Each channel has a diminishing-returns response curve — the classic saturating-exponential model used in advertising response analysis: the more you already spend on a channel, the smaller the extra conversions the next dollar buys.
Conversions(spend) = Cap * (1 - e^(-spend / τ))
Marginal return: dConversions/dSpend = (Cap / τ) * e^(-spend / τ)
Cap is the channel's ceiling of daily conversions as spend grows without limit; τ ("tau") is the spend scale at which the channel is about 63% saturated — a small τ means a channel saturates fast on a small budget, a large τ means it keeps paying off at high spend.
The equal-marginal-value rule from microeconomics says a fixed budget is spent optimally when the marginal return per dollar is equal across every channel — if one channel's next dollar buys more conversions than another's, money should move there first. Optimize Allocation implements this directly: it hands out the budget in small increments, always giving the next increment to whichever channel currently has the highest marginal return, until the whole budget is spent. The result is the spend split that maximizes total conversions for that budget.
- Total budget slider — the daily ad spend available to allocate across all four channels.
- Per-campaign sliders — manually override any channel's spend; the towers and curves update live.
- Optimize Allocation — runs the marginal-value algorithm above and animates the sliders to the optimal split.
- Split Evenly — a naive baseline (budget / 4 per channel) to compare against the optimized split.
Real-world relevance: this is the same principle media-buying teams and ad platforms' automated budget optimizers use to shift spend between campaigns and channels as returns diminish.