension that fails inspection more than once a month, a go/no-go gauge for that dimension will pay for itself quickly. If you do not have a lathe, a local engineering sub-contractor will make you a simple gauge for well under £50.

The Principle Behind All of It

The common thread is this: the correct action should be easier than the incorrect action. Not labelled as correct. Not described as correct. Physically easier. An asymmetric connector that only plugs in one way is more reliable than a label that reads “this side up.” A shaped slot that accepts only the right part is more reliable than a work instruction that lists the part number.

When you walk your cell looking for poka-yoke opportunities, ask one question at each station: what would have to go wrong for an error to occur here, and can I make that physically impossible? If you cannot make it impossible, can I make it immediately visible?

You will usually find three or four candidates in an hour. Pick the one with the highest impact and build it this week.

If you want a structured way to find where errors are actually costing you, the losses calculator is worth running before you start. It tends to surface the stations that look fine on paper but are quietly eating margin. Once you know where to focus, the diagnostics tool can help you see where poka-yoke sits alongside your other cell priorities.