A campaign earns an average 5.0 ROAS, so the business doubles the budget and expects the same efficiency on the added spend. That expectation assumes the next opportunities resemble the average of the opportunities already captured.
Marginal analysis asks a different question: what additional value is expected from the next amount of spending? The answer can differ materially from the campaign's average result and should be evaluated against the business's contribution and capacity.
Separate average from additional return
Average ROAS divides total attributed value by total advertising cost under a defined reporting view. A marginal-return estimate relates a change in value to a change in spend over a relevant range.
Write the numerator carefully. Additional platform-attributed value is not necessarily additional company revenue caused by advertising. The attribution and incrementality guide explains that evidence distinction.
Use “observed change” when comparing ordinary before-and-after periods, and reserve stronger causal language for a design or analysis that supports it.
Work through the arithmetic
Consider an illustrative campaign with $10,000 of spend and $50,000 of reported value. A later comparable view shows $12,000 of spend and $56,000 of value. Average ROAS changes from 5.0 to about 4.67.
The difference is $6,000 of value for $2,000 of additional spend, giving an observed change ratio of 3.0. That additional tranche looks less efficient than the original average, even though total value is higher.
This arithmetic does not prove the budget increase caused the entire $6,000 difference. Promotions, seasonality, customer mix, and reporting maturity may also have changed. The calculation is a starting point for the decision, not its causal conclusion.
Translate additional revenue into contribution
Use the contribution rate appropriate to the added business, which may differ from the prior average product mix. The break-even ROAS guide explains the relationship between contribution and advertising cost.
If the illustrative additional $6,000 represented incremental retained revenue with a 40% pre-ad contribution rate, it would contribute $2,400 before the added $2,000 media cost. The modeled contribution after that media cost would be $400, before any additional costs excluded from the model.
If the added orders had a lower contribution rate or required extra operational expense, the conclusion could change. The ratio alone does not settle the economics.
Choose an evidence source for the response
| Evidence source | What it can contribute | Limit to retain |
|---|---|---|
| Historical budget changes | Context about observed responses | Other changes may explain the difference |
| Platform simulation | Estimated opportunities under its model | Forecast and attribution assumptions |
| Controlled experiment | A defined causal spend comparison | Tested scope and uncertainty |
| Marketing mix model | Response estimates under model assumptions | Identification and extrapolation limits |
Google's target-adjustment guidance describes simulators as planning aids for potential outcomes. Treat them as estimates rather than a commitment that the added spend will produce a particular business result.
Keep the range local to the decision
A response estimated around a modest budget increase may not describe a much larger expansion. New spending can change auction opportunities, customer mix, inventory needs, and operational capacity.
State the spend range the evidence concerns. Avoid extending a short local response line indefinitely across a chart and calling it a forecast.
If the proposed increase is materially larger than the observed range, use scenarios or a staged decision with explicit review points. The uncertainty should influence the size of the commitment.
Verify the constraints around the budget
Check bid strategy, target, conversion goal, audience scope, and current budget status. Increasing a nominal budget does not guarantee additional actual spend if another constraint remains binding.
For Google Ads, use the target ROAS change plan when a target adjustment is also proposed. Budget and target changes communicate different instructions and should have separate rationales.
Also verify product availability and sales capacity. Additional demand that cannot be fulfilled can make the marginal economic result worse than the advertising report suggests.
Compare alternatives on the same basis
When choosing between channels or products, align revenue definitions, cost scope, uncertainty, and outcome maturity. Do not compare one channel's causal lift estimate with another channel's full attributed ROAS as if they were the same quantity.
Consider the next available spending tranche in each option, not only each option's historical average. The best average performer may already be operating near a different part of its response curve.
Keep business constraints explicit. The decision can involve contribution, customer mix, cash timing, and risk tolerance rather than one ratio alone.
Define the review before increasing spend
Record the approved amount, expected response range, source of the estimate, operating limits, and review window. Include concrete early-intervention conditions such as a broken offer or unauthorized exposure.
After the change, compare actual spend with the intended increase. Then review mature outcomes and the relevant business contribution. If spend did not change meaningfully, the test may not have evaluated the proposed response at all.
Report the next decision honestly
State the observed additional spend and value, the estimated causal contribution if supported, the uncertainty, and the resulting action. Do not present a falling average ROAS as automatic failure or rising revenue as automatic success.
Marginal analysis is useful when it directs attention to what the next dollar can reasonably accomplish. Its strength comes from matching the arithmetic, evidence, and business economics to the actual budget decision.