Loading tutorials…
Loading tutorials…
Performance Max is Google's most opaque ad product — and for accounts with the right ingredients, also one of the highest-ROAS. The catch: every default Google offers is tuned for Google's revenue, not yours. Here is the build that gives you back control.
Who this is forOwners with $3K+/mo Google Ads spend, working conversion tracking, and either an e-commerce catalog or a strong landing-page asset library. PMax is not for brand-new accounts — it needs 30+ days of conversion history to perform.
What you'll need
Step 1
PMax needs 30+ conv/mo, a clear conversion value, and either an e-com feed or rich creative assets. Without these, it will underperform a focused Search campaign.
PMax minimum requirements before it makes sense to run:
☐ 30+ conversions per month in your account, all sources combined. Below this, the algorithm cannot find your conversion patterns and PMax behaves like a generic spending campaign.
☐ A primary conversion action with a consistent value (e.g., purchase revenue, lead score). PMax does not optimize well against pure event counts — it needs value signals.
☐ Either a Google Merchant Center product feed (e-com) OR a strong asset library (10+ high-res images, 5+ logos in different ratios, 5+ short videos, 5+ headlines, 5+ descriptions, sitelinks, callouts).
☐ Existing Search and Shopping campaigns running, so you have baselines to compare PMax against and so PMax has somewhere to send its overflow learning.
☐ At least $1K/mo committed to the PMax campaign. Below that, the algorithm spreads spend too thin across Search, Display, YouTube, Discover, Gmail, and Maps inventory.
If any of these are missing, focus on Search campaigns first. PMax will still be available in 60-90 days when the prerequisites are in place.
Step 2
Campaigns → "+ New campaign" → Sales (or Leads). Choose "Performance Max." Set ONE primary conversion goal — secondary goals dilute PMax optimization more than in any other campaign type.
Click Campaigns → "+ New campaign." Pick the objective: Sales (for e-com), Leads (for lead-gen), or Local store visits (for physical locations). Avoid "Without a goal" for PMax — the campaign objective shapes what inventory PMax can access.
Pick Campaign type: Performance Max.
On the conversion goals screen, override Account-default Goals if your account has multiple conversion actions. Set ONE primary action for this PMax campaign. Mark others as Secondary or exclude them.
Why this matters: PMax optimizes hardest toward whatever it can find. If you let it chase both newsletter signups AND purchases as primary actions, it will overspend on the cheap, easy-to-find signups and starve the purchases that actually drive revenue.
Name the campaign with the structure "PMax-[Region]-[Vertical]-[Date]" so it is easy to identify later among your other campaigns.
Step 3
Bidding: Maximize Conversions (no target) for first 30 days, then Target CPA or Target ROAS. CRITICAL: turn OFF "Final URL expansion" or PMax will route traffic to arbitrary pages on your site.
Bidding: Maximize Conversions with NO Target CPA for the first 30 days. PMax has a longer learning period than Search — usually 21-30 days. Target during learning prevents the algorithm from finding the conversion baseline.
After 30 days of unrestricted spend, switch to Target CPA (for cost-controlled) or Target ROAS (for revenue-controlled). Start the target at 1.25-1.5x your baseline, never tighter.
Daily budget: minimum $50/day ($1.5K/mo) for PMax to spread effectively across the inventory mix. Below this, PMax under-serves Search inventory where most conversions come from.
CRITICAL setting — "Final URL expansion": this is buried in the campaign settings under "Additional settings." Default is ON, meaning PMax can send traffic to ANY page on your domain it thinks is relevant — including blog posts, About pages, and orphaned drafts.
Turn Final URL expansion OFF for any campaign where you have specific landing pages. Then add the specific URLs as Final URL exclusions or use 'Page feeds' to constrain PMax to a defined set of URLs.
Customer acquisition: under "Audience-related settings," choose "Bid higher for new customers" if you sell repeat-purchase products and want PMax to focus on net-new buyers. This is a recent setting most older guides do not mention.
Step 4
Asset groups are PMax's ad creative containers. Build one per audience theme — "First-time buyers," "Loyalty repeat," "Premium seekers." Each asset group needs its own creative + audience signals.
Asset groups are the closest thing PMax has to ad groups. Each asset group bundles: a set of creative assets (images, videos, headlines, descriptions, sitelinks) + audience signals + listing groups (for e-com).
Build asset groups around audience themes, not product themes. Example: "First-time buyers" (asset group 1) with creative that emphasizes onboarding/free shipping/easy returns, vs. "Loyalty repeat" (asset group 2) with creative that emphasizes new product launches or exclusive access.
Per asset group, provide: 5+ headlines (30 char max each), 5+ long headlines (90 char max), 5+ descriptions (90 char max), 1+ short description (60 char max), 4+ landscape images (1200x628 or 1.91:1), 4+ square images (1200x1200 or 1:1), 1+ portrait image (960x1200 or 4:5), 5+ logos in different ratios, 1+ video (Google will auto-generate one from your images if you do not provide one — provide your own for better quality).
