Whimsical LogoWhimsical LogoWhimsical Logo

Brand
Get all logo versions.
Download

Root Cause Analysis

A root cause analysis (RCA) digs past a problem's first apparent cause to the condition that actually produced it, the way a doctor treats the illness rather than the fever. This template shows one worked end to end: a bad espresso traced through three symptom branches, off taste, low pressure, no crema, that all converge on one root cause, a grinder nobody maintained. Run the same structure on your outages, defects, and recurring failures.

A worked RCA: three symptom branches, one shared root cause.

What's included

  • A complete worked example. Bad Espresso at the root, real cause chains below it, no placeholder boxes to decode.
  • Three symptom branches. Off taste, low extraction pressure, and no crema, each drilled level by level.
  • A convergence story. All three branches implicate the same grinder, which is how real root causes announce themselves.
  • An explicit root-cause node. The chain ends at 'Root cause: grinder not maintained or calibrated', the thing you'd actually fix.
  • Mind-map mechanics. Swap in your own symptom and branch with Enter and Tab; the structure does the discipline.

Why run a root cause analysis?

  • Fix it once. Restarting the server ends the incident; finding the missing health check ends the incidents.
  • Convergence finds the real culprit. When three independent symptoms all trace back to the same node, the argument about what to fix is over.
  • Evidence over instinct. A written cause chain can be challenged node by node; a hunch in someone's head can't.
  • Blameless by structure. Branches end at processes and conditions, not at whoever was on call.
  • The artifact outlives the incident. Six months later, the map still shows what you ruled out and why, which is half the value of having done it.

How to use this template

  1. Write the symptom at the root. Stated plainly and measurably: 'checkout conversion dropped 12% on Tuesday', not 'site issues'.
  2. Branch every plausible cause. Level 1 holds the candidates; resist crowning a favorite before the branches exist.
  3. Ask why, twice more. Each level-1 cause gets level-2 causes, and again; three levels is where guesses run out and investigation starts.
  4. Mark verified versus suspected. Color-code nodes you've confirmed with evidence; the rest are hypotheses, not findings.
  5. Stop at actionable. A branch is deep enough when the cause is something your team can change.
  6. Attach the counter-measures. The causes worth fixing each get an owner and a follow-up; an RCA without actions is a diary.

Root cause analysis vs troubleshooting

Troubleshooting gets things working again; root cause analysis stops them from breaking the same way. Troubleshooting is fast, experience-driven, and stops at the first fix that holds: restart the service, replace the part. RCA is slower and systematic: it maps the causal chain past the immediate trigger to the condition that allowed it. You usually need both, in order. Troubleshoot to recover, then run the RCA before the lesson evaporates.

Frequently asked questions

  • A root cause analysis (RCA) is a systematic process for finding the underlying cause of a problem instead of patching its symptoms, so the fix prevents recurrence rather than postponing it. It's standard practice after incidents, defects, and recurring process failures. RCA is the methodology; techniques like the 5 whys, fishbone diagrams, fault tree analysis, and Pareto charts are the tools used inside it.

  • The board on this page is one: a cafe's espresso turns bad. Three symptoms get their own branches: the taste is off (sour or bitter shots), extraction pressure is low (shots pull too fast), and no crema forms (weak, watery body). Drilling each branch, the same culprit keeps appearing: inconsistent grind from a grinder nobody maintained or calibrated. Fix the cleaning schedule, and all three symptoms go with it.

  • Often, and sometimes the opposite happens: several problems share one. Complex failures usually have contributing causes across process, tooling, and environment, which is why branching beats a single chain. But watch for convergence, like this template's espresso example, where three separate symptoms all trace to one unmaintained grinder. When independent branches keep landing on the same node, you've found the fix with the highest payoff.

  • Root cause analysis is the umbrella process; the 5 whys and the fishbone diagram are techniques inside it. The 5 whys drills one cause chain by asking why repeatedly, which suits simple, linear problems. The fishbone spreads causes across categories, which suits messy multi-causal ones. This mind-map template sits between them: it branches like a fishbone but drills like the 5 whys.

  • A symptom is what you observe: the error rate spiked, the part failed, the customer churned. A root cause is the condition that produced it, usually two or three whys deeper: the missing monitor, the worn process, the unowned handoff. The test: if you fix it and the problem can still come back the same way, you fixed a symptom. The smoke-versus-fuel distinction is the whole discipline.