Did the Campaign Cause the Sale? An Introduction to Incrementality

The sale happened but what caused it

A customer sees an advertisement and buys. The dashboard gives the campaign credit. That does not tell you whether the customer would have bought anyway. Marketing incrementality asks the harder question: how much did the intervention add compared with a credible version of events without it? That distinction matters most when a campaign reaches people who are already close to buying. High attributed sales can coexist with little additional business. The point of an incrementality study is to make the next investment decision less dependent on a convenient story. For a wider view of the options, explore our marketing frameworks and measurement guide.

Define the intervention precisely

A useful study names what changes and what stays comparable. Are you adding a channel, increasing spend, changing creative or withdrawing an existing campaign? Each intervention answers a different question. If budget and creative change together, the result is about that combined package unless the design separates them. Geographic testing is one approach when location-based intervention and credible comparisons are feasible. Google’s GeoX guidance describes preparation, design, execution and counterfactual analysis. ¹ It is a research process, not a setting that turns every campaign into a proven success. For the wider picture, see our guide to attribution vs incrementality.

Choose an outcome the business can trust

Use a defined commercial outcome such as fulfilled orders, qualified appointments or net revenue. The outcome should be measured consistently across treatment and comparison groups. A media platform’s conversion total may omit purchases through other routes or apply different attribution rules. Store transactions and CRM outcomes can provide a broader view when their quality is sufficient. Decide how returns, cancellations and duplicate customers will be handled. A study that increases gross orders while worsening contribution margin needs a different reading from one that adds profitable orders. Select the primary metric before viewing the result.

A counterfactual needs evidence

Randomised assignment can make comparison groups more credible, but implementation still matters. Customers can cross boundaries, media can leak and other campaigns can differ between groups. Where randomisation is unavailable, a modelled comparison relies on assumptions that must be tested and disclosed. Marketing incrementality therefore includes a design review, not just a lift calculation. Ask whether the intervention can create a detectable change, whether the groups had sufficiently comparable pre-test behaviour, and whether operational events could break the comparison. A study with too little signal may produce uncertainty rather than a dependable winner. If that is your next question, read our guide to geo experiments.

Report both the difference and its uncertainty

Incremental outcomes are estimated actual outcomes minus estimated counterfactual outcomes for the stated population and period. Relative lift divides that estimated difference by the counterfactual baseline. Incremental ROAS divides incremental revenue by the relevant incremental advertising spend. Define every part of the calculation. A lift of 8% with a wide interval has a different meaning from the same estimate with a narrow one. An interval compatible with no effect does not prove the campaign has no effect; it means the study did not resolve the question at that precision. Avoid converting an uncertain point estimate into a firm budget promise.

Build a learning programme rather than one verdict

Start with a decision that has enough financial consequence to justify the research. Document the design, data checks, exclusions and reporting plan before launch. After analysis, keep the result attached to its intervention, period and market conditions. A test at a small spend level does not establish the return from a much larger budget. Quiamo can help connect the campaign brief with a measurement design and work with qualified analysts where the study warrants it. The useful deliverable is a defensible recommendation about what to test or fund next, including what the evidence still cannot resolve. Marketing incrementality becomes more useful as each test informs the next one. Marketing mix modelling is explained in a separate guide.

Agree the decision rule before the test

The team should know what result would justify expansion, continued testing or withdrawal. That rule needs commercial context and uncertainty, rather than a demand for any positive percentage. Ask finance for contribution economics and ask operations whether the business can fulfil additional demand. Write down what happens if the interval remains too wide to resolve the question. You might extend a future study, change the intervention or choose a lower-risk budget rather than announce success. Preserve the original analysis alongside any exploratory follow-up. Subgroup findings can suggest the next test, but selecting the most favourable region after the result arrives is not the same as demonstrating a planned effect. Good governance keeps the learning useful even when the campaign does not provide a clean answer. It also stops a disappointing result from being buried beneath attributed conversions.

Put the measurement plan into practice

Keep a one-page study record with the original question, intervention, population and decision rule. Add dated notes about implementation changes so the interpretation does not depend on someone’s memory after the campaign. When a result is inconclusive, state the practical consequence: the team may need another study, a smaller risk or a different intervention. Do not rewrite the original objective as awareness if the agreed test concerned profitable orders. Communication measures can still be informative, but they answer a separate question. A disciplined record makes it easier to explain both a successful experiment and a campaign that did not generate enough evidence to justify expansion.

Before the campaign goes live

Use this checklist to organise evidence about additional profitable outcomes rather than assigned conversion credit. Assign owners to the tasks and keep the supporting evidence with the campaign record.

• Write the investment decision and intervention in plain language.
• Choose the commercial outcome and reconciliation rules before launch.
• Assess feasible assignment, contamination and detectable effect.
• Prespecify duration, analysis method and stopping rules.
• Monitor implementation without changing the question to favour a result.
• Report uncertainty and contribution economics with the lift estimate.

Questions people ask

Bring the measurement question into the brief

The campaign’s credit and its additional contribution are different questions. Marketing incrementality makes the second question explicit and preserves uncertainty instead of hiding it behind a favourable ratio. Start with the intervention, the outcome and the decision that follows. Quiamo can help connect that research brief with the creative and response journey. If the available data cannot resolve the investment question, the next step is to improve the evidence or limit the risk, rather than make a causal claim the campaign did not earn. Explore our campaign strategy and measurement support to see how the work can be structured. You can discuss your campaign measurement brief with Quiamo.

Sources

1. Google Meridian introduction to GeoXGeographic experiments compare marketing interventions and require historical geographic data and counterfactual analysis.
2. Google Analytics attribution settingsReporting attribution settings determine conversion credit allocation and lookback windows.

Worked examples on this page are fictional and labelled. Search phrases and figures from external sources are cited; nothing here describes a Quiamo client result.