What this fixes
A problem keeps coming back. Each time, you find a plausible cause, patch it, and move on - and a few weeks later, it returns, in a slightly different form.
That’s the trap. The first cause you land on is usually the comfortable one. It explains the symptom, patching it feels productive, so the drilling stops. But a cause you can patch isn’t always the one driving the problem. The first cause is often where curiosity stops, not where the problem starts.
This tool makes you keep drilling - and, crucially, keep proving. You trace the problem down one evidenced step at a time, and you’re not allowed to move to the next cause until the current one is backed by evidence. The chain ends at a root you can actually remove.
How this works
Imagine this: your team’s releases keep slipping. The obvious cause is “we underestimate the work”. You could stop there and add buffer time - and slip again next release.
You build a chain, one link at a time. Start with the symptom, then lay the first cause beneath it. Before you can add the next cause down, you have to show evidence that the current one is real - a log, a number, an example, not a guess. Then ask what caused that proven cause, add the next link, and prove it. Repeat, going down, until removing the bottom cause would stop the recurrence.
The evidence is the whole point. The moment you can’t back a cause, the chain stops - and that’s usually where the room realises it’s been guessing, not drilling. Unlike Five Whys, every answer must be backed by evidence before you move on.
You’ve reached the root when removing it would stop the recurrence - not just reduce this instance. Then you read the finished chain back from the root up to the symptom: the proof that this cause really does drive the problem.
Team: one person acts as the facilitator - they run the activity using the facilitation guide while everyone else takes part.
Time required
Solo ✓ Team ✓ Solo: about 20–30 minutes. Team: 40–60 minutes.
The time required is indicative and will depend on how many layers sit beneath the symptom.
The facilitator should read the short facilitation guide first - about 6–8 minutes. It is a downloadable PDF on this page.
When to use
Bring one real problem that has come back after being “fixed” at least once - a live recurrence, not a hypothetical.
Use it whenever:
What you get
You’ve done it well when: every link in your chain has evidence behind it, and the root is something you could actually remove - not another symptom to patch.
What you’ll notice
Why this works
Recurring problems survive because people stop at the first plausible cause. It feels productive to patch it, so the investigation ends too early - even though the information to go deeper is usually right there. Requiring evidence at every link keeps the team drilling until the real cause is found, instead of the first comfortable explanation.
What you need
One recurring problem. A specific problem with a real symptom you can point to. Choose one that has already been returned after being fixed at least once. Bring a live one, not a hypothetical.
The capture sheet (supplied). The linked cause boxes where the chain is built: symptom at the top, then a cause and its evidence at each step down, ending in the named root. An unfinished chain is meant to look unfinished. Leave space to add more links if the chain grows.
Evidence to hand. Logs, numbers, timelines, screenshots, examples - whatever lets you back each cause. Part of the activity is noticing where you don’t have it.
Roles - Team: a facilitator (runs the flow, holds the evidence gate, writes the chain) and participants (propose causes, bring evidence, decide the root). Solo: you run it yourself from the capture sheet.
Steps
Output / Takeaways
You leave with:
Important rules
Planning to facilitate this activity? Access the resources below to get started.