Concept/Operations and Risk/No. 0556
Leading vs. Lagging Indicator
Leading and lagging indicators are measures that signal future conditions or record past results. Used in economics, safety and performance measurement, the distinction depends on the outcome tracked. A leading indicator needs evidence of predictive value, not just an early report.
Also called Leading and Lagging Indicators · Leading vs. Lagging Indicators
- Evidence
- Well established
- Read
- 6 min
- Links
- 9 connections
- Useful when
- Forecasting · Metrics and incentives · Reading data and statistics · Risk and safety · Running projects
01You've seen this when…
- at work
Customer cancellations stay low, so the monthly report looks healthy. Meanwhile, the oldest unanswered support ticket goes from two days old to nine.
- in life
Your final exam is three weeks away. Hours spent studying look reassuring, but the mistakes on your practice tests show which topics still need work.
- out in the world
A utility reports fewer water-main breaks last year. Its engineers also want to know how many high-risk pipes are overdue for inspection.
02The idea
A result tells you what happened. A warning gives you a chance to change what happens next. A useful dashboard makes room for both.
A lagging indicator records an outcome already realized, from customers lost and injuries sustained to revenue earned. A leading indicator aims to signal a future condition: support tickets may pile up, or deteriorating equipment and falling orders may warn of trouble ahead.
The distinction is about the relationship to a particular outcome, not just when a report arrives. A cancellation count is lagging even if it updates instantly. An equipment inspection can provide advance warning even if its report takes a week to reach you.
Leading indicators also come in different forms. Some track actions expected to improve results, such as completing safety-critical maintenance. Others track emerging conditions, such as rising temperatures inside a machine. Neither earns the label merely by appearing earlier on a dashboard.
To use one responsibly, finish this sentence: This measure helps predict this outcome over this time horizon. Then test that relationship. Study hours might predict exam performance poorly; scores on representative practice questions might do better. A plausible story is a starting point for validation. Testing must follow.
03Why it matters
Waiting for the final result can leave you with excellent records and no remaining choices. By the time customers cancel, the service problem may have been building for months. This is the practical cost of feedback delay.
But watching only activity creates the opposite problem: everyone looks busy while the outcome fails to improve.
- Advance warning creates room to act. A warning matters when it arrives early enough to investigate, repair, redirect resources, or stop.
- Lead time determines usefulness. Predicting tomorrow’s failure five minutes beforehand may be accurate but useless.
- Outcomes check whether the effort worked. Training completed and inspections performed measure preventive work. Lagging measures help reveal whether the proposed preventive actions deliver safer operations.
- The pair separates execution from effectiveness. If maintenance completion falls, investigate whether the work is getting done. If completion stays high but failures rise, investigate whether the maintenance itself addresses the right risks.
The strongest arrangement connects an outcome to a small set of warnings and preventive actions. A dashboard with twenty activity counts still needs an outcome scoreboard.
04A worked example
In March 2005, an explosion at BP’s Texas City refinery killed 15 people. The independent panel that subsequently reviewed BP’s U.S. refineries found serious weaknesses in process-safety management, including performance measurement. It warned against treating progress on personal injury rates as evidence that major industrial hazards were under control.
What it looks like Fewer ordinary workplace injuries suggest that the refinery is becoming safer overall. The headline safety number appears to support confidence in the operation.
What’s actually going on Ordinary workplace injury rates include injuries from slips and falls. They give a limited picture of the risk of a large release of flammable material, and a refinery can improve those rates while remaining vulnerable to a release. The injury rates were lagging indicators. Using that narrow measure to judge a different risk was a problem of construct validity.
Catastrophic accidents are rare enough that a quiet period gives limited reassurance. Without measures of whether critical safeguards are working, the final outcome may be the first unmistakable signal of failure.
What would have helped The panel recommended effective leading and lagging indicators for process safety. That means tracking relevant outcomes, such as releases of hazardous material, alongside warnings about degraded safeguards. Overdue safety-critical inspections or unresolved hazardous conditions are possible candidates whose predictive value needs testing, even when they sound sensible. Each should connect to a specific hazard and response, as in bow-tie analysis. A dashboard can make deteriorating controls visible before an accident. Prevention remains uncertain even with a dashboard.
