Retail Media Metrics

The practical question

A marketplace can credit an advertisement for a sale the shopper was already likely to make. Retail media metrics therefore need two views: what the platform attributes and what additional business the advertising creates. Both are useful, but they answer different questions. This guide explains attributed ROAS, new-to-brand measures and incremental sales alongside margin and digital-shelf conditions. It gives a brand team a way to discuss performance without assuming that every tracked purchase represents a newly won customer.

Keep the platforms attribution rule attached

An attributed purchase means it met the platform’s rules for eligible interaction, product scope and time window. Record whether credit follows a click, a view or both, and whether the report includes only the promoted product or other products from the brand. If two platforms claim the same sale, adding their attributed revenue can overstate total business. Preserve platform reports for optimisation while maintaining a reconciled commerce view. Do not quietly switch windows when comparing campaigns. A wider window may increase credited sales without changing actual purchases. Explain the rule in the report so finance and the brand team can assess what the number supports.

Illustrative example · fictional

Read new to brand with its lookback window

Amazon Ads determines its standard new-to-brand purchase classification by reviewing whether the shopper purchased from the brand during the previous twelve months.¹ That is a platform-history definition, not proof the person has never bought the brand anywhere. Someone may be an existing customer in a physical store or may have bought more than a year ago. State the platform, eligible ad product and current reporting definition. New-to-brand order share uses eligible attributed new-to-brand orders divided by the relevant eligible attributed orders. Cost per new-to-brand purchase uses the corresponding advertising cost and count. Keep the denominator visible because availability of these measures varies by reporting product.

Sales credit and causal lift are different

Incremental sales estimate purchases above what would have happened without the advertising. NielsenIQ notes that retail media interacts with pricing, promotions, distribution and other channels, making attribution alone insufficient for that question.² Where available, use an appropriately designed holdout or another credible causal approach. Define the treatment, population, period, outcome and assumptions before interpreting the result. Cross-device activity, organic substitution and competitor effects can complicate the design. A statistically uncertain estimate is still informative if its uncertainty is shown. Avoid treating a platform’s modelled incremental result as self-explanatory; ask what counterfactual was used and whether the outcome captures the business decision you need to make. Our guide to marketing incrementality covers the next question. If that is your next question, read our guide to attribution vs incrementality.

A fictional campaign shows the difference

Imagine a fictional household-product campaign spends INR 100,000 and receives INR 700,000 in attributed sales under a stated click window. Revenue ROAS is seven. A valid experiment estimates INR 180,000 in additional net sales, with a wide uncertainty interval. Incremental revenue ROAS on that estimate is 1.8. If contribution before advertising on those extra sales is 35 percent, estimated contribution is INR 63,000 before the ad cost. Retail media metrics now prompt a different commercial discussion. The illustration assumes the experimental estimate is valid and uses a simplified margin. It is not evidence of actual campaign performance or a universal threshold for marketplace advertising. ROI vs ROAS is explained in a separate guide.

The digital shelf can change the result overnight

Stock availability, price, reviews, delivery promises and buy-box status all affect whether a shopper can complete the purchase. NielsenIQ identifies these commerce conditions as part of credible retail-media analysis.² Record their changes alongside the media schedule. Advertising an unavailable product can create interest that converts to a competitor. A deep promotion can improve attributed ROAS while reducing contribution. A review problem may depress conversion across all channels. Compare suitable periods and products, then investigate the specific constraint. The advertising team needs a working relationship with marketplace operations because media optimisation cannot repair a product page that misstates the product or a stock system that repeatedly fails. For the wider picture, see our guide to D2C marketing metrics.

Talk it through

Look beyond attributed sales

Share the platform reporting rule, product-sales view and stock context. Quiamo can help identify the questions that attributed ROAS leaves unanswered before a larger media commitment.

