A full tray of chocolate-chip cookies went from golden-brown to charcoal while the head baker at Sweet Crumb Bakery was on the phone. The kitchen filled with smoke. The oven had to cool down and get scraped out. Forty-five minutes of Saturday revenue vanished into the ventilation hood.
It is the kind of incident everyone understands. That is exactly why we are using it to announce something we have just shipped: 3-legged 5-Whys are now built into RCA Map.
You will see it on the "why" nodes directly under an incident: a small label that reads Occurrence, Detection, or Systemic. That label is the hinge that turns a flat chain of "why" into a 3-legged investigation. If you want the full methodology, our guide "What is a 3-legged 5-Whys analysis?" covers it. Below, we are going to walk the cookie incident leg by leg.
The incident
Here is the map. The incident sits at the top: the bakery burned a full batch of cookies during the Saturday morning rush.
Leg 1 — Occurrence: what physically failed?
The Occurrence leg is the one most people are already comfortable with. It asks: what actually happened?
The answer, once you dig, is almost insultingly straightforward. The baker put the tray in, got pulled away by a phone order, and never set the timer. The oven kept doing what ovens do. The cookies kept doing what cookies do when left unattended. Chemistry took over, and the kitchen learned what smoke smells like.
This leg is satisfying because it points to a concrete fix: set the timer. But if you stop here, you have answered only what happened. You have not asked why the failure was allowed to continue until it became a disaster, or why the bakery was organized so a single missed step could ruin a batch.
Leg 2 — Detection: why did nobody catch it in time?
This is where investigations usually get interesting. Almost nothing is a true surprise; there is almost always an earlier signal that got ignored.
The smell of smoke was the alarm. But by the time anyone smelled it, the batch was gone. There was no check during the critical window because the kitchen was short-staffed and no one had been assigned to watch the oven while the baker was away. The team was busy, the floor was noisy, and the oven did not beep because the timer had never been set.
The fix on this leg is not about the cookies; it is about the signal. Maybe the oven needs an automatic audible alert. Maybe the Saturday shift needs a posted oven-monitoring rotation. The Detection leg does not stop the oven from failing; it stops the failure from staying hidden until the kitchen fills with smoke.
Leg 3 — Systemic: what shaped the decision?
The Systemic leg is the uncomfortable one. It asks: why did the organization make it easy for this to happen?
The baker was alone on the oven during the busiest shift of the week. The shop had no backup protocol for oven coverage when the primary baker was interrupted. Training assumed one person could juggle ovens, phones, and counter service without ever missing a step. The process was built as if Saturdays were quiet Tuesdays, and the process did not notice the mismatch until the kitchen filled with smoke.
This is the leg that keeps incidents from repeating. It points to the schedule, the staffing model, and the assumption that one person can reliably do three jobs at once. Fixing it might mean a second pair of hands on Saturday mornings, a phone-policy that routes calls away from the person managing active ovens, or a staffing model that does not assume one person can reliably juggle ovens, phones, and counter service at the same time.
Why all three matter
Look at what each leg gives you:
- Occurrence says: set the timer. Good. That addresses the immediate mechanism.
- Detection says: make sure someone notices before the smoke. Better. This leg guards against a wider set of failures — a wrong temperature, a timer that was set but not heard, or any other deviation that lets cookies stay in the oven too long.
- Systemic says: design the Saturday shift so the oven is never left unattended. Best. The whole class of incidents shrinks.
That is the point of the 3-legged structure. A single chain of "why" explains the trigger. The other two legs explain why the trigger became a disaster and why the disaster was allowed to happen at all. Drop a leg and your fix lands on the symptom.
How to try it
This is now available in RCA Map. Create an investigation, and watch the AI branch the incident into three legs automatically.
If you want the methodology behind the labels, read the full guide: "What is a 3-legged 5-Whys analysis?". It walks the Challenger loss as a 3-legged tree — a more serious example, but the same shape.
We are starting with the cookies because the shape of a good investigation does not depend on the size of the disaster. A burned batch and a billion-dollar recall both split the same way: what failed, what let it through, and what decided it.
Happy investigating.