Ad Spend Calculator

How the figure is built

Ad spend required for a revenue target is target revenue ÷ target ROAS. For an order target it is orders × cost per acquisition.

Both are the same identity rearranged, and which form is useful depends on whether the objective is stated in revenue or in customers.

The figure only means anything alongside the margin, since spend that hits a revenue target at a losing ROAS has succeeded at the wrong thing.

Putting numbers to it

A $100,000 monthly revenue target at a 2.5 ROAS requires $40,000 of ad spend.

At a 55% contribution margin that revenue carries $55,000 of contribution, leaving $15,000 after the spend.

The same target at a 2.0 ROAS requires $50,000 of spend and leaves $5,000. At 1.82 it leaves nothing.

So the revenue target is achievable at several spend levels and only some of them are worth achieving, which is why a revenue goal without a ROAS constraint is an instruction to lose money efficiently.

Where the figure deceives

It assumes ROAS holds as spend rises, and it does not. Additional budget reaches progressively less responsive audiences, so the marginal ROAS falls even when the average looks stable.

A plan built on constant ROAS at triple the current spend is the single most common failure in advertising forecasting.

Acting on it

Model spend increases in steps with declining marginal ROAS rather than as a straight multiple. If the current spend returns 2.5, assume the next tranche returns 2.1 and the one after 1.8, and find where the marginal return crosses break-even.

Then set the budget where marginal contribution is still positive, not where average ROAS still looks acceptable.

Marginal versus average return

An account reporting a 2.5 average ROAS at break-even 1.82 is comfortably profitable on average and may be losing money on its last few thousand dollars of spend.

The way to see it is to look at the incremental result of the most recent budget increase: additional revenue divided by additional spend. That figure is nearly always well below the average and it is the one that decides whether to spend more.

Accounts managed on average ROAS therefore tend to overspend, because the average conceals the marginal. Running the calculation on increments rather than totals is a small analytical change that frequently finds a meaningful share of spend that is producing nothing.

Spend increases take time to reach steady performance, because algorithms re-enter learning and audiences take time to respond. A budget raised on the first of the month will not produce its eventual performance until well into it.

Planning increases a few weeks ahead of when the volume is needed avoids the pattern of raising budget for a peak and getting the improved performance after the peak has passed.

The same lead time applies to reductions: cutting budget sharply produces a performance dip before the saving appears.

Modelling the ramp explicitly, a fortnight of unstable performance after any significant change, makes forecasts considerably more accurate than assuming an immediate step.