Leading and Lagging Indicators

A lagging indicator measures an outcome after it has happened, and a leading indicator measures something earlier that is expected to produce it. Revenue, renewals and churn are lagging: they are the things you actually care about, they are unambiguous, and by the time they move the decision that caused them was made months ago. Activation rates, usage depth, support contact rates and pipeline coverage are leading: they move quickly enough to act on and none of them is the thing you want. The trade is always the same. Lagging indicators are trustworthy and late; leading indicators are timely and provisional, and treating either as though it were the other is where the trouble starts. Two failures are common. The first is running a business entirely on lagging numbers, which produces a management rhythm where every correction arrives a quarter after it would have helped, and where nobody can tell whether this quarter's work is going well until it is over. The second is the opposite: promoting a leading indicator to a target and forgetting that its whole claim to relevance is a correlation that was measured once, possibly on a different product, possibly on a different set of customers. A leading indicator is a bet on a causal path, and bets get re-examined. The useful discipline is to keep them in pairs. For each leading indicator write down the lagging outcome it is supposed to predict, and check the relationship periodically rather than assuming it holds — annually is often enough, and after any significant change to the product or the mix of customers is better. When the pair comes apart, the leading indicator is the one to retire, not the outcome. A related trap is treating a leading indicator as an early version of the same number. It is not the outcome arriving sooner but a different measurement with its own errors, which is why a leading indicator can be up all quarter and the lagging one still land flat without either being wrong.