Lesson 2 of 5 · 9 min · Metrics that mean something

Leading and lagging metrics

Covers: Standard analytics practice with SQL and pandas examples. Updated 13 Nov 2026. Part of the Data Analyst roadmap.

What you will learn

  • The difference between lagging and leading metrics
  • Examples from sales, product, support and education
  • How to find leading indicators using cohorts and comparisons
  • How to combine both kinds in a dashboard
  • The traps: weak predictors, gaming and treating correlation as cause

In simple words

When you drive, the speedometer tells you how fast you are going now, but the road ahead tells you what will happen next. A lagging metric is like looking in the rear-view mirror: it reports a result that has already happened, such as last month's revenue. A leading metric is like looking ahead: it moves first and hints at what the result will be, such as the number of demos booked this week.

Words to know

  • Lagging metric: a measure of an outcome that has already happened.
  • Leading metric: an early measure that tends to predict a later outcome.
  • Input metric: something a team directly does or controls, such as calls made.
  • Output metric: the result that follows, such as revenue.

The two kinds

Lagging metrics measure outcomes: revenue, profit, churned customers, exam pass rate, employee attrition. They matter because they are the goal, but they arrive too late to change. By the time churn rises, the customers have gone.

Leading metrics measure early behaviour that predicts the outcome: trial users who complete setup, demos booked, practice tests completed in the first week, resumes sent this week, support tickets about a recurring bug. You can act on them while there is still time.

Area Lagging (the result) Leading (the early signal)
Sales monthly revenue, deals won qualified leads, demos booked, proposals sent
Subscription product churn, renewals first-week activity, features used, login frequency
Education exam pass rate, placement rate weekly practice hours, mock test scores, attendance
Job seeker offers received applications sent, interviews attended, mock interview scores
Support customer satisfaction score first response time, backlog size, reopen rate
Operations deliveries on time (monthly) orders late today, stock-outs, driver availability
Hiring cost per hire, retention at 1 year number of qualified applicants, time to first interview

Most businesses need both: the lagging metric says whether the strategy is working; the leading metrics say whether you are on track and what to change.

Input, output and the chain between them

Think of a chain: inputs (actions you control) lead to outputs (early results) that lead to outcomes (the business goal).

Calls made -> Meetings booked -> Proposals sent -> Deals won -> Revenue
(input)       (leading)          (leading)        (lagging)    (lagging)

Teams control inputs directly; outcomes they only influence. Managing by the lagging end alone ("hit the revenue number") gives no guidance on Monday morning, so teams also track the earlier steps.

Finding leading indicators in your data

A leading metric must be early, measurable now and genuinely related to the outcome. Do not guess; test with your own data using cohorts.

  1. Pick the outcome, such as "customer is still paying after 90 days."
  2. List candidate early behaviours in the first 7 days: number of sessions, whether feature X was used, whether they invited a teammate.
  3. Compare outcomes between users who did and did not do each behaviour:
SELECT used_feature_x_in_week1,
       COUNT(*)                               AS users,
       ROUND(100.0 * AVG(retained_90d::int), 1) AS retention_pct
FROM user_summary
GROUP BY 1;

Suppose users who used feature X in week one have 62 percent 90-day retention versus 31 percent for those who did not. Feature X in week one is a promising leading indicator.

  1. Check it is not just a proxy for engaged users in general: compare within segments (channel, plan, device), and look at how the effect changes with the amount of usage (does 3 uses beat 1?).
  2. Set a threshold where the effect jumps ("at least 3 completed tests in 14 days"), which becomes your activation definition.
  3. Test it: change something to raise the leading metric (a better first-week guide) and see whether the outcome follows. This is the only way to learn whether it is a lever or only a sign.

A balanced view

A useful dashboard shows a few of each, with targets:

  • 1 to 3 lagging outcomes (monthly revenue, 90-day retention).
  • 3 to 5 leading indicators that predict them (weekly activations, first-week usage, pipeline).
  • 1 to 2 guardrails (refund rate, complaints).

Put leading metrics on a short cycle (daily or weekly) and lagging ones on a longer cycle (monthly or quarterly). Review them in the right order: when a lagging number surprises you, look back at which leading metric should have warned you, and fix your indicators.

Traps

  • A leading metric that does not predict. Many things correlate with success only because successful customers do everything more (they log in more, click more). Verify with comparisons and experiments.
  • Gaming. If you reward "calls made", people make short pointless calls. Pair activity metrics with quality ones ("meetings booked per 100 calls").
  • Acting too soon on noise. Daily leading metrics bounce around; look at a weekly average or a trend, and mind sample sizes (see the statistics course).
  • Confusing lagging with useless. Revenue lags, but it is the truth. Leading metrics are tools for steering; lagging metrics are the score.
  • Different lags. A leading metric may predict churn 30 days ahead for one product and 6 months for another. Find the lag from the data.
  • Seasonality. January signups behave differently from August signups; compare like with like (the cohort lesson).

Common mistakes

  • Reviewing only lagging metrics and discovering problems when it is too late.
  • Assuming every early behaviour predicts the outcome.
  • Rewarding activity numbers without quality.
  • Treating a leading indicator as a cause without testing.
  • Mixing time scales on one chart.

Try it yourself

  1. For a business you know, list five lagging and five leading metrics and draw the chain between them.
  2. Using any dataset with a later outcome (repeat purchase, churn), compare the outcome between users with and without an early behaviour.
  3. Define an activation threshold from that comparison and write the SQL for it.
  4. Pick a team metric that could be gamed, and add a quality metric to pair with it.

Check yourself

What is the difference between leading and lagging metrics?

Lagging metrics report results that have already happened; leading metrics are early signals that tend to predict those results.

Why is revenue a lagging metric?

It records the outcome of earlier activities (leads, demos, deals), so by the time it changes it is too late to influence that period.

How can you check that an early behaviour is a genuine leading indicator?

Compare later outcomes between those who did and did not do it, check within segments, look for a dose effect, and ideally run an experiment to raise it.

Why can reward systems based on a leading metric backfire?

People optimise the number itself, for example making many low-quality calls, so you need quality or guardrail metrics alongside.

Interview questions

Give an example of a leading and a lagging metric for a subscription product.

A leading metric is the share of new users who complete setup and use a key feature in their first week; a lagging metric is 90-day retention or monthly churn.

How would you find which early behaviours predict retention?

Define retention at a fixed point, compute it for users with and without each early behaviour in a cohort, control for segments, check the size of the effect and the lag, then validate with an experiment.

Why do you need both kinds of metrics on a dashboard?

Lagging metrics show whether the goal is being achieved, while leading metrics provide early warning and tell the team what to do before the lagging numbers move.

Summary

  • Lagging metrics measure results; leading metrics predict them and can be acted on early.
  • Teams control inputs, influence outputs and are judged on outcomes.
  • Find leading indicators by comparing later outcomes across early behaviours, then test them.
  • Balance with guardrails, beware gaming, noise and weak predictors.

Going deeper

See how much of this you can explain

A free mock interview asks you new questions on Metrics, KPIs and business analysis for Data Analyst and scores every answer, with what to fix.