Build a retail media scorecard that goes beyond ROAS

Give every campaign a job

Our recommended planning split is established demand, new customer discovery and product launches. Define the question each budget should answer. A campaign capturing shoppers already searching for your brand should not be assessed exactly like a campaign introducing an unfamiliar product. Use separate reporting where the platform supports it, without fragmenting a small budget into too many thin tests.

Read the sales definition before comparing returns

Amazon's attribution rules vary by ad product and advertiser type. A report can include purchases of products other than the item advertised, and recent results can remain incomplete until the lookback window ends. Confirm the reporting window, click or view basis, eligible products and sales definition before comparing campaigns or platforms. Amazon ad campaign attribution ↗

Use one scorecard for media and retail

Combine media results with your commercial and stock records. Blinkit Brand Central describes city targeting and campaign reporting; Amazon provides search-term reports that can inform targeting decisions. The available controls differ, so adapt the scorecard to each account rather than assuming every platform exposes the same data. Blinkit Brand Central ↗ · Amazon search-term reporting ↗

• Media: spend, impressions, clicks and attributed purchases, using each platform's definitions.
• Commercial: total net sales, realised margin, fees, discounts and contribution after ads.
• Retail: live assortment, availability, stock cover and promotion changes.
• Learning: the hypothesis tested, changes made and what remains uncertain.

Separate attribution from additional demand

Our interpretation rule: attributed revenue is not, by itself, proof that every sale would disappear without advertising. Where feasible, use a controlled experiment or a carefully matched comparison to estimate additional demand. If you rely on a before-and-after comparison, record price changes, stockouts, promotions and seasonality. These can alter sales even when the campaign is unchanged.

A better budget conversation

Illustrative case: Campaign A reports higher ROAS but mostly serves shoppers already seeking the brand. Campaign B has lower ROAS and reaches a broader relevant audience. Do not automatically move everything to A. Check contribution, stock capacity and evidence of additional sales. Keep a defined learning budget for B, or reduce it if the evidence does not support its purpose. A high ROAS is a useful signal, not a complete growth strategy.