Leading vs lagging indicators: measure what can still change

The short answer

A lagging indicator confirms the result after it happens; a leading indicator measures behavior or movement that predicts it early enough to act. Use one outcome metric, one conversion metric and one controllable input metric, then review whether the assumed relationship actually holds.

Build a three-level measurement chain

1. 1. Outcome Name the result that ultimately matters, such as retained customers, revenue or completed training sessions.
2. 2. Conversion Measure the nearest externally verified step that must happen before the outcome, such as activated trials or qualified proposals accepted.
3. 3. Input Track the controllable behavior most likely to create that step, such as completed onboarding sessions or decision-maker conversations.
4. 4. Assumption Write the causal belief connecting the three levels. This is a hypothesis, not a permanent truth.
5. 5. Threshold Decide in advance what result makes you continue, change or stop the input.

Do not optimize a proxy after it breaks

A leading metric is useful only while it predicts the desired result. Reply rate can rise while qualified meetings fall; hours worked can rise while delivery slows. When the relationship weakens, replace the proxy rather than defending it.

Keep missing data separate from zero. No recorded conversations does not prove that no conversations happened. It proves that the system cannot yet distinguish execution from missing measurement.

Questions

How many indicators should one goal have?

Usually three: one outcome, one conversion step and one controllable input.

How often should indicators change?

Only when evidence shows the metric no longer predicts the outcome or when the bottleneck moves.

Is activity a leading indicator?

Only when the activity has a demonstrated relationship with a useful result. Volume alone is not enough.