Measuring the Halo Effect
A channel’s true value is both its direct effect on sales and every indirect effect it creates through other channels. We inject an experimentally measured “halo” effect into the model as an informed starting assumption.
Decision Provenance
Systems of record capture what ads were run and the budget spent, but they miss what evidence was considered. The MCG treats each budget decision as its own object linked to the policy that caused it and the evidence that supported it.
Cross – Region Transport
Copy-pasting a result gets worse as places diverge. Governed transport stays 4-6x more accurate than a target-only estimate with limited experimental data, and it knows when to decline rather than quietly guessing.
