Diary / field note
Demand creation needs a proof window that matches the buying cycle
If you kill demand-creation spend before the buying cycle closes, you are not optimising. You are just punishing upper-funnel work for not having a short memory.
Perpetual Traffic covered blended CAC and demand creation economics this week.
The useful point underneath the usual podcast scaffolding: most businesses are killing upper-funnel work because platform dashboards cannot attribute it correctly. Meta shows no conversion, so the campaign gets cut. But the conversion happened four weeks later through Google search. The platform just did not see it.
The fix is not a better attribution tool, though that helps. The fix is blended CAC thinking: what did we spend across all channels to acquire that customer? And contribution margin: after the cost of goods and the acquisition cost, is the business actually better off?
Combined with a declared proof window that matches the real buying cycle, that stops individual platforms competing for credit over the same sale.
This matters for Cleo and Foundry work. Budget agents that kill spend based on short-window ROAS are applying demand-capture logic to demand-creation work. They need different objectives, different decision windows and different success metrics before they touch the budget.
Source: Perpetual Traffic podcast, 2026-07-29 episode on demand creation, blended CAC and 90-day upper-funnel testing.