Work Understanding: Root cause
Objective. Read the root-cause view as a set of observed contributions with confidence, so a reader presents a likely driver of delay as a strong candidate to validate rather than a proven cause.
Root cause answers 'what appears to be driving the delay or rework we see?' It is where the integrity of the whole story is tested - observed contribution is not proven causality.
See it in the demo
What this screen answers
The customer question is: what seems to be causing this delay, friction or rework? A Sales person uses it to move from a symptom to a credible driver with evidence. A CSM uses it to prioritise which candidate driver to test first with an experiment and an owner.
The task
Scope to the finding in question, then read the contributing factors the view associates with it and the evidence behind each. The disciplined move is the integrity line: 'this is an observed contribution to the problem with a confidence level - a strong candidate to validate, not a proven root cause. Here is the experiment we would run to confirm it.'
How to read it
A root-cause factor is an observed contribution with a confidence, not proven causality. The classic misread is presenting the top factor as the confirmed cause and committing to a fix. Correlation in observed signal is a lead, not proof. Read the strongest factor as the first thing to test, and be explicit that confirmation comes from a measured before/after, not from the ranking itself.
Honest limits
The view does not prove causation, and overlapping factors cannot simply be added into one savings number. It does not replace a validation experiment. Confirm detection confidence and coverage with product, and validate the leading candidate with the customer through a before/after before it becomes a claim.
Interpretation check
A root-cause factor is an observed contribution with a confidence, not a proven cause - present it as the strongest candidate to validate with an experiment, never as settled causality.
Practical task
On the root-cause view for one finding, name the leading candidate driver, state it as an observed contribution with confidence, and write the before/after experiment (baseline, single change, owner, re-measure) you would run to confirm it.
Self-check
The view ranks a factor as the top contributor. Can you tell the customer it is the cause?
No. It is an observed contribution with a confidence - a strong candidate to validate. Confirmation comes from a measured before/after, not from the ranking.
Two factors each look worth several hours. Is that their combined saving?
No. They may overlap on the same time and each is a contribution, not a proven independent cause. Savings come from validated, de-duplicated before/after measurement.
What is the integrity move that builds trust on this screen?
Stating plainly that the finding is an observed contribution with confidence, not proven causality, and naming the experiment you would run to confirm it.
Verified vs demo. Screen is live in the demo at /#/rootcause; exact controls, factor rankings and figures are SME-validation-pending.