A lagging indicator reports what has already happened. A leading indicator moves before the result does. A team deciding what to watch needs both, because each one answers a question the other cannot. A strategy committed under uncertainty is a bet, and the team has to find out early which way the bet is going. That holds whichever posture the team took, whether it set out to shape the market or only to reserve the right to play. The distinction between the two came out of economics. Wesley Mitchell and Arthur Burns published a list of statistical series for the National Bureau of Economic Research in May 1938, and they sorted the series into three groups. Some turned before a recovery, some turned with it and some turned after it.
Robert Kaplan and David Norton carried the idea into management measurement in Harvard Business Review in 1992. They argued that financial measures report the results of actions already taken, so a company watching only those measures is reading its own history, and measures of what drives future performance have to sit beside the financial ones.
The first two sections below cover what a lagging indicator confirms and what a leading indicator warns about. The third shows how an indicator is built backwards from the assumption it tests. After that come the lag between a leading indicator and its lagging pair, and the indicators that move without meaning anything. The page ends with the rhythm for reading both sets.
Lagging indicators and what they confirm
A lagging indicator measures a result the organisation wanted. Revenue, retention at twelve months, gross margin and market share are all lagging, because each one records what customers did long after the decisions that influenced them.
A lagging indicator is the honest test of a strategy, because the strategy existed to move exactly these numbers. The cost is time. Retention at twelve months takes twelve months to read, so a strategy waiting for it has spent a year before it learns anything. A team measuring only lagging indicators is steering by a result it can no longer change.
Leading indicators and what they warn about
A leading indicator measures behaviour earlier in the same chain of cause. If the lagging indicator is retention at twelve months, the leading indicator might be the share of new accounts that import their own data in the first week. Importing data is what makes a team's own work depend on the product, so an account that has done it has a great deal more to unpick before it can leave.
A leading indicator is a claim about cause. The claim is the part teams skip. Most indicators earn their place on a dashboard because they update weekly and move visibly, and nobody has argued that they sit upstream of the result. A weekly active user count moves for reasons that have nothing to do with retention. Watching it closely therefore gives a team confidence and no warning.
Building an indicator from the assumption it tests
The way to stop an indicator giving confidence and no warning is to build it backwards from an assumption. A strategy's assumption list already names the claims the whole strategy rests on. Each load bearing assumption then earns one measure, and that measure is the one that would move first if the claim behind the assumption turned out to be false.
| The load bearing assumption | The leading indicator built from it | The lagging indicator it ends in |
|---|---|---|
| Operations managers will adopt with no sales conversation | Share of new accounts reaching a first report inside seven days | Revenue from self serve accounts as a share of new revenue |
| Teams that import their own data stay | Share of new accounts importing data in week one | Retention at twelve months |
| The price holds at the top tier | Share of deals closing at list price | Average revenue per account |
| Partners will sell the product without help | Deals a partner closes alone each quarter | Share of new revenue arriving through partners |
Each row reads in one direction. If the leading indicator fails to move, the assumption behind it is in question, and the team knows which choice in the strategy has lost its support. An indicator with no assumption behind it cannot do any of that. A number that falls with no claim attached starts an argument about the number itself.
The lag between a leading indicator and its lagging pair
Every pair in that table carries a gap in time, and naming the gap is part of choosing the pair. If data import in week one leads retention at twelve months, the team is reading a signal eleven months ahead of the result. That lead is what makes the pair worth having, and it is also where the risk sits. A leading indicator running that far ahead depends on a longer chain of cause, so more can happen along the way to break the link between the early signal and the result it was supposed to predict.
The practical rule is to state the expected lag when the indicator is chosen, along with how big a move would count as a signal. An indicator with no expected lag and no threshold turns every review into a debate about whether the number has moved enough yet. Settling that debate is what the indicator was for.
Indicators that move without carrying meaning
A number can move for reasons that have nothing to do with the strategy. Seasonality moves it, a marketing campaign moves it and a change to the instrumentation moves it. If nobody has written down what else could produce the movement, the team reads every rise as evidence and every fall as noise. The direction depends on what the team was hoping for.
The protection is a baseline and a stated threshold, both agreed before the number is read. If a leading indicator is expected to reach a quarter of new accounts inside a quarter, a reading of eighteen per cent is a miss worth acting on, and a reading of twenty four per cent is noise near the line. Both readings tell the team something, because somebody set the expectation in advance. A chart with no expectation against it tells the team nothing.
Reviewing indicators on a rhythm
Indicators chosen well still need a meeting. A monthly reading of the leading set and a quarterly reading of the lagging set give the team a rhythm. That rhythm is what turns a change of direction into a decision somebody took on a date. Without it, the change arrives as a panic in the week the annual number lands.
Watching a strategy costs far less than running one. A set of indicators tells a team whether one product's bet is working. An organisation running several products against one strategy has a further question that no indicator settles, and that question is how much of the money and how many of the people each product should have.