Audience signals: tell PMax who the asset group is FOR. Click "+ Add audience signal" and add: Customer Match lists (existing customers, abandoned cart, high-value leads), Custom segments (people searching for specific terms or visiting specific competitor sites), Demographics, Interests, and Detailed demographics that match the theme.
Audience signals are NOT targeting in PMax — they are hints to the algorithm about who to find. PMax will still expand beyond them, but they dramatically improve learning speed.
Build 2-4 asset groups per PMax campaign. More than 4 fragments your creative testing; fewer than 2 means the algorithm has no theme variation to optimize across.
Step 5
For e-com: in each asset group, set up Listing groups to control which Merchant Center products PMax can advertise. For lead-gen: use Page feeds to constrain PMax to specific URLs.
E-COMMERCE — Listing groups:
If you have a Merchant Center feed, each asset group can have a "Listing groups" section that controls which products PMax advertises.
By default, listing groups include "All products." This is fine if your catalog is uniform-quality, but most e-com sites have hero SKUs, mid-tier SKUs, and clearance SKUs that should be optimized separately.
Split your listing group by Brand, Category, Product Type, or Custom Label (configured in Merchant Center). Asset groups with high-margin products should bid more aggressively than asset groups with clearance items.
LEAD-GEN — Page feeds:
Tools (wrench icon) → Setup → Business data → Page feeds. Create a new page feed with the specific landing page URLs PMax should drive traffic to.
Attach the page feed to the campaign (Settings → Page feeds → select the feed).
This is the lead-gen equivalent of constraining PMax to specific products — without it, PMax will route traffic to whatever pages on your site it thinks are relevant.
Step 6
Search themes (per asset group): give PMax explicit search intent hints. Brand exclusions (campaign-level): keep PMax out of brand searches that your Search campaign already covers.
Search themes — per asset group, add 5-15 search themes (each up to 4 words). These are explicit hints to PMax about which search intents this asset group serves.
Examples for an e-com running shoes business: "running shoes for women," "trail running shoes," "marathon training shoes," "best running shoes 2026."
Search themes do NOT work like keywords — PMax uses them as hints, not strict matches. But strong search themes meaningfully improve the Search portion of PMax delivery.
Brand exclusions — Campaign settings → Brand exclusions. Add your own brand name as an exclusion to PMax so it does not bid on brand searches that your dedicated brand Search campaign already handles. This prevents PMax from cannibalizing cheap brand clicks and claiming credit for them.
Account-level Brand exclusion lists (Tools → Brand exclusions) let you maintain these centrally across campaigns. Set this up once, apply to every PMax campaign.
Add negative keywords to the PMax campaign — Settings → Negative keywords. PMax now accepts negative keyword lists (this was rolled out in late 2025). Apply the same negative keyword lists you use on Search campaigns to prevent obvious junk queries.
Step 7
After launch, do not touch the campaign for 14 days. Then check Insights & reports → Insights tab weekly for asset performance, audience expansion, and search categories.
Click "Publish campaign." PMax campaigns usually take 6-12 hours to go through asset and policy review before serving begins.
Do not change ANYTHING for the first 14 days. Bidding adjustments, audience changes, asset swaps — all of them restart the learning period. PMax learning takes 21-30 days at minimum.
After day 14, start a weekly Insights review: Insights & reports → Insights tab. Look at:
— Asset performance ratings ("Best," "Good," "Low"): replace any "Low" assets with new variants. Google does not tell you exact CTR per asset (this is a PMax opacity issue) but the ratings are directionally correct.
— Audience insights: which audience signals are converting. Strengthen those, deprioritize others.
— Search categories: this is the closest PMax has to a Search Terms report. It shows broad categories of searches you matched. Use this to update Search themes and brand exclusions.
— Conversion path: where the conversion came from in PMax's mix. If you are getting 90% of conversions from Search inventory, consider whether you should be running Search instead and saving PMax for fresh customer acquisition.
Monthly cadence: refresh assets, update audience signals based on Customer Match lists, and tighten Target CPA/ROAS by 5-10% if performance is hitting target consistently.
Common mistakes
Running PMax with Final URL expansion ON
What goes wrong: PMax routes traffic to arbitrary pages on your site — blog posts, About pages, draft pages that should not get traffic. Conversion rate from those pages is 0-5% of your real landing pages. Wastes $300-1,500/wk for a $5K/mo campaign with no obvious failure signal in the UI.
How to avoid: Settings → Additional settings → Final URL expansion → OFF. Use Page feeds (lead-gen) or Listing groups (e-com) to define exactly which destinations PMax can use.
Not excluding brand searches from PMax
What goes wrong: PMax bids on your brand searches, wins them at low CPC, takes credit for the conversion, and reports a great ROAS. Meanwhile, your dedicated brand Search campaign is starving for impressions and the apparent PMax performance is mostly inherited brand traffic.
