Question 1 of 8: The Five Whys and Poka-Yoke

Question 1: The Five Whys and Poka-Yoke (20 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

(a) “Ask Why Five Times”: Effect and Reason

The technique referred to is the Five Whys, developed within the Toyota Production System under Taiichi Ohno as TPS matured through the 1950s and into the early 1960s. It is a root-cause-analysis method: when a problem or defect occurs, the investigator asks “why did this happen?”, and then asks “why?” again of the answer just given, repeating the question (traditionally five times, though the number is a guideline, not a rule) until the causal chain reaches a true, actionable, process-level root cause rather than stopping at the first, most visible, or most convenient explanation. A representative chain for a stopped machine might run: the machine stopped → why? a fuse blew from an overload → why? the bearing was insufficiently lubricated → why? the lubrication pump was not circulating enough oil → why? the pump's intake was worn → why? there was no strainer fitted to keep metal shavings out of the pump — a root cause (a missing strainer) entirely different from, and far more useful than, the symptom first reported (“the machine stopped”).

The effect is that each successive “why” forces the investigation one causal layer deeper, moving from an equipment symptom toward a process, design, or maintenance-system deficiency that, once corrected, prevents not just this failure but the whole family of failures the same root cause would otherwise keep producing. It works because most problem investigations naturally stop at the first plausible cause, which is usually only a proximate symptom of a deeper systemic issue; fixing that symptom (replacing the blown fuse, in the example) restores production immediately but leaves the true cause untouched, so the same failure recurs. The technique is deliberately cheap and simple — no statistical tooling or specialist training is required — which is precisely why it works as a shop-floor discipline: it can be applied by any operator or engineer, immediately, at the point where the problem actually occurred, which pairs naturally with jidoka (stopping the line the instant an abnormality appears) so the causal evidence is still fresh and directly observable rather than reconstructed later from memory or paperwork.

(b) Poka-Yoke

Poka-yoke (Japanese for “mistake-proofing,” literally “avoiding inadvertent errors”) is a technique developed by Shigeo Shingo at Toyota for designing a process or a piece of tooling so that a human error either becomes physically impossible to commit or is caught and signalled the instant it occurs — before a defective part can travel any further downstream. Shingo's own framing distinguishes it sharply from inspection-after-the-fact quality control: rather than sampling finished output and rejecting defects statistically, poka-yoke intervenes at the source of the error itself, so the defect rate it targets is not “reduced” but driven toward zero.

Poka-yoke devices fall into two functional categories. A control (prevention) poka-yoke physically stops the process from continuing once an error condition is detected — for example, an asymmetric locating pin on a fixture that only allows a part to be clamped in the single correct orientation, so an operator cannot even begin the next operation on a misloaded part. A warning (detection) poka-yoke does not halt production automatically but immediately alerts the operator with a light, buzzer, or counter mismatch — for example, a parts tray with one recessed slot per fastener, where any slot still full at the end of a cycle signals a missed screw before the assembly leaves the station. Within either category, Shingo further classifies the sensing method as a contact method (a limit switch or photoelectric sensor that detects a part's physical shape, size, or presence — e.g., a probe that will not seat unless a hole was actually drilled), a fixed-value method (a counter that confirms the correct number of repeated actions occurred, such as the exact number of bolts torqued per unit), or a motion-step method (a sensor confirming that every step of a standardized work sequence was actually performed, in order, before the next station can begin). A now-familiar consumer-facing example of the same principle is a USB-C or SD-card connector shaped so it physically cannot be inserted upside down.