Performance Max asset-group analysis works best as a drill-down, not a leaderboard. Start with the campaign’s business result. Compare asset groups against that baseline. Then inspect channel splits, creative strength, individual assets, and top combinations before deciding what to change.
With HireOtto connected to Google Ads, an AI assistant can pull these PMax reports, aggregate the rows, calculate CPA or ROAS, identify patterns, and prepare a reviewed action plan. The practitioner still decides whether the data is mature enough, whether the groups are genuinely comparable, and whether a structural or creative change makes commercial sense.
The short answer
Use four layers of evidence:
Campaign performance: Is PMax meeting the account’s conversion and value goals overall?
Asset-group performance: Which themes or landing-page groups are driving spend, conversions, conversion value, CPA, and ROAS?
Asset-group strength: Does each group have enough creative variety, and what does Google currently recommend adding?
Individual assets and combinations: Which asset types are serving, converting, or appearing in top combinations?
Do not pause an asset group merely because its average CPA is higher. Google explicitly cautions that higher-CPA or lower-ROAS groups can still contribute to the campaign’s overall goal. Treat the report as diagnostic evidence, not a delete button.
Why PMax asset-group reporting needs a different mental model
Search campaigns give you familiar layers: campaign, ad group, keyword, ad, and search term. Performance Max does not use standard ad groups or ads in the same way. Its core creative and thematic unit is the asset group.
An asset group usually represents a coherent product, service, audience, language, or landing-page theme. Google combines its headlines, descriptions, images, logos, and videos across eligible inventory.
That changes the question. You are not asking, “Which ad group deserves more budget?” Asset groups do not have independent budgets. You are asking:
Which themes are contributing to the campaign goal?
Where is efficiency or volume materially different?
Is the difference caused by creative coverage, the offer, the landing page, the audience signal, the channel mix, or simply insufficient data?
What is the smallest useful change we can test without confusing the next read?
What the PMax reports actually tell you
Campaign performance
Use this as the commercial baseline. Review spend, conversions, conversion value, CPA, and ROAS for the whole PMax campaign. If the campaign’s conversion setup is unreliable, every downstream comparison inherits that problem.
Asset-group performance
HireOtto can return each asset group’s spend, impressions, clicks, CTR, average CPC, conversions, cost per conversion, and conversion value. The current report also includes the Google Ads network type.
That last detail matters: the same asset group may appear in multiple rows because performance is split across networks. Ask the AI to aggregate each asset group first, then preserve a separate channel breakdown. Otherwise, you may accidentally compare a YouTube row with an all-channel total.
Asset-group strength
This is a current-state quality view, not a performance trend. It returns status, primary status, Ad Strength, and Google’s action items. Use it to spot missing creative coverage or eligibility issues. Do not treat “Excellent” as proof that the group is profitable, or “Poor” as proof that it is unprofitable.
Individual-asset performance
HireOtto can retrieve asset type, serving status, primary-status details, network, impressions, clicks, CTR, cost, conversions, and conversion value for assets linked to an asset group. This is useful for finding assets with meaningful delivery but weak outcomes, or assets that are not eligible to serve.
Google’s current reporting supports quantitative asset-level metrics. One important caveat remains: conversions are not divided across every component of an assembled ad. Do not add conversions across headlines, images, and videos and expect the sum to equal campaign conversions.
Top asset combinations
The combinations view shows which sets of text, image, and video assets Google is surfacing among the top combinations. Use it to understand the message and creative mix Google is favouring. It is not a clean conversion-attribution report for each combination.
A six-step PMax asset-group analysis workflow
Step 1: Validate the measurement before ranking anything
Confirm the campaign is optimizing toward conversions that represent real business value. Check whether value tracking is consistent, whether recent tracking changes affect the comparison, and whether each asset group serves the same commercial goal.
A lower CPA for a lightweight lead and a higher CPA for a qualified opportunity are not directly comparable. Neat tables can still tell untidy lies.
Step 2: Pull the campaign baseline and asset-group totals
Start with a 30-day window when the campaign has regular conversion volume. Use 60–90 days for low-volume accounts, while checking whether material changes make the longer period misleading.
Copy-paste prompt:
Get PMax campaign performance and PMax asset-group performance for campaign [CAMPAIGN_ID] in account [CUSTOMER_ID] for the last 30 days. Use summary_and_csv.
First aggregate every asset group across network rows.
Then calculate CPA and ROAS where possible.
Compare each group with the campaign average, but do not recommend pausing or removing anything yet.Step 3: Keep the channel split as a second view
After calculating totals, inspect how each group behaves across Search, YouTube, Display, Discover, Gmail, and other reported networks. Channel mix can explain why two groups with similar themes show different CTR, CPC, or conversion behavior.
Ask for two outputs:
one row per asset group for the executive comparison;
one asset-group-by-network table for diagnosis.
Step 4: Add strength and eligibility evidence
Now retrieve the asset-group strength report. Flag Poor or Average groups, current action items, non-eligible primary statuses, and groups with thin creative coverage.
Copy-paste prompt:
Get current asset-group strength for campaign [CAMPAIGN_ID].
Join it to the asset-group performance summary by asset_group_id.
Separate performance evidence from creative-coverage evidence.
