View-Through Revenue in Retail Media: What It Counts and Why It Inflates ROAS

Of all the measurement differences between retail media platforms, view-through attribution is the least visible and the most consequential. Unlike attribution window length—which is usually disclosed in platform documentation—view-through counting is often buried in methodology footnotes. Planners who don’t actively look for it will include it in their ROAS figures without knowing it’s there.

The result is a systematic inflation of reported ROAS on platforms that count view-through conversions, compared against platforms that don’t. When you compare those numbers directly, you’re comparing different measurement realities. Budget follows the inflated number. Performance suffers.

What a View-Through Conversion Actually Is

A view-through conversion occurs when a consumer is served an ad impression—meaning the ad appears in their browser or app—but does not click on it, and then completes a purchase within the platform’s attribution window. The platform credits the conversion to that ad impression, attributing the sale to the campaign even though no direct interaction took place.

This is distinct from a click-through conversion, where the consumer clicks the ad and converts. Click-through attribution has a clear causal chain: the consumer expressed explicit interest by clicking, then purchased. View-through attribution asserts causality without an explicit signal—the consumer may have been influenced by the ad, or they may have been going to buy the product anyway.

The legitimacy of view-through attribution depends entirely on whether the ad meaningfully influenced the purchase decision. For a brand a consumer has never purchased, a display ad that drives awareness before an organic purchase can represent real incremental value. For a brand the consumer already buys habitually, that same ad impression is likely taking credit for a purchase that would have happened regardless.

Which Platforms Count View-Through Revenue

Platform behavior on view-through attribution varies significantly and changes over time. The table below reflects current default configurations as of early 2026.

Amazon Sponsored Products and Sponsored Brands use a click-only attribution model by default. A consumer must click an Amazon ad for the resulting purchase to be credited to the campaign. This makes Amazon’s reported ROAS more conservative than platforms that include view-through conversions in their headline figures.

Walmart Connect counts view-through conversions in its default reported revenue. A shopper who views a Walmart Connect display ad and then purchases within 14 days—without ever clicking the ad—is counted as an attributed conversion. This is included in the ROAS figure unless the planner specifically extracts and excludes view-through revenue from the export.

How Much ROAS Inflation View-Through Causes

The inflation magnitude depends on three factors: the product category, the campaign type, and the platform’s view-through window length. For display-heavy campaigns on platforms with generous view-through windows, the effect can be substantial.

For a fast-moving consumer goods brand running display campaigns on Walmart Connect, view-through conversions typically represent 15–25% of total attributed revenue. For a category with high organic purchase frequency—household essentials, grocery staples, personal care—the number can reach 35–40%, because the platform is crediting purchases that would have happened regardless.

For sponsored product campaigns—where the consumer is already searching for relevant products—view-through inflation is lower, typically 8–15%. The consumer’s intent is already established; view impressions are less likely to change behavior.

Why Direct Exclusion Is Not Enough

The obvious solution is to exclude view-through revenue from all platforms before comparison. This is better than doing nothing, but it introduces its own distortion.

Not all view-through revenue is worthless

View-through attribution does capture real incremental lift in some circumstances. A consumer who has never purchased a brand before, sees a display ad, and later purchases in the same session or within a short window was plausibly influenced by the ad. Blanket exclusion removes this signal along with the inflated portion.

The correct adjustment is a view-through discount factor that reflects the incremental lift attributable to ad exposure, not a binary include/exclude decision. This factor varies by brand, category, and campaign type.

Platform exports don’t always separate cleanly

Some platforms report view-through revenue as a combined figure with click-through revenue in the headline ROAS metric, requiring the planner to calculate the split from raw export fields. Others report it in a separate column that must be manually subtracted. The extraction process is error-prone at scale, and the column names are inconsistent across platforms and export formats.

The comparison is only valid within a single platform

If you exclude view-through from Walmart and then compare to Amazon (which already excludes it), the comparison is directionally correct but not precisely normalized. The remaining click-through attribution still uses different window lengths and attribution models. View-through exclusion corrects one variable while leaving others intact.

A Normalized Approach to View-Through

Structural normalization treats view-through as one of several correction factors rather than a binary include/exclude decision. The process involves three steps.

Isolate view-through revenue by platform

Every platform that includes view-through conversions provides some mechanism to extract them separately. For Walmart Connect, the view_through_revenue column in campaign exports is the primary source. For Criteo, the post-view sales field serves the same purpose. The first step is extracting this figure cleanly for each platform.

Apply a category-calibrated discount factor

Rather than removing 100% of view-through revenue, apply a discount that reflects the estimated incremental contribution. A conservative starting point for FMCG brands is a 70–85% discount on view-through revenue—retaining 15–30% as an estimate of genuine incremental lift. This factor should be validated against holdout tests where available.

Combine with window and model corrections

View-through normalization produces a corrected revenue figure that must still be adjusted for attribution window length and model type to produce a truly comparable ROAS. A platform with a 30-day window and view-through inclusion requires both corrections applied simultaneously—sequential correction introduces compounding errors.