How to read an MMM saturation curve

A response curve is just a function from spend to revenue. Per channel, you fit one, and then the budget conversation becomes a math problem instead of a politics problem.

Three parameters define the curve we use, and once you can read them, you can read the whole story:

α (alpha) — the ceiling

How much revenue this channel could deliver at infinite spend. If α is small relative to your goals, the channel will never be the answer to your growth question — no matter what you do.

K (kappa) — the bend

The spend level at which you've captured half of α. K tells you where saturation starts to bite. If you're above K, the next dollar is buying less than the previous dollar. If you're below K, you're leaving easy revenue on the table.

γ (gamma) — the steepness

How sharply the curve bends. High γ means the channel goes from "working" to "saturated" over a narrow spend range — the cliff is real. Low γ means the bend is gradual; you have room to maneuver.

The single most useful number isn't α

It's marginal ROAS at your current spend— the slope of the curve at the point where you're standing. Every dollar of budget moves toward whichever channel has the highest marginal ROAS. The math is unambiguous; the only ambiguity is in the curve fits.

That's why we report fit quality alongside the curves. If the data is too sparse to identify γ, we say so out loud rather than letting an over-confident allocator push you into a corner.