Green Dashboard, Red Reality
A government digital services programme reports 87% on-time delivery against its plan. A new assurance reviewer notices that the plan was re-baselined twice in the last year, each time moving milestone dates later, and that 'on-time' is measured against the most recent baseline. Two of the largest deliverables have been marked 'complete' when only the first of three releases shipped. Defect backlog has grown from 40 to 310 open items over six months, and average resolution time has doubled. The programme director says the dashboard is accurate because it follows the agreed methodology. Citizen service outages have risen from 2 to 9 per month.
1. Diagnose why the reported performance diverges from operational reality, and identify the measurement flaws.
Read the worked answer
The dashboard measures conformance to a moving plan, not delivery of value. Re-baselining twice without flagging it converts schedule slippage into apparent on-time performance; the metric is technically consistent but substantively misleading. Marking a multi-release deliverable 'complete' after one of three releases is a definition-of-done failure: progress is counted at the wrong granularity, so 87% overstates completion.
The defect trend is the strongest signal. A backlog growing from 40 to 310 while resolution time doubles indicates quality is deteriorating faster than it is being fixed, and the rise in outages from 2 to 9 per month confirms the effect is reaching citizens. The director's defence ('we follow the methodology') confuses procedural compliance with measurement validity. A metric that cannot distinguish improvement from deterioration is not controlling anything.
2. Propose a revised measurement set and governance response that would surface the true position without punishing honest reporting.
Read the worked answer
Revise the measures to be outcome- and quality-anchored: report schedule variance against the original baseline alongside the current one, so re-baselining is visible; define completion by release-level acceptance criteria, not by deliverable label; and add defect inflow versus outflow, ageing, and citizen-facing outage rate as primary indicators with thresholds.
Governance response: require any re-baseline to be approved and disclosed with its reason, so it is a decision rather than a quiet adjustment. Separate the reporting conversation from the blame conversation, because the current dynamic rewards optimistic status. The director should present the revised position with the trend and a recovery plan, and the assurance function should confirm the new definitions independently. The aim is a dashboard that would have shown red months ago and that people are not afraid to populate honestly.