Search "performance max best practices google ads help" and you get the same checklist on every page: upload fifteen images, add audience signals, let it run for six weeks. That advice is fine for someone launching their first campaign. It is useless if you already spend two or three thousand dollars a month on Google and your ROAS looks great while your bank balance does not move.
This guide is for that second reader. If you want the full account-level picture, start with our Google Ads economics guide. Below, three practices that actually change your profit.
Best practice 1: set your tROAS target from margin, not from a blog
The single most common Performance Max mistake is setting your Target ROAS to "what a good number sounds like." A target ROAS above what your account has historically delivered does not force efficiency — Google's own documentation warns it "may limit the amount of traffic your ads may get" (Google Ads Help). Raising the target is a volume decision wearing an efficiency costume.
So derive the target from arithmetic instead. Your break-even ROAS is 1 ÷ contribution margin. Say you sell a mug at $24. Your Printify base and shipping run $13, and payment processing takes about $1. That leaves $10 of contribution, a margin of 10 ÷ 24 = 41.7%. Break-even ROAS = 1 ÷ 0.417 = 2.4x. If you want a profit buffer, set tROAS to roughly 2.9x–3.6x (break-even × 1.2–1.5), not whatever a case study bragged about.
Then move it in small steps. Practitioners commonly adjust targets by 10–20% at a time and wait a conversion cycle or two before judging, because any significant goal change re-enters the learning phase (Triple Whale). Jerk the target from 300% to 500% and watch spend collapse — that is the system working exactly as designed, not breaking.
Best practice 2: exclude your brand before you scale a dollar
Performance Max chases the cheapest conversions it can find, and your own brand searches are the cheapest of all. Left unchecked, PMax absorbs orders that would have come from a branded search, from organic, or direct — and then reports a gorgeous ROAS for "winning" them.
Practitioner analyses estimate 8–15% of PMax budget leaks to brand queries in unprotected accounts, inflating apparent ROAS by 15–30% (GrowthSpree). The agency smec describes the cost mechanism bluntly: you end up paying "$1.50 for a click that you could have bought for $0.20" (smec).
The fix exists and is free. Turn on account-level brand exclusions for PMax — available since 2024 and covering brand variants and misspellings — then run a dedicated brand Search campaign to catch that traffic cheaply (AdNabu). Now your PMax number reflects new customers, not recycled ones.
The honest test of whether PMax is working is not its reported ROAS. It is total business orders divided by total ad spend across the account, period over period. If PMax ROAS climbed while your brand Search impressions quietly fell, that is reallocation dressed up as growth.
Best practice 3: treat the feed as the campaign
The multi-channel pitch (Search, YouTube, Display, Gmail) is mostly marketing for retailers. smec, which manages spend for hundreds of retailers, reports that 74–97% of PMax cost goes to feed-based Shopping-style ads (smec). Your "creative" is your product feed.
That means Shopping has no keywords — the product title is the query-matching surface. Front-load the defining attributes: product type, material, color, size. Keep price and availability matched to the landing page, because a feed-to-site mismatch is the fastest route to a Merchant Center disapproval, and a disapproval on a top-20% SKU silently removes the products carrying the account.
Price competitiveness acts like a hidden bid multiplier. Google compares your price against other sellers of the same item, and an uncompetitive price suppresses impression share no matter how high you bid. Most "feed optimization" posts stop at titles; the price lever is where the impressions actually hide.
Should you run PMax alone? Usually not
The old "PMax replaced Standard Shopping" framing is dead. The two now compete on equal Ad Rank in the same account, and because Google runs a second-price-style auction, running both does not mean bidding against yourself. Optmyzr's study of 24,702 campaigns found 82% of advertisers run PMax alongside Shopping or Search, and split-budget accounts posted the strongest returns in the data (Optmyzr).
A practical use for Standard Shopping is rescuing "zombie" products — SKUs PMax starves of impressions because they have no conversion history, which smec defines as products under about ten clicks a month (smec). Pull them into a clicks-oriented Standard Shopping campaign to force data collection, then graduate the proven ones back. Budget that campaign like R&D — it buys information, not profit. For the full portfolio argument, see our breakdown of the latest Performance Max update.
Know your feed before the first click
Every best practice above assumes you know your true per-order profit. Most store owners do not, because the numbers live in four places: Shopify shows revenue, Google shows ad spend, Printify or Printful shows base cost, and the processing fee hides inside the payout.
This is the job PodVector AI built Victor for. Victor is an AI employee who connects your Shopify store, Google Ads, Meta Ads, and your Printify, Printful, or Gelato supplier accounts, and computes the true profit on every order — product cost, shipping, fees and ad spend netted out. He works inside Google Ads as a full operator: he sets your tROAS target from that real margin, turns on brand exclusions, pauses the SKUs bleeding money, and raises bids on the products that actually clear your break-even.
Because Victor knows your base cost from Printify or Printful before the first sale, he can tell you a product's margin on day one — no sales history required. He delivers the reports to Google Drive and drafts the customer-support emails too. Victor is not a dashboard you log into to read charts; he does the work in the account.
Put Victor on your Google Ads account and see your true per-order profit before you touch another tROAS slider.
If you decide the Shopping side needs a specialist's hand as you scale, our guide to choosing a Google Shopping ads agency walks through what to ask.
FAQs
What is the single most important Performance Max best practice?
Deriving your tROAS target from break-even math instead of a benchmark. Break-even ROAS is 1 ÷ contribution margin, so a 40% margin means a 2.5x break-even. Setting the target above what your account has demonstrated suppresses volume rather than creating efficiency (Google Ads Help). Everything else — feed, brand exclusions, assets — amplifies a target that is already correct.
How many conversions does Performance Max need to work?
Google's technical floor for Target ROAS on Search and Shopping is 15 conversions with values in the past 30 days (Google Ads Help). The practitioner floor for stable behavior is higher — roughly 50 conversions a month is the widely repeated field standard (Search Engine Land). For PMax specifically, 30–50 conversions per campaign per month is the convention for exiting learning cleanly. Below that, expect volatility.
Is Performance Max still a black box?
Not since 2025. PMax now has a channel performance report breaking spend across Search, Display, YouTube and the rest, full search-terms reporting, and campaign-level negative keywords (Google Ads Help). Any advice telling you "you can't see inside PMax" is outdated. Read the channel report to learn what share of your conversions came from Search versus Display.
Why does my PMax ROAS look great but my sales are flat?
That is the signature of brand cannibalization. PMax absorbed branded-search, organic and direct orders you were already getting, reports them as its own, and shows a spectacular ROAS while adding zero incremental sales. Field estimates put the apparent-ROAS inflation at 15–30% in unprotected accounts (GrowthSpree). Turn on brand exclusions and judge PMax on non-brand performance.
Should I chase a higher impression share?
No. Impression share is a diagnostic, not a goal. Chasing it buys auctions your tROAS math already rejected. When impression share is low, first split "lost to budget" from "lost to rank" — they have opposite fixes, budget versus bid, quality and price. For Shopping specifically, low impression share with a healthy budget usually traces to price competitiveness or feed quality, not bids.
Does rising CPC mean I'm doing something wrong?
Often it is the market, not your account. WordStream's 2025 data found about 87% of industries saw year-over-year CPC increases, averaging around a 13% rise (Search Engine Land). If your CPC climbed while CTR and conversion rate held flat, that points to auction pressure, not a campaign problem. Check Auction Insights for new competitors before you start changing settings.