The situation
When trading performance moved unexpectedly, investigation was fragmented across teams and data sources. People repeatedly entered reactive investigation or war-room mode before they understood whether the change came from demand, conversion, price, supply and capacity, operations, or data quality.
The problem was not simply a lack of dashboards. It was the absence of a shared diagnostic path. Each team could see part of the system, but nobody had a consistent way to localise the change, compare it with a meaningful baseline, and decide where investigation should begin.
Organise signals around the questions
The work brought the relevant commercial and operational signals into one decision-support view. Signal selection began with the questions stakeholders needed answered: Is the change broad or local? Is traffic different, or is conversion moving? Has price or product mix shifted? Is supply constrained? Could the apparent movement be a data-quality problem?
Each measure used a comparison that made the change interpretable, rather than presenting an isolated number. The diagnostic sequence then moved from detection to localisation, likely cause and action. This created a common starting point for commercial, operational and data teams without pretending that a dashboard could replace judgement.
- Detect
- Localise
- Diagnose
- Act
Show where to look first
The executive view was intentionally selective. It separated a headline trading movement from the signals that could plausibly explain it, then highlighted the first investigation path rather than asking everyone to inspect everything at once.
Trading movement diagnostic
Illustrative data- DemandStableWithin expected range
- ConversionWatchSoft in one segment
- PriceStableMix-adjusted movement limited
- Supply / capacityPriorityConstraint localised to peak periods
- Data qualityClearCore checks passing
The recommendation
Start investigation where the signal diverges most clearly from its expected range, after first confirming that core data-quality checks pass. Give the relevant owner a specific operational question, while other teams monitor their guardrails instead of launching parallel, unfocused investigations.
The value came from agreeing which signals mattered, establishing useful comparisons, and giving teams a common order of operations. The tool made the first decision faster: who needs to investigate what, and what evidence would change that direction?
Outcome
The trading-health approach reduced emergency investigation and meeting effort by roughly 80%.
It demonstrates commercial analytics as decision support: identifying useful signals, communicating likely causes clearly, and helping senior stakeholders move from an unexpected result to a focused action.