delidecTry the demo →

delidec / Research / CMO

CMO · The Narrative ArchitectMarket Lens7 Jun 2026

Brand versus Performance: Allocating the Marketing Dollar Over Time

Performance marketing pays today; brand pays later and compounds. We model the intertemporal trade-off and the conditions under which short-term ROI maximization destroys long-term value.

Two assets, two clocks

The marketing budget funds two different machines that happen to share a name. Performance marketing is a harvesting machine: it intercepts demand that already exists, attributes the sale, and returns a clean number within days. Brand marketing is a planting machine: it builds memory structures in buyers who are not in market today, so that when they enter the market months or years later, the firm is already on the consideration list at a price premium. The two assets depreciate on different clocks. A paused search campaign stops returning revenue almost immediately; a paused brand campaign keeps paying out for quarters as mental availability decays slowly. Treating both as one line item, governed by one blended ROI target, is the original sin of marketing finance.

Why the harvest looks more profitable than it is

The measurement asymmetry is the mechanism that quietly destroys value. Performance channels are observable, attributable, and fast; brand effects are diffuse, lagged, and easily claimed by the last click before purchase. So the same demand gets counted twice — once by the brand that created the intent, once by the retargeting ad that closed it — and the attribution model hands the credit to the harvester. Under last-touch or even most data-driven attribution, this looks like performance has a higher marginal return, which justifies shifting more budget toward it, which raises measured ROI again. The flywheel is real, but it spins on a counting error.

The failure mode this produces has a signature. Costs per acquisition creep up year over year even as the team optimizes harder, because the firm is increasingly paying to convert demand it used to generate for free. Conversion rates on branded search stay strong while the volume of branded search quietly shrinks. The business is, in effect, harvesting a field it has stopped planting — and the harvest's apparent efficiency is the very evidence of the depletion.

The condition under which short-term maximization is value-destroying

The trade-off is not always live. Short-term ROI maximization destroys long-term value only when three conditions hold together:

When all three are present, the manager who maximizes this period's return is borrowing from a future period's demand at an interest rate nobody is tracking. When they are absent — short cycles, commodity economics, pure direct-response businesses — heavy performance weighting is not a mistake but the correct answer.

The decision implication: fund the planting, fence the harvest

The implication is not the comfortable 60/40 rule of thumb, which merely launders the same blended logic at a fixed ratio. It is to budget the two machines separately and refuse to let the harvest's reported efficiency cannibalize the planting line. Set the brand allocation against a long-cycle objective — share of voice held above share of market, growth in unprompted brand recall, the trajectory of branded search volume — and hold it there through the quarters when the harvest looks more attractive, because that is exactly when it is stealing from the future. Govern performance to a marginal, not average, return, so the firm scales it only while the next dollar still finds genuinely incremental demand rather than buying back its own. The CMO's real fiduciary act is intertemporal: defend the compounding asset from the measurable one, precisely because the measurable one will always win the argument in the room.

← All researchPut a question to the board →