Budget Allocator

Make the next advertising dollar a deliberate choice.

Role: Product design & developmentStack: TypeScript · Response modelling · Constrained optimisation

Budget Allocator channel comparison: average ROAS beside estimated marginal return

A planning tool that fits response curves to weekly channel data and compares budget allocations using estimated marginal returns.

The problem

A channel with a strong average return does not necessarily offer the best return on the next dollar. As spend increases, returns can flatten. Budget Allocator makes that trade-off visible before someone changes the media plan.

Start with the observed data

Import weekly channel spend and revenue from a CSV. Compare fitted response curves, including logarithmic, power, and Hill models, to understand how the historical relationship changes across spend levels.

Allocate within real limits

Set a total budget and minimum or maximum spend for each channel. The allocation solver compares estimated marginal returns while keeping recommendations inside the observed spend range. It does not extrapolate beyond the evidence supplied.

Inspect the assumptions

Compare the modelled allocations with the current plan and review channel saturation. Constraints and uncertainty stay visible so the output can inform a planning conversation instead of becoming an unexplained recommendation.

What the model can tell us

These curves describe historical associations between spend and revenue. They do not establish causal lift, guarantee future returns, or replace experiments. The tool is a way to explore a decision under explicit assumptions.