Explain where they agree, where they conflict, and what additional evidence is needed.Step 5: Drill into individual assets only where the group-level evidence warrants it
Do not dump every asset into the first response. Pick the groups that deserve investigation: high-spend underperformers, groups with deteriorating results, or groups whose strength and performance disagree.
Copy-paste prompt:
For asset groups [ASSET_GROUP_IDS], pull individual PMax asset performance for the last 30 days and the top asset combinations.
Group assets by field type and network.
Flag serving or eligibility issues, high-delivery assets with weak conversion outcomes, and top combinations worth preserving.
Do not sum asset-level conversions into a campaign total.Step 6: Produce a reviewed action plan
The final output should separate observation, hypothesis, and action. For example:
Observation: one asset group has a materially higher CPA after enough conversion volume.
Hypothesis: the group’s channel mix, landing page, or creative message may be reducing intent.
Next check: compare the channel split, strength actions, asset-level delivery, search terms, and landing-page relevance.
Proposed action: add missing creative formats, refresh a weak message, improve the landing page, or restructure only after review.
Approval gate: make no Google Ads changes until the practitioner approves the exact assets or structural edits.
Worked example: do not confuse efficiency with contribution
Suppose a lead-generation PMax campaign has three asset groups over 30 days:
Core category: $4,200 spend, 84 conversions, $50 CPA, 3.0 ROAS.
Industry use case: $2,700 spend, 30 conversions, $90 CPA, 2.0 ROAS.
Competitor alternatives: $800 spend, 20 conversions, $40 CPA, 4.0 ROAS.
The tempting conclusion is to remove the industry group and “shift budget” to the competitor group. That skips several realities:
PMax budget sits at campaign level, not asset-group level.
The efficient group has much less volume and may not scale at the same rate.
The industry group may reach incremental demand that other groups do not.
Different network mixes or landing pages may explain the gap.
Conversion quality may differ even when the platform’s conversion count does not.
The better next move is a drill-down. Check whether the industry group has weak creative coverage, irrelevant queries, a mismatched landing page, or one network absorbing spend without value. Then make one reviewed change and measure again.
A practical decision framework
High volume, strong efficiency
Protect the working theme. Review top combinations and asset coverage before refreshing anything. Add variety without deleting the evidence that is already working.
High volume, weak efficiency
Prioritize diagnosis. Inspect channel mix, conversion quality, search terms, landing-page relevance, strength actions, and individual assets. This group has enough data to justify attention, not necessarily removal.
Low volume, strong efficiency
Treat it as promising but unproven. Ask whether the theme is too narrow, the group lacks coverage, or the available demand is genuinely limited. Avoid assuming its current efficiency will hold at higher volume.
Low volume, weak efficiency
Check eligibility and creative completeness first. If the group serves a strategically important theme, improve the inputs. If it duplicates another group or lacks a clear job, consider consolidation after review.
Mistakes to avoid
Ranking asset groups before validating conversion tracking.
Comparing groups with different goals or lead quality as if they are identical.
Forgetting that network segmentation can create several rows per asset group.
Treating Ad Strength as a profitability score.
Adding asset-level conversions together.
Treating top combinations as clean conversion attribution.
Recommending budget changes at asset-group level.
Making several creative, signal, and structural changes at once.
Where HireOtto fits
HireOtto lets you run the full diagnostic inside an AI assistant instead of opening and exporting each PMax view manually. You can pull campaign, asset-group, strength, asset, combination, search-term, placement, and feed reports; request inline summaries or CSV exports; and continue from analysis into reviewed asset or signal changes.
For changes, HireOtto can add or reuse assets, manage asset groups, update signals and search themes, and remove assets while checking Google’s minimum asset requirements. The useful boundary is simple: let the AI do the retrieval, calculation, and mechanical execution; keep business judgment and approval with the practitioner.
Related guide on managing PMax campaigns with AI here.
Get started with HireOtto: https://docs.hireotto.com/quickstart
FAQ
What is the difference between asset-group and asset performance?
Asset-group performance measures the combined results of a themed group of creative assets. Asset performance reports on individual linked headlines, images, descriptions, logos, or videos. Use the first to locate the problem; use the second to investigate it.
Can I calculate ROAS by asset group?
Yes, when conversion value is tracked. Divide conversion value by cost. Ask the AI to calculate it after aggregating network rows for each asset group.
Can I assign budget to one PMax asset group?
No. Budget is controlled at campaign level. If a theme needs independent budget or bidding control, a separate campaign may be the structural answer – but that is a larger decision with its own trade-offs.
How much data should I use?
Use enough data to represent normal conversion volume. Thirty days is a practical starting point for active campaigns; low-volume campaigns may need 60–90 days. Shorten or split the window when tracking, budgets, offers, or structure changed materially.
Should I remove every low-performing asset?
No. Check delivery volume, asset type, channel, eligibility, recent changes, creative coverage, and whether the asset appears in top combinations. Replace assets deliberately, and avoid removing so much coverage that the group becomes weaker or invalid.
Why do asset-level conversions not add up to campaign conversions?
A PMax ad can contain several assets, and Google does not split one conversion cleanly across every component. Asset metrics help comparison and diagnosis, but they are not additive attribution.
About Me
I’m Suyash – badminton junkie, ex‑GroupM ad‑ops grunt, first marketer at a B2B SaaS startup, and creator of Hiretto: Google Ads MCP Server.
My mission: less clicking, more thinking.
Let’s build leverage together.