Media incrementality: the question ROAS cannot answer on its own

Incrementality is the additional result caused by an intervention, compared with what would have happened without it. In media, the question is how much the campaign added. Attributed ROAS relates credited revenue to ad spend, but it does not show on its own that the revenue depended on the ad.

A campaign can reach people who would have bought anyway. It can also influence purchases that do not appear in the platform's attribution window. For large advertisers, confusing these two readings can steer budget toward channels that capture existing demand and undervalue actions that create future demand.

Start with the decision, not the test format

Define the choice the study should support: keeping an investment, expanding a region, or comparing levels of advertising pressure. Specify the outcome, population, and horizon. A vague question produces a test that ends with a number, but not with a decision.

Whenever feasible, random assignment between groups helps build a causal comparison. In other contexts, geographic tests or quasi-experimental methods can be considered, with their assumptions and limitations. The choice requires statistical, operational, and commercial assessment.

What must be agreed before the campaign

• Primary metric: sales, contribution margin, qualified opportunity, or another clearly defined outcome.
• Minimum relevant effect: the difference that would justify changing the investment decision.
• Duration and size of the test: consistent with variability, the buying cycle, and the ability to detect the effect.
• Contamination: cross-exposure, promotions, price changes, and parallel actions that could affect the comparison.
• Analysis rule: what will be concluded given a positive, negative, or inconclusive result.

An inconclusive result also needs interpretation

Failing to detect a difference does not prove that the effect is zero. The study may have low precision, inadequate duration, or operational conditions different from those planned. Present uncertainty intervals and explain what the data allow you to claim. Avoid turning a point estimate into financial certainty.

In a hypothetical example, a retail chain might test a media increase in comparable markets while tracking price, availability, and promotions. If stock fails in only some of the markets, the reading stops being a simple comparison of advertising pressure.

How to combine the evidence

Attribution helps observe recorded paths. Experiments investigate causal effects under defined conditions. Marketing mix models help analyze aggregate relationships and scenarios, depending on the assumptions. The Meridian documentation presents a modeling structure that can incorporate prior knowledge and experimental results. No method removes the need to assess data quality.

Organize a testing agenda by the value of the decision. It is not necessary to test every channel at once. Prioritize where there is significant investment, material doubt, and the possibility of acting on what is learned. Keep designs, results, and conditions on record to avoid repeating studies without accumulating knowledge.

Does a high ROAS mean high incrementality?

Not necessarily. The attributed result may include purchases that would have happened without the campaign. They are measures with different interpretations.

Can everything be measured by experiment?

Not always. Constraints of scale, time, and execution call for combining methods. Learn about the Media practice and how to structure budget allocation.