Compounding needs a measurement window longer than a month

Compounding needs a measurement window longer than a month

You cannot see compounding on a monthly dashboard. Month to month, a compounding loop and a flat one look almost identical, because the gains per period are small by design. Judge by the month and you will kill your best assets for being unexciting and chase your worst ones for being loud. The measurement has to match the timescale of the effect, or it lies to you.

So the first move is to measure the rate, not the level. A flat revenue line that is growing 4 percent month on month is a compounding line in disguise, and a dashboard that only shows the level hides that completely. Track the growth rate, the retention rate, the cost-per-outcome trend over quarters, the share of new customers coming from existing ones. These rates are where compounding shows up; the absolute numbers lag behind by months.

Cohorts are the right lens. Watching a single blended number averages your improving new customers together with your older ones and washes out the signal. Splitting by cohort, the customers who joined in January versus March versus June, lets you see whether each new group retains better, refers more, and costs less than the last. That trend across cohorts is the truest picture of whether your loops are actually compounding.

Set the review cadence to match. A monthly check is for operations, keeping the machine running. The compounding question, are the loops bending the curve, belongs in a quarterly review where you can see across enough time for the small gains to add up to a visible trend. Asking "is it compounding" every Monday is like weighing yourself every hour: the noise drowns the signal and you make bad decisions on it.

The discipline this buys you is the confidence to hold. When you can see, in the cohort data, that each new group is a little better than the last, you can sit calmly through a flat-looking month because the rate tells you the level will follow. Without that view, every soft month is a crisis, and crises are where compounding goes to die.

INTERVIEW EWOUD: What do you actually look at to know whether something is compounding rather than just ticking along? Name the specific rates or cohort views, and a moment they told you the truth a monthly number was hiding.