How to avoid: Settings → Brand exclusions → add your brand. Better yet, set up an account-level Brand exclusion list (Tools → Brand exclusions) and apply it to every PMax campaign.
One giant asset group with everything in it
What goes wrong: When all 30 images, 15 headlines, and 5 videos are in a single asset group, PMax cannot test variation across audiences. Asset reporting is flattened — you cannot tell which creatives work for which audiences. Learning is slow because the algorithm has no theme structure.
How to avoid: Build 2-4 asset groups per campaign, each themed around a distinct audience or intent. Each asset group gets its own audience signals, creative variants, and (for e-com) listing group.
Tight Target CPA from day one
What goes wrong: PMax learning period is 21-30 days — longer than Search. A tight Target CPA before day 30 prevents the algorithm from exploring inventory. The campaign appears 'Limited by bid strategy' and spends 20-30% of budget. You think PMax does not work; actually you choked it.
How to avoid: Maximize Conversions (no target) for the first 30 days. Then add Target CPA at 1.25-1.5x baseline. Tighten by 5-10% every 2 weeks once stable.
Skipping audience signals because PMax does not require them
What goes wrong: PMax with no audience signals takes 30-60 days to find its conversion patterns instead of 14-21 days. You burn 2-3 weeks of budget on exploratory delivery that could have been informed.
How to avoid: Always add audience signals per asset group: Customer Match lists, Custom segments, Demographics, Detailed demographics. They are hints, not targeting — but strong hints save weeks of learning time.
Mixing high-margin and low-margin products in one listing group
What goes wrong: PMax allocates spend toward whatever converts easiest. If your $5 clearance items convert at 8% and your $80 hero items convert at 2%, PMax will starve the hero items because the easy money is in clearance — even though the clearance has zero profit margin.
How to avoid: Use Custom Labels in Merchant Center to tag products by margin tier or strategic priority. Split asset groups by label and set different bid emphasis per asset group.
Recap
Done — what's next
How to create your first Google Ads Search campaign
Read the next tutorial
Hand it off
PMax is the campaign type where the gap between DIY and specialist performance is widest. The defaults are tuned for Google's revenue. The good settings are buried. The reporting is opaque. A specialist who has run PMax across 50+ accounts knows which campaign exclusions, asset group structures, and audience signal combinations actually work. EverestX Google Ads specialists typically run $400-1,200/mo for accounts with active PMax campaigns, at $14-16/hr.
See specialist rates
No — PMax works best as an addition to Search, not a replacement. Search campaigns give you keyword-level control and visibility that PMax does not. Most well-run accounts have 60-70% of spend in Search/Shopping and 30-40% in PMax, with PMax used for net-new customer acquisition and inventory PMax has unique access to (YouTube, Discover, Gmail).
PMax intentionally hides Search Terms — it is Google's most aggressive 'trust the algorithm' product. You can see broad search categories under Insights & reports → Insights, and Search themes give you hints about intent, but query-level data is not available. This opacity is the main reason PMax is harder to manage than Search.
Minimum $1,500/month dedicated to the PMax campaign. Below that, PMax cannot spread effectively across Search + Display + YouTube + Discover + Gmail + Maps inventory and learning stalls. Most accounts running PMax effectively are at $3K+/month in that single campaign.
21-30 days for the initial learning period (longer than Search). Significant changes during learning (bidding, asset groups, audience signals) restart the clock. Major performance changes are normal in days 7-21 — do not over-react. By day 30, the campaign should be in stable performance.
Technically yes, but you lose the theme variation that helps PMax learn. The minimum recommended structure is 2 asset groups — one for primary intent and one for secondary intent or audience segment. Three to four asset groups is the sweet spot for most accounts.
Yes — Customer Match lists are the single most powerful audience signal for PMax. Upload your existing customer list and use it as a positive signal for retention asset groups, or use it as a negative for new-customer-only asset groups. Refresh the list at least monthly.
Google Ads
A Search campaign is the highest-intent ad format Google offers — people typing exactly what they want, right when they want it. Setting it up well is also the workflow Google works hardest to flatten into 'just trust the algorithm.' Here is the real build.
Google Ads
Remarketing is the lowest-friction conversion lift in the platform — but only when the audiences are built cleanly. Stale lists, leaky audience definitions, and missing exclusions kill the lift. Here is the setup specialists actually use.
Google Ads
Conversions firing in your store but Google reporting zero? You're not alone. This walks through the right setup path for your stack — including the GA4 import workflow that 80% of DIY tutorials skip.
Google Ads
CPC creeping up while conversion rate stays flat means you're paying more for the same outcome. These six tactics drop CPC 20-40% on most accounts. Some take 1 hour. Some take 1 month. All matter.
Google Ads
DIY Google Ads is a great idea — until it isn't. This is the honest framework: when the cost of self-managing exceeds the cost of hiring, and how to tell which side you're on.