What is root cause analysis?

The discipline of asking why something went wrong until the answer stops moving, so that fixing the root removes the symptoms above it.

When should a group do root cause analysis instead of an analyst?

When the causes are spread across the people involved and no single analyst holds them all. Structured Democratic Dialogue collects causes from everyone and lets everyone judge the links at an 80% bar.

How large should a root cause analysis group be?

Twelve to twenty. In simulation panels of fifteen find 4.1 of five known causes and panels of fifty 3.6, because a fixed 80% bar gets harder to reach as a panel grows. Larger rooms are split into groups of fifteen and their maps merged.

Root cause analysis example: delivery delays, from cards to root causes

Root cause analysis example

Fifteen people who work in and around a regional delivery operation — drivers, dispatchers, a depot manager, two customers — answer one question. The stages, what each produced, and the map at the end. Wording is the group's own, lightly shortened.

The question

What is driving delivery delays across the region?

Generate → 61 cards.

A sample: loads are dispatched before the paperwork is checked · drivers get routes they cannot finish · route plans ignore school-run traffic · depot radios are on the wrong channel · yard slots are booked twice on the same hour · handovers between shifts lose key details · customs forms are filled by the wrong desk · agency drivers never see the site notes · fuel cards are shared between drivers · the scanner app logs out every hour.

Clarify → 61 cards, 4 merged.

"Handovers lose key details" and "the 6am handover has no checklist" turned out to be one claim. "Routes they cannot finish" was clarified to mean planned stops, not driving time — which changed what people linked it to later.

Cluster → 7 clusters.

Paperwork and customs · route planning · yard and slots · handovers and briefing · equipment and radios · agency staff · customer-side constraints.

Select → 15 on the shortlist, 2 by seat.

Spreadthink 41% — 25 of 61 cards got at least one sticker, which is the usual range. "Customs forms are filled by the wrong desk" got no stickers but a dispatcher nominated it with the customs office's own returns figures; two people agreed the source was relevant and it took a seat.

Structure → 210 pairs, 86 asked, 79 settled, 7 contested.

The rest were implied. The contested pairs were all about whether agency drivers cause or merely suffer the other problems; the room split 8–7 on those and they stayed on the record as split.

The map.

Level 4 (symptoms): late deliveries · missed customer windows · overtime
Level 3: routes drivers cannot finish · yard slots double-booked · scanner app logging out
Level 2: route plans ignore school-run traffic · handovers between shifts lose details · agency drivers never see site notes
Level 1 (root causes): loads are dispatched before the paperwork is checked · customs forms are filled by the wrong desk

Two roots. The second was the seated card that got no votes. Every symptom the room had ranked highest — late deliveries, overtime — sits three levels above them. Nobody had come in believing paperwork was the problem; the links, judged one at a time, put it at the bottom.

Actions → 3 judged against the roots.

"Paperwork check before dispatch, with a hard stop" moved both roots at 80%. "Give agency drivers the site notes" moved neither — a good action, not a root-cause action, and the matrix said so. "Fix the scanner app" was withdrawn by its author once the map showed it three levels up.

What a five-whys session would have found: the depot manager's chain, which ran from late deliveries to route planning and stopped there — level 2 of this map. The paperwork root sat in a dispatcher's head and a customs officer's figures, and no analyst was going to ask either.

Run the same seven stages · Methods compared · The same process for any number of groups