Review my retail media evidence

Measure customer growth beyond the first order

Follow eligible acquired customers or shopper cohorts where lawful platform tools and data access permit. Examine repeat purchase, product switching and realised contribution over a relevant consumption period. Keep platform customer identifiers within the approved environment and use aggregate reporting for planning. A high new-to-brand share can be useful when recruiting category buyers, but it may arrive at a poor margin or fail to produce repeat demand. A low share may support another legitimate job, such as defending visibility during a competitive promotion. State that job before evaluating the score. The media objective should guide the measurement rather than being rewritten after the most flattering number appears.

Use the evidence to set the next investment

Separate three questions in the review: did the platform deliver the intended audience, did commerce conditions support purchase, and did the spending create sufficient additional contribution? Some programmes will have strong evidence for the first two and limited causal evidence for the third. Say so. Quiamo can help shape a marketplace communication and measurement brief covering product-page clarity, creative variants and reporting definitions. Retail operations and finance should approve stock, net-sales and cost inputs. The scope can also identify where an experiment is needed before scaling spend. It should not promise an incremental return simply because platform attribution produces a generous ratio.

A team packs rows of branded white and black boxes on a long table, seen from above, illustrating retail media metrics. — Conceptual editorial image with fictional subjects, not evidence of results.

Before interpreting the result

Rows of identical cream jars stocked on retail shelves, illustrating retail media metrics. — Conceptual editorial image with fictional subjects, not evidence of results.

A hand holds a phone showing a coral drinks can beside the same can on a retail shelf, illustrating retail media metrics. — Conceptual editorial image with fictional subjects, not evidence of results.

Before a spend review, save the platform’s current reporting definition and note any change in attribution settings. Reconcile a sample of credited orders with the commerce report, then ask whether halo products, returns and tax treatment are included. Inspect product availability during the advertising period. A short stockout or lost buy box can distort a comparison with another week. If an incrementality study is supplied, ask which shoppers or markets formed the control and whether spillover was considered. Keep the uncertainty interval with the estimate when it is available. The report should then identify a next decision: fix the product page, adjust the audience, test a placement or investigate profitability. A large attributed ratio without that context is a starting point for questions, not a sufficient investment recommendation.

Questions people ask

Does new to brand mean a completely new customer?

Not necessarily. It refers to the platform definition and purchase-history window. On the standard Amazon definition, the lookback is twelve months. Purchases elsewhere or earlier may not be captured by that classification.

Can attributed ROAS still be useful?

Yes. It helps compare activity under stable reporting rules and supports tactical optimisation. Use it with net sales, contribution and causal evidence when making broader investment decisions.

Should we stop advertising products that already rank organically?

That requires evidence. Paid placements can create or defend additional demand, but may also substitute for organic purchases. Test the relevant product and market conditions rather than assuming one outcome everywhere.

What should we ask before scaling a marketplace campaign?

Check mature net sales, contribution, stock availability and the reporting window. Ask whether the result reflects additional demand or likely substitution from organic purchases. Where causal evidence is limited, define a sensible test rather than assuming the attributed ratio will persist at higher spend or across different products and markets.

The next useful step

Retail media metrics should help distinguish a credited sale from additional profitable demand. Bring Quiamo the platform’s attribution definition, a reconciled product-sales view and the creative being used. We can scope the content and measurement work around the investment question that attributed ROAS alone cannot answer. Discuss the communication brief with Quiamo.

Next step

Fix the product story before scaling spend

Describe the marketplace, product and customer uncertainty. We can scope product-page content and creative variants alongside reporting definitions that preserve margin and attribution boundaries. Remove personal and confidential information from any example you share.

Brief my marketplace content

Sources

Amazon Ads new to brand reportingNew-to-brand purchase classification uses prior twelve-month platform purchase history.
NielsenIQ retail media incrementalityAttribution differs from incrementality and commerce conditions affect outcomes.

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.

Retail Media and Marketplaces: Sales Credit vs Incremental Sales