05Where people trip up
- They count effort as progress. Calls made, lessons watched, and audits completed describe activity. They become useful leading indicators only when they tell you something about the desired outcome. Include quality: an inspection that misses the defect is still a completed inspection.
- They confuse prediction with causation. A measure’s ability to forecast a result is separate from its power to change that result. Umbrella sales may signal rainy conditions; selling more umbrellas leaves rainfall unchanged. Before acting on a relationship, examine alternative explanations and confounding.
- They skip the prediction test. State the expected direction, outcome, and time horizon before checking results. Compare the signal with a simple baseline, such as the outcome’s recent history. Test on later observations to check whether the relationship holds beyond the records used to select it. Check missed warnings and false alarms as well as successful predictions.
- They reward the warning until it becomes unreliable. Paying for completed checks can produce hurried checks. Once a leading indicator becomes a target, Goodhart’s law applies just as it does to an outcome measure. Audit both the count and the underlying work.
- They respond to every wiggle. More timely measures are often noisier. Decide how large or persistent a change must be before it triggers investigation. Otherwise a poor signal-to-noise ratio turns advance warning into constant distraction.
- They collect warnings without assigning responses. For each important signal, name who checks it and what they can do. A red dashboard reports a problem; preventive use requires an owner and an available action.
06When it isn’t one or the other
A number’s role as a leading or lagging indicator depends on the outcome being considered. Today’s support backlog is an outcome of earlier staffing and demand. It may also warn of future cancellations. A practice-test score records learning so far while helping predict a later exam result.
Economists also use coincident indicators: measures that move alongside current conditions rather than clearly before or after them. Not everything needs to fit into two buckets.
A useful leading indicator can track conditions beyond your direct control. An external warning can still help you prepare. The key questions are whether it adds dependable information and whether that information arrives while you still have choices.
07Roots
Wesley Mitchell and Arthur Burns worked at the National Bureau of Economic Research in the 1930s. They were trying to recognize economic recovery before it became obvious. After the Great Depression, their work had practical stakes. Businesses and governments needed to distinguish an approaching revival from another temporary improvement.
They compared economic series around earlier business-cycle turning points. Orders, production, employment, and other measures turned at different times. Some moved earlier; others followed. Their 1938 report, Statistical Indicators of Cyclical Revivals, helped establish the systematic search for advance signals. The economic tradition developed the familiar grouping that places coincident indicators between leading and lagging ones.
Safety management and performance measurement later applied the same timing distinction to different problems. Accident totals and financial results describe what an organization has already experienced. Measures of safeguards, developing hazards, and operating activity might offer earlier opportunities to intervene. The question remained the same: which observations help us predict what is coming?
08How solid is this?
The distinction is well established in economics, safety management, and performance measurement. Whether a particular indicator reliably predicts an outcome requires separate evidence, even if it is early, actionable, or easy to count.
09Connections
- Helps counter Feedback Delay
- Part ofBalanced Scorecard, Construct Validity
- See also Confounding, Signal-to-Noise Ratio, Bow-Tie Analysis, Goodhart’s Law, Normalization of Deviance, Scenario Planning
10Origin and sources
The economic tradition was developed systematically by Wesley C. Mitchell and Arthur F. Burns at the National Bureau of Economic Research, including their 1938 work on indicators of cyclical revival. Safety management and organizational performance measurement later adopted the distinction.
- [1]Mitchell, W. C., & Burns, A. F. (1938). Statistical Indicators of Cyclical Revivals. National Bureau of Economic Research.
- [2]Burns, A. F., & Mitchell, W. C. (1946). Measuring Business Cycles. National Bureau of Economic Research.
- [3]BP U.S. Refineries Independent Safety Review Panel. (2007). The Report of the BP U.S. Refineries Independent Safety Review Panel.
- [4]Occupational Safety and Health Administration. (2019). Using Leading Indicators to Improve Safety and Health Outcomes.
Suggest an edit· Updated 2026-10-02