If you run a store with real sales history, you have probably heard that Performance Max is a black box. That was true. It is not anymore.
Since the channel performance report rolled out with data available from June 2025 onward, and full search-terms reporting shipped the same year, PMax exposes most of what you need to run it like an operator. The problem now is not missing data. It is that the ranking guides teach you to read the insights page, not to decide with it.
This walks through every report on the insights page, then does the part they skip — turning those numbers into a per-order profit call.
Where the Performance Max insights page lives
Open Google Ads, select your PMax campaign, and click Insights & Reports. Google refreshes most of this daily.
You will find several distinct reports there, and each answers a different question. Treat them as a diagnostic sequence, not a dashboard you glance at.
The four that matter for a store are the channel performance report, search terms insights, asset and audience insights, and the search-categories view. Our Google Ads economics guide covers how these fit the wider account; here we go deep on the reports themselves.
The channel performance report: where your money actually goes
This is the report that killed the black-box complaint. It breaks your PMax spend across the seven surfaces PMax serves — Search, Display, YouTube, Discover, Gmail, Maps, and Search partners — with impressions, clicks, conversions, conversion value, and cost for each.
Here is the insight most guides bury. PMax markets itself as multi-channel, but for retailers it is overwhelmingly a Shopping engine. smec, which manages ads for hundreds of retailers, reports that between 74% and 97% of PMax cost goes to feed-based Shopping-style ads.
So when you open the channel report and see most of your spend on Search and Shopping surfaces, that is normal. The practical takeaway: your product feed is the load-bearing input, not your YouTube creative.
Watch the report month over month. If your feed-based conversions drop while Display spend climbs, the system is buying cheaper, lower-intent impressions — a signal to check your feed for disapprovals or price parity, not to celebrate the extra reach.
Search terms insights: the report that catches what PMax is really doing
PMax now shows you the actual queries your ads matched, grouped into search categories with clicks and conversions attached. This is the single most useful report on the page, and it is where the profit story starts.
The reason: PMax chases the cheapest conversions it can find, and branded searches are the cheapest of all. Left unchecked, it eats queries your organic listings or brand Search campaign were already winning, then reports a spectacular return for conversions that were coming anyway.
Practitioner analyses estimate that roughly 8% to 15% of PMax budget leaks to brand queries in unprotected accounts, inflating apparent ROAS by around 15% to 30%. smec puts the cost mechanism plainly: you end up paying "$1.50 for a click that you could have bought for $0.20".
So read the search terms report with one question: how many of these conversions are on your own brand name? If a chunk are, your headline ROAS is fiction. The fix is account-level brand exclusions plus a dedicated brand Search campaign — the same controls covered in today's Performance Max updates.
Asset and audience insights: useful, but not a KPI
The insights page also rates your creative assets and shows which audience segments engage with which text, image, and video. An outdoor store might learn that fitness-minded shoppers respond to lifestyle photos over plain product shots.
Use this to refine asset groups and audience signals — the hints you feed PMax about who converts. But remember signals are suggestions, not fences; Google expands past them.
Do not over-index on asset ratings. A "low" rated asset that happens to sit in your best-converting group is still earning its keep. Judge groups on conversions and value, not on Google's asset grades.
Reading the insights page like an operator: a worked example
Numbers on a screen mean nothing without your margin. Here is how an operating store turns the insights page into a decision.
Say you sell a mug at a $24 price. Your base cost, shipping share, and payment fees come to $15, leaving $9 of contribution — a 37.5% margin. Your break-even return on ad spend is 1 ÷ 0.375 = 2.67x. Below that, every order loses money before a single dollar of profit.
Now say the campaign runs 420 conversions a month on $3,200 of spend at a reported 4.0x ROAS. That looks great — until the search terms report shows 130 of those conversions came from your own brand name. Strip those and the non-brand math is 290 orders ÷ $3,200, at roughly the same value per order, and your true acquisition return drops well under the headline.
That gap — between reported ROAS and non-brand ROAS against your 2.67x break-even — is the only number that decides whether to scale. The insights page hands you the pieces; the arithmetic is on you.
Using insights to set your target, not chase a benchmark
The most common mistake is setting your tROAS to "what a blog said is good." Your target should come from your break-even math times a profit buffer, then stepped toward your goal — not from a benchmark table.
Benchmarks are context, not decisions. LocaliQ's 2026 data puts the average retail-category cost per click around $4.14 with a conversion rate near 4.01%, and WordStream found that roughly 87% of industries saw year-over-year CPC increases, averaging about a 12.9% rise. Those tell you the auction is getting pricier; they do not tell you your break-even.
The insights page matters here too: your campaign needs enough conversion volume to bid stably. Google's documented floor for tROAS on Search and Shopping is at least 15 conversions with values in the past 30 days, while the practitioner convention is closer to 30 to 50 a month per campaign before the system settles. Check your conversion count on the insights page before blaming settings for volatility.
The zombie-SKU signal hiding in your reports
One last read most guides miss. PMax funnels impressions toward proven converters, so a SKU with no history gets starved — no impressions, no data, no way to prove itself. smec defines these zombies as products under roughly ten clicks a month.
If your insights show a shrinking set of products carrying all the spend, that is the pattern. The fix is a Standard Shopping campaign to force data collection on the neglected SKUs — a hybrid approach that the Optmyzr study of tens of thousands of campaigns found 82% of advertisers already run, with split-budget accounts posting the strongest returns. Whether to bring in help for that is what our note on choosing a Google Shopping ads agency walks through.
Where the insights page stops — and profit begins
Every report on the insights page is denominated in ROAS and conversion value. None of them subtract your product cost, fulfillment, and fees. That is the gap between a good-looking campaign and a profitable one.
This is the work PodVector AI built Victor for. Victor is an AI employee, not a dashboard — he connects your Shopify store, Google Ads, Meta, and your print partner, computes the true per-order profit behind those PMax conversions, and delivers the report to your Google Drive. Every write action he takes on your account is approval-gated, so he drafts the change and you approve before anything runs.
The insights page tells you what happened in the auction. Knowing whether it made you money is a different question — and the one that actually runs your store.
FAQs
What are Performance Max insights?
They are the reports inside a PMax campaign, found under the Insights & Reports tab. The set includes a channel performance report, search terms insights, asset and audience insights, and search-category groupings, most refreshed daily.
Can I see search terms in Performance Max now?
Yes. Full search-terms reporting shipped in 2025, so you can see the actual queries your ads matched, grouped into categories with performance metrics — the same visibility Standard Shopping and Search campaigns have long offered.
Is Performance Max still a black box?
No, and any guide saying so is out of date. The channel performance report exposes spend across all seven surfaces with data available from June 2025 onward, search terms are visible, and campaign-level negative keywords exist.
What is the most important Performance Max insight for a store?
Whether your conversions are new orders or brand traffic you were already getting. Read the search terms report for your own brand name first; a high ROAS built on cannibalized brand searches can hide zero incremental growth.
How do I turn insights into a profit decision?
Start from your break-even ROAS, which is one divided by your contribution margin. Then strip brand conversions from the reported numbers and compare your non-brand return against that break-even — that comparison, not the headline ROAS, tells you whether to scale.
Does a high ROAS on the insights page mean the campaign is working?
Not by itself. A PMax campaign absorbing branded search can show a spectacular return while adding no new orders, so judge on non-brand performance and total business results, not per-campaign ROAS. The difference between PMax and other campaign types is worth understanding too, which our Demand Gen versus Performance Max comparison covers.