Two kinds of metric, both needed
Lagging metrics measure what already happened: tolerance breaches, incident duration, regulatory findings. They tell the board whether resilience held. Leading metrics measure capability before the incident: plans exercised, restore tests passed, crisis roles trained, supplier concentration. They tell the board whether resilience will hold. A dashboard with only lagging metrics is a rear-view mirror; with only leading ones it is a promise.
The ten metrics
| # | Metric | Type | Green | Red |
|---|---|---|---|---|
| M-01 | Impact tolerance breaches per quarter | Lagging | 0 | 2 or more |
| M-02 | Prioritised activities with a plan exercised in 12 months | Leading | 90-100 % | below 70 % |
| M-03 | Critical services where tested recovery time exceeds RTO | Leading | 0 | 3 or more |
| M-04 | Critical systems with a successful restore test in 6 months | Leading | 100 % | below 90 % |
| M-05 | Critical services depending on a single supplier | Leading | 0 | 3 or more |
| M-06 | Corrective actions overdue | Leading | below 10 % | above 25 % |
| M-07 | Crisis roles with trained primary and deputy | Leading | 100 % | below 80 % |
| M-08 | Mean time to detect, minutes | Lagging | below 15 | above 60 |
| M-09 | Mean time to activate the plan, minutes | Lagging | below 30 | above 60 |
| M-10 | Open regulatory findings | Lagging | 0 | 4 or more |
Thresholds in the table are starting points for a mid-size organisation in the Gulf; the file keeps them in editable columns. The rule that matters more than the numbers: every red threshold has a named reaction, otherwise the metric is statistics.
How to measure each metric
A threshold is useless until the number behind it is defined the same way every quarter. Below is the calculation for each of the ten: what goes above the line, what goes below it, over what period, and which log the figures come from. Nothing here needs new tooling.
| # | How it is calculated | Period | Source |
|---|---|---|---|
| M-01 | Count of occasions when a critical service stayed disrupted beyond its impact tolerance | Quarter | Incident log |
| M-02 | Prioritised activities exercised at least once, divided by all prioritised activities | Rolling 12 months | Exercise log |
| M-03 | Count of critical services where the last tested recovery time exceeded the agreed RTO | Last test per service | Exercise and test reports |
| M-04 | Critical systems with a successful restore from backup, divided by all critical systems | Rolling 6 months | IT restore test reports |
| M-05 | Count of critical services that depend on one supplier with no substitute inside the tolerance | Quarter | Supplier register |
| M-06 | Corrective actions past their due date, divided by all open actions | Month end | Action log |
| M-07 | Crisis roles with both a trained primary and a trained deputy, divided by all crisis roles | Rolling 12 months | Training records |
| M-08 | Sum of minutes from event start to detection, divided by the number of incidents | Quarter | Incident log |
| M-09 | Sum of minutes from detection to plan activation, divided by the number of activations | Quarter | Incident log |
| M-10 | Count of regulatory findings not yet closed at the reporting date | Quarter | Findings register |
Three rules keep the measurement honest. Count the event, not the ticket: one disruption that spawned forty tickets is one breach. Measure from when the event started, not from when someone opened the log, otherwise detection time flatters itself. Keep the denominator fixed for the quarter, because a scope that grows mid-period turns a worsening ratio into an improving one.
A metric with no named owner and no agreed reaction to red is a statistic. Each of the ten gets one person who reports it and one action that happens when it crosses the threshold.
Download the template
Metrics sheet with ten indicators, definitions, data sources, frequency, owners and thresholds; a one-page board view that pulls current values and statuses. Xlsx.
Download xlsx (operational-resilience-metrics-template.xlsx)
Status for M-01 is calculated by formula; for percentage metrics enter the status or adapt the formula to your thresholds. No registration, no forms. Need it adapted to your organisation or a full programme: see how we work.
Where the data comes from
Nothing in the ten requires new data collection. Incident and exercise logs give M-01, M-02, M-08 and M-09; IT test reports give M-03 and M-04; the supplier register gives M-05; the corrective action log gives M-06; training records give M-07; the findings register gives M-10. If one of these logs does not exist, that absence is the first finding, and the internal audit checklist will record it.
How to present them
One page, ten rows, three colours, trend arrows against last quarter. The board does not need the definitions; it needs to see which rows are red and what is being done. The structure of the full report is on the board reporting page, with a template.
Frequently asked questions
What are operational resilience metrics?
Measures that show whether an organisation can keep delivering critical services through disruption: lagging ones such as tolerance breaches and time to detect, and leading ones such as plans exercised, restore tests passed and crisis roles trained. The ten on this page cover both kinds.
How many resilience metrics should a board see?
Ten or fewer on one page, each with a threshold and a trend. More becomes an operational dashboard that belongs to management, not the board.
What is an impact tolerance?
The maximum level of disruption of a critical service the organisation is prepared to accept, expressed in time or volume. Breaches of tolerance are the single most important lagging metric and the one regulators ask about first.
How often should the metrics be updated?
Monthly for operational owners, quarterly for the board. Leading metrics change slowly; lagging ones are reported as incidents occur.