Amazon Ads Attribution and Measuring Incrementality

What attribution means

Attribution assigns a sale to an ad according to rules, such as a purchase within a set number of days after a click. Windows and rules differ by ad type and can change. Always read the definition in the reporting you use.

Attributed is not incremental

Some customers would have bought anyway, such as those searching for your brand name. Ad-attributed sales therefore include both incremental sales and sales that would have occurred without advertising. Judging advertising by attributed sales alone can overstate its effect.

Use total-level measures

TACoS relates advertising spend to total sales and helps show whether advertising is building overall business. Watch it alongside ACoS, and look at trends over time.

• TACoS trend over time
• Organic sales trend
• Share of branded versus generic search spend
• New customer versus repeat sales where available

Branded versus non-branded

Branded terms usually convert well but may produce less incremental value. Non-branded terms introduce new customers but cost more. Review them separately and decide on defensive spend deliberately.

Run simple tests

You can test the effect of advertising by changing spend on a product for a defined period and observing total sales, remembering that other factors also change. Keep tests clear and limited, and avoid conclusions from short or noisy periods.

Consider halo effects

Ads on one product can influence sales of related products. Some reporting shows purchases of other products from the same brand after an ad click. Include these effects in your analysis.

Be honest about uncertainty

Attribution is imperfect. Present results with caveats and avoid claiming precise causal effects you cannot prove.

A worked example

A seller cuts spending on branded terms for a few weeks to test. Total sales stay stable while advertising costs fall, suggesting much of the branded spend was not incremental. They keep a smaller defensive budget to protect against competitors.

Measurement checklist

Know attribution windows, track TACoS, separate branded and generic, record changes with dates, run limited tests and report with caveats.

Questions to ask your agency or team

How are results measured? What portion of sales would have happened anyway? How do we judge brand terms? What tests have we run? Look for reasoned answers, not certainty.

Reporting results honestly to stakeholders

When you present advertising results, state clearly which attribution window and metric you are using and what is not captured. Show total sales and TACoS as well as ad-attributed sales. Where you have run tests, describe them and their limits.

Stakeholders usually appreciate honest uncertainty more than precise-looking numbers that later prove misleading. It also builds trust that your recommendations are based on evidence.

Your next three steps

Improve how your team reads advertising results.

• Write down the attribution window for each ad type you use.
• Add TACoS and organic sales to your weekly advertising report.
• Plan one small, time-limited test on branded spend and record the result.