Work Output: Root cause
Objective. Open root cause on a specific output shortfall and read the surfaced causes as candidates to validate - correlations and leads, not proven causality.
Root cause answers 'why did this output shortfall happen, and what should I check first?' It connects a shortfall to candidate explanations so an investigation starts with leads instead of a blank page - but a listed cause is a hypothesis, not a finding.
See it in the demo
What this screen answers
A CSM opens root cause after spotting a shortfall on another screen to get a shortlist of where to look - mix, complexity, backlog, source gaps, quality or effort. Sales uses it to show that the platform moves from 'what slipped' to 'why, probably' with evidence attached. The customer question is 'where do we direct the investigation first?'
The task
Start from a specific, scoped shortfall - one process or team, one period - so the causes are about something real. Read the candidate causes with their supporting evidence and rank them by plausibility, then validate the top one against the underlying transactions or quality data before you accept it. Resist attributing the shortfall to effort until mix, complexity, backlog and source coverage have been ruled out, because those are the usual true culprits.
How to read it
The surfaced causes are candidates and correlations, not proven causality - this is the honest core of the screen. A cause appearing high on the list means 'investigate this', not 'this is why'. The classic misread is lifting the top candidate into a conclusion and putting it in a customer deck. Treat every listed cause as a hypothesis to test with the evidence underneath it, and remember that two plausible causes can both be partly right, so you cannot simply pick one and move on.
Honest limits
Root cause does not prove causality and cannot replace a human checking the evidence. It depends on the same source coverage and definitions as everything else - a gap there produces a confident-looking but wrong candidate. Confirm with product how causes are derived, and present any cause to a customer as a validated finding only after you have checked it, never as the screen's verdict.
Interpretation check
Surfaced causes are candidates and correlations to validate, not proven causality; the top candidate is where to look first, never a conclusion to put in front of a customer.
Practical task
Take one scoped shortfall, list the candidate causes, validate the most plausible against the underlying transactions or quality data, and write a one-line finding distinguishing the validated cause from the unverified candidates.
Self-check
The top candidate cause is 'lower effort'. Do you report it?
No - validate it against the evidence first, and rule out mix, complexity, backlog and source gaps, which are the usual true causes. A high-ranked candidate is a lead, not a finding.
Why start from a scoped, specific shortfall?
Causes are only meaningful about a real, bounded event. Run root cause on a vague average and you get candidates for nothing in particular.
Can you pick the single top cause and move on?
Not safely - two causes can both be partly right, and the list is correlational. You validate with the underlying data before concluding.
Verified vs demo. Screen is live in the demo at https://work-output.demo.timechamp.io/#/rootcause; exact controls, labels and on-screen figures are SME-validation-pending.