You already run ads and have real conversion history, so skip the setup checklists. The question that matters is whether your Performance Max spend is building incremental profit or just repackaging sales you'd have gotten anyway. These best practices are ordered by how much money they move, not by how often they appear in other posts.
Set your tROAS from margin, not a benchmark
The most common mistake is treating target ROAS as a wish. Raising your tROAS target does not force Google to be more efficient — it tells the system to decline auctions it predicts won't clear the bar, so volume collapses. Google's own documentation warns that a target far above your account's history "may limit the amount of traffic your ads may get" (Google Ads Help — About Target ROAS).
So derive the target from arithmetic instead. Break-even ROAS equals one divided by your contribution margin — this is identity math, not a benchmark.
Say you sell a mug for $24. Your Printify base cost, shipping share, and payment fees total $15, leaving $9 per order. That is a 37.5% contribution margin, so your break-even ROAS is 1 ÷ 0.375 = 2.67x. Set tROAS at break-even times a profit buffer you choose — if you want a 30% cushion over break-even, multiplying 2.67x by 1.3 puts your target near 347%, not whatever a "good ROAS" article told you.
This is the whole game, and it is covered in depth in the Google Ads economics guide. A 5.0x ROAS can still lose money on a thin-margin product once COGS and fees come out, which is why ROAS is never the same thing as profit.
Run PMax and Standard Shopping together
Treating "PMax vs Standard Shopping" as an either/or is behind the field. Google confirmed the two now compete on Ad Rank in the same account, so the higher-ranked campaign wins each impression rather than PMax always outranking Shopping (smec).
The hybrid is the norm, not a hack. Optmyzr's study of 24,702 PMax campaigns found 82% of advertisers run PMax alongside Shopping or Search, and split-budget accounts posted the strongest returns in the study (Optmyzr). Treat the exact ROAS figures as study aggregates, not an expectation.
Feed quality matters more than the multi-channel pitch suggests. smec, which manages spend for 350-plus retailers, reports 74–97% of PMax cost goes to feed-based Shopping-style ads (smec). For a deeper breakdown of when each type does which job, see what the Performance Max campaign type is built for.
Exclude your brand so PMax can't pad its ROAS
Here is the trap that makes PMax look like a hero while your business stays flat. PMax chases the cheapest conversions available, and branded searches are the cheapest of all. It absorbs conversions that organic, direct, or your brand Search campaign were already winning, then reports a spectacular ROAS for sales you never paid for before.
smec frames the cost plainly: you end up paying "$1.50 for a click that you could have bought for $0.20" (smec). The honest test is total business orders against total ad spend, not per-campaign ROAS.
The controls exist. Account-level brand exclusions for PMax have been available since 2024 and cover brand variants and misspellings (AdNabu). Pair exclusions with a dedicated brand Search campaign and a regular search-term audit — the full playbook is in how Performance Max brand exclusions work.
Read the channel and search-term reports weekly
"PMax is a black box" is a stale complaint, and content repeating it loses. The channel performance report now breaks PMax results out by Search, Display, YouTube, Discover, Maps, Gmail, and Search partners, with data viewable for any date range after June 6, 2025 (Google Ads Help).
Full search-terms reporting and campaign-level negative keywords also shipped for PMax in the 2025 update wave. Use the channel report to answer a concrete question: what share of your PMax conversions came from Search versus Display? If a tROAS cut pushed incremental spend into high-CPC Search auctions, that is the system buying harder auctions to hit the volume your lower target unlocked.
Feed quality is the load-bearing input
Because Shopping has no keywords, your product title is the query-matching surface. Front-load brand, product type, and the defining attribute, and keep the title matched to the landing page — stuffed titles risk disapproval, the most expensive outcome of all.
Price competitiveness acts like a bid multiplier. Shopping surfaces compare your price against competitors selling the same GTIN, so an uncompetitive price suppresses impression share no matter how high you bid. Watch for "zombie" SKUs too — products PMax starves of impressions for lack of conversion history; smec defines the cohort as products under roughly ten clicks a month (smec).
Diagnose before you touch anything
When ROAS drops, check change history first, then conversion tracking, then your brand mix — before blaming the campaign. The market is often the cause: WordStream's 2025 data showed roughly 87% of industries saw year-over-year CPC increases, averaging about a 12.9% rise (Search Engine Land).
Benchmarks are context, not decisions. LocaliQ's 2026 data puts the retail-category Search CPC near $4.14 with a 4.01% conversion rate (LocaliQ) — useful for orientation, but your break-even math decides whether a $4 click is cheap or ruinous. And after a product disapproval, expect reprocessing and impression recovery to take days to about two weeks once you fix the data (HawkSEM).
Give the volume question a real answer
Value-based bidding needs volume to behave. For Search and Shopping, tROAS technically requires at least 15 conversions with values in the past 30 days (Google Ads Help), but the practitioner floor for stable behavior sits closer to 50 conversions a month (Search Engine Land). Below that, expect volatility no matter how clean your settings are.
When you do adjust the target, move it in small steps — the commonly cited cadence is 10–20% changes at a time, then let the system recalibrate for one to two conversion cycles (Triple Whale).
Where Victor fits for operating stores
Every best practice above depends on one number most tools can't give you: true per-order profit. Victor, the senior AI employee from PodVector AI, sees every order, every ad dollar, and every supplier cost together, so he computes real profit on each product before setting any target. He knows your Printify, Printful, and Gelato costs before the first sale, so break-even ROAS is never a guess.
From there Victor works inside the accounts directly. He adjusts tROAS targets and budgets in Google Ads, adds brand exclusions and negative keywords, pauses the losing asset group, and reprices thin-margin SKUs in Shopify — then delivers the weekly profit report to your Google Drive. He is not a dashboard you read; he is the operator who makes the change. See how this compares to hiring out the work in the case for a Google Shopping ads agency, or put Victor on your account.
FAQs
What is the single most important Performance Max best practice in 2025?
Deriving your tROAS target from your break-even margin instead of a benchmark. Raising the target to a number you "want" suppresses volume rather than creating efficiency, which Google documents directly. Start from one divided by your contribution margin, add a profit buffer, then step toward your goal.
Should I replace Standard Shopping with Performance Max?
No. They now compete on equal Ad Rank footing, 82% of studied advertisers run both, and split-budget accounts performed best in the Optmyzr data (Optmyzr). Think portfolio, not replacement — PMax for scale, Standard Shopping to force data on neglected SKUs.
How do I stop Performance Max from stealing credit for my brand sales?
Turn on account-level brand exclusions, which have covered brand variants and misspellings since 2024 (AdNabu). Add a dedicated brand Search campaign and audit PMax search terms for branded queries. Judge PMax on non-brand performance and total business orders, not its reported ROAS.
How many conversions does Performance Max need to run well?
The technical floor for tROAS on Search and Shopping is 15 conversions with values in the past 30 days (Google Ads Help). Stable, low-volatility behavior is widely observed to need closer to 50 a month (Search Engine Land). Below the floor, consolidate campaigns rather than splitting budget thin.
Is Performance Max still a black box in 2025?
Not anymore. The channel performance report, full search-terms reporting, and campaign-level negative keywords all shipped in 2025, with channel data viewable from June 6, 2025 onward (Google Ads Help). The best practice is to actually read those reports every week.
Does a high Performance Max ROAS mean it's working?
Not until brand traffic is excluded or segmented. A PMax campaign eating branded search can post a huge ROAS while adding zero incremental orders. Reconcile total business orders against total ad spend to see the truth.