Incrementality: attributed versus caused

An incremental signed case is one that would not have happened without the spend. Incremental cases are estimated only in aggregate, from a test that meets the holdout-test rules, and always with an interval. No individual case is ever labeled incremental.

An attributable case is different: a signed case whose originating lead can be matched to a marketing source under the attribution rules. Attributed is not caused.

When a test is required

• Causal words need a test. Any statement that spend “caused”, “created”, “drove” or “added” signed cases needs one. Without it, our word is “attributed”.
• Before a large increase. We propose a test before recommending that any channel’s budget rise by more than 30% on attribution alone. If no test is possible, we say so in writing.

The measures

• Incremental cost per signed case = the change in media a test made ÷ the incremental signed cases it produced, with its interval. In a holdout, the change is the spend withheld; in a step-up, the spend added.
• Incrementality factor = incremental signed cases ÷ attributed signed cases, in the same areas and period. The incrementality-adjusted cost per signed case = media cost per signed case ÷ the factor, shown with the test’s date and interval.
• Marginal cost per signed case describes the next dollar: the change in cost ÷ the change in signed cases between two spend levels. A budget rises only while it stays below the firm’s target and intake has capacity; where either fails is the spend ceiling.

Using a result

• A factor applies only to the channel, areas and conditions tested.
• It expires after 12 months, or sooner if the channel’s budget moves by more than 50%, its campaign strategy changes, or the platform changes how the channel works. Expired factors are labeled.
• Tests step budgets rather than going dark wherever they can, never cut intake coverage, use only attorney-approved, filed copy, and stop whenever the firm asks.

Why it matters

Brand search, retargeting-style formats and channels that sit at the end of a buyer’s path collect credit for cases that other spending created. Without a test, the most-credited channel can be the least necessary one.

Cite as: The Probative Standard v1.0, Incrementality. Last reviewed October 24, 2026, by the Probative editorial team.