How to Audit Performance Max Asset Group Reporting

An asset group should represent a coherent product, service, offer, or audience theme supported by matching text, images, video, products, and landing pages. Before opening performance columns, write the intended customer question and commercial job for each group. Record its campaign, conversion goals, product filters, URLs, audience signals, search themes, and active dates. If two groups have indistinguishable jobs, reporting cannot tell the team which theme caused a result because delivery can overlap inside the same optimized campaign.

Google describes asset groups as collections of creative assets centered on a theme and recommends continuing to evaluate the campaign at campaign level. Treat the group report as a contribution view within that system. It can reveal delivery and outcome differences that deserve investigation, but it does not create a randomized comparison. Preserve budget, bidding, attribution, conversion windows, and major site or feed changes for the selected period before turning a row into a creative recommendation.

Open the asset-group table for one Performance Max campaign and add the columns needed for its stated objective. Useful starting fields include impressions, clicks, cost, conversions, conversion value, cost per conversion, and conversion value per cost when those measures are configured correctly. Fix the date range, time zone, attribution setting, conversion actions, and currency. Export or record the filtered view so a later reviewer can reproduce the comparison instead of relying on a screenshot with missing settings.

Use sufficient completed data and label recent or delayed periods. Do not compare a newly launched group with a mature group as though both had equal opportunity. Note product availability, policy status, asset approval, and budget constraints. Segment only where the report supports the requested question. If a group contains different products, regions, or margins, calculate commercial acceptability outside the ad interface using approved source data rather than treating platform ROAS as a complete profit measure.

The asset-group row summarizes outcomes associated with the group. The asset report shows individual creative assets and available performance information. The combinations report displays top combinations of text, image, and video and previews how those combinations may appear. These surfaces answer different questions. A top combination is not a controlled proof that every included asset caused the result, and an individual asset label is not a customer-level causal measure.

Inventory every required format and approval state before replacing creative. Google says the asset report can be filtered and downloaded and identifies whether an asset was advertiser supplied or created or enhanced with Google AI. Capture the asset ID or exact content, type, source, approval, group, and date. When a combination exposes a misleading pairing or product mismatch, fix the underlying asset set and destination contract. When it merely looks unfamiliar, verify actual eligibility and serving evidence before editing.

Google explicitly cautions that asset groups with higher CPA or lower ROAS can still contribute to campaign goals. Check whether the row has enough volume, whether products were available, whether conversions are valid, whether the group serves a distinct funnel or catalog job, and whether other groups would cover that job after removal. Compare search themes, landing pages, product filters, geographic eligibility, asset completeness, policy limits, and channel diagnostics. A weak row may expose a feed or eligibility problem rather than poor creative.

Create a hypothesis with a falsifiable observation. For example, an incomplete video set may limit delivery on a channel, or an outdated product image may contradict the landing page. Change one bounded component or add a missing format, then record the exact asset versions and time. Avoid deleting a whole group and refreshing every asset simultaneously. That destroys the baseline and makes any later movement impossible to interpret.

For each review, store campaign and group IDs, theme, objective, date range, report filters, conversion definitions, product availability, assets, landing pages, channel notes, finding, proposed action, owner, approval, and review date. Attach exports without customer-level identifiers. Distinguish observed platform metrics from business calculations and from hypotheses. If the report changes after attribution or conversion repairs, preserve the earlier snapshot rather than rewriting the decision history.

Review after the documented learning and conversion windows relevant to the account, not on a fixed daily impulse. Confirm the edited assets are approved and serving, the destination still matches, and no new feed or tracking failure explains the outcome. Roll back when the change creates a material customer mismatch or removes required coverage. Report improvements as observed changes with concurrent factors, never as guaranteed causal impact. The useful result is an asset system the team can explain and maintain.

An asset-group review can end without an authorized edit. Record insufficient volume, overlapping themes, incomplete conversion evidence, or an unresolved feed problem as the reason. Name the next eligible observation date and the evidence needed to reopen the decision.

This prevents daily reviewers from reaching different conclusions from the same partial row. It also protects useful coverage from being deleted merely because a report invites action before the campaign, catalog, or conversion window can support it.