Inspection dashboards tend to count activity: inspections performed this month, reports issued, readings taken. Activity metrics reward volume, and volume is not the goal. A program can hit every activity target while overdue vessels accumulate and findings age quietly in a backlog. The KPIs worth leadership attention measure program health: is the schedule honest, is the data trustworthy, and does work actually close?
Schedule health
- Overdue inspections, by consequence. The raw overdue count is noise if a lubricator sits next to a high-consequence vessel in the same number. Split by criticality. The trend on high-consequence overdues is the single most important line in the program, and the target is not zero this month; it is a credible downward slope with named owners.
- Schedule attainment. Of inspections due in the period, what share was completed in the period? This is the honesty check on the plan. Chronic attainment below plan means the schedule is aspiration, not commitment, and intervals are drifting in practice regardless of what the calculations say.
- Deferral rate and age. Deferrals are legitimate tools that become risk when they stack. Track how many active deferrals exist, how many are on their second or third extension, and whether each has a documented basis. Repeat deferrals on the same equipment are the program telling you something.
Data health
- Finding-to-report cycle time. Days from field capture to reviewed, approved report. This is where inspection value evaporates: a finding that takes six weeks to reach a decision-maker was, for six weeks, not a finding. It is also the metric most directly improved by capturing once in the field instead of transcribing.
- Re-inspection rate from data problems. Repeat visits caused by missing readings, wrong locations, or lost context, as distinct from repeats driven by findings. Each one is a technician-day plus access costs, and the trend measures whether capture quality is improving.
- CML coverage integrity. Share of scheduled CMLs actually read on completed inspections, and the count of readings landing without required context. Coverage gaps compound: a skipped CML this cycle weakens the rate calculation for the next decade.
Closure health
- Recommendation aging. Open recommendations by age band and consequence, with owners. Findings that never close are the program's quiet failure mode, and an aging chart makes the quiet visible. Watch especially the items that cross audit-interval boundaries with no status change.
- Audit demand response time. How long to produce, on request, a complete answer to a standard auditor question: the overdue list, a vessel's inspection basis, a recommendation's closure evidence. Measure it with an internal drill once a quarter. This number is a proxy for the entire records system.
Using them without gaming them
Every metric invites its own distortion, so pair them: attainment pairs with deferral rate, so the schedule cannot be "attained" by deferring it. Cycle time pairs with coverage integrity, so speed cannot come from skipping CMLs. Review the set monthly at the program level, and resist adding more than these eight until the first eight are trusted, because a forty-tile dashboard is where attention goes to die.
The precondition for all of it is structural: these numbers must fall out of the system of record as a byproduct of normal work. A KPI that requires a monthly spreadsheet expedition will be estimated, then fabricated, then abandoned.
That precondition is a fair test for any inspection platform, intelliSPEC included: ask to see these eight measures computed live from working records, not mocked in a slide. If the system runs the program, the program's health should be one query away.
