When Google first introduced Performance Max (PMax), it was met with a mixture of awe and profound frustration. Advertisers were handed a powerful engine but were locked out of the controls. For years, PMax operated as a notorious black box. It was a campaign type that generated incredible top-line numbers but often cannibalized branded search, spent budget on obscure placements, and offered very little transparency.

The landscape has dramatically shifted since then. After years of feedback from media buyers, Google has steadily opened up the system. Today, Performance Max has evolved into a highly sophisticated, multi-channel tool that requires active management. Advertisers must actively pilot the system to achieve the best results.

Today, you must understand how to manipulate the inputs. While PMax still doesn’t allow full control over exactly where and when your ad placements show, it does have 10 powerful levers to guide performance. Here is a deep dive into those critical campaign-level settings you can use to optimize Performance Max, reduce wasted spend, and drive profitable growth.

1. Conversion Goals: Tell the Machine Exactly What Matters

The absolute foundation of a successful PMax campaign is your conversion tracking. PMax is a goal-based campaign. The algorithm will ruthlessly hunt for whatever action you tell it to value. High-quality data inputs are required for optimal performance.

Advertisers need to be surgically precise by focusing on bottom-of-funnel actions.

For Ecommerce: Ensure your primary goal is strictly set to “Purchases” with dynamic conversion values attached. You want the algorithm optimizing for actual revenue generation.

For Lead Generation: Optimizing for basic form submissions often yields poor results because PMax will scrape low-quality ad placements for cheap clicks. Advertisers must implement Offline Conversion Tracking (OCT) to focus on deeper funnel actions. By integrating your Google Ads account directly with your CRM software like Salesforce or HubSpot, you can feed data back to the algorithm when a lead actually closes. Setting your PMax conversion goal to “Sales Qualified Lead” or “Closed Won Deal” forces Google to look for users who match the profile of your actual paying clients.

Refining your Conversion Goals at the campaign level instantly upgrades the quality of the audience the AI targets.

2. Budget and Bidding Strategy: Fueling the Engine

Your bidding strategy dictates how aggressive PMax will be in the auction. You generally have two paths: Maximize Conversions (with an optional Target CPA) or Maximize Conversion Value (with an optional Target ROAS).

Setting overly restrictive targets early on can limit the campaign’s ability to learn and spend. The algorithm requires liquidity and data volume to build its predictive models. A best practice for launching a new PMax campaign is setting a Target ROAS or Target CPA 10% to 15% above your actual goal.

After allowing the campaign to gather volume for two full weeks, you can gradually adjust the target by small increments of 5% at a time. This methodical approach helps you reach your ideal efficiency range without shocking the algorithm and halting your ad delivery.

3. New Customer Acquisition (NCA) Rules

One of the most powerful updates to the PMax ecosystem is the New Customer Acquisition setting. Early iterations of PMax often prioritized retargeting past purchasers to inflate ROAS metrics artificially.

Today, you can force the algorithm to prioritize genuine business growth. You need a robust First-Party audience list uploaded via Customer Match to define your existing customer base. Once Google processes your customer list, you have two specific options:

• Bid Higher for New Customers: PMax will still serve ads to past buyers, but it will bid more aggressively when it spots a user who has never bought from you.
• Only Bid for New Customers: This strict exclusion limits ad delivery exclusively to cold traffic. It is ideal for aggressive acquisition campaigns. It will typically result in a lower overall ROAS since you are removing easy repeat purchases from the data pool.

4. Campaign-Level Negative Keywords

Applying negative keywords to PMax previously required backend assistance from Google representatives. Now, campaign-level negative keywords are fully accessible and serve as your best defense against wasted spend.

B2B software companies must apply negative keywords like “free,” “student,” or “login” to filter out unqualified traffic looking for customer support or free trials. Businesses selling high-end auto detailing supplies to commercial buyers must block consumer terms like “auto parts near me” or “single car wash towel.” Building a robust negative keyword list forces PMax out of irrelevant search auctions and improves your click-through and conversion rates immediately.

5. Brand Exclusions: Protecting Your Turf

Brand Exclusions specifically prevent PMax from bidding on your branded search traffic. Left unchecked, PMax will recognize that bidding on your exact brand name is the easiest way to generate cheap conversions.

Applying a Brand Exclusion list at the campaign level ensures the PMax algorithm focuses entirely on finding new top-of-funnel traffic. Advertisers can then capture their brand searchers efficiently in a separate, tightly controlled Standard Search campaign. This guarantees you are not paying premium PMax rates for customers who were already actively looking for your specific website.

6. Final URL Expansion: The Double-Edged Sword

Final URL Expansion allows Google’s AI to look beyond the specific landing pages provided in your Asset Groups. If the algorithm believes a user’s search query is highly relevant to another page on your site, it will dynamically generate a headline and send the user there.

Leaving Final URL Expansion enabled allows Google’s AI to act similarly to a Dynamic Search Ad. This is highly effective for massive ecommerce sites with thousands of categories. It helps catch long-tail queries that you might not have explicitly targeted.

Conversely, advertisers driving traffic to a specific funnel or promotional landing page should disable this setting. Disabling URL Expansion keeps traffic focused on the intended conversion path and prevents Google from sending paid clicks to a privacy policy or a generic contact page.

7. Automatically Created Assets (ACA)

The Automatically Created Assets setting allows the system to scrape your landing pages and generate its own text headlines, descriptions, and video assets.

Allowing Google to create text assets can be useful for scaling. The system will dynamically create headlines that exactly match a user’s search query to boost ad relevance scores.

Advertisers must provide their own high-quality video creative. If no video is uploaded and the video ACA setting is enabled, Google will automatically stitch static images together with royalty-free music. These auto-generated videos often look unprofessional and perform poorly on YouTube. Disabling auto-generated videos ensures better control over brand messaging and ad quality.

8. Location and Language Targeting

Because PMax spans the entire Google network, a small targeting error can result in significant spend in the wrong region.

Advertisers should verify their specific location options within the campaign settings. Selecting “Presence: People in or regularly in your targeted locations” provides tighter geographic control than the default interest-based targeting. Actively using location exclusions also prevents ad spend in regions you cannot service profitably due to logistics, shipping constraints, or regional legal compliance.

9. Ad Scheduling (AKA Dayparting)

While PMax thrives on continuous data, certain business models require strict Ad Scheduling.

B2B service providers relying on inbound phone calls can use Ad Scheduling to restrict delivery strictly to actual business hours. There is zero value in paying for clicks at 3:00 AM if your sales team is not available to answer the phone.

Ecommerce brands running campaigns continuously can analyze historical data to identify low-converting hours. If the data shows a massive drop in conversion rates between midnight and 5:00 AM, advertisers can restrict bidding during those specific times to reallocate budget to peak evening shopping hours.

10. Product Feed Integration and Segmentation

For ecommerce brands, the Google Merchant Center feed is the beating heart of Performance Max. Segmenting large product catalogs is crucial for PMax success. The algorithm naturally favors the path of least resistance. If you place 5,000 products into a single campaign, the system will find 10 easy-to-sell products, spend the entire budget on them, and ignore the remaining 4,990 items.

Advertisers must group products by business metrics using Custom Labels within their Merchant Center feed.

A highly effective strategy involves mapping out your inventory using Custom Label 0 for profit margins. You can tag items as “High Margin,” “Medium Margin,” and “Low Margin.” You then create three separate Performance Max campaigns.

The “High Margin” campaign can be set with a lower Target ROAS to push aggressive volume. The “Low Margin” campaign must have a strict, high Target ROAS to protect business profitability.

You can use Custom Label 1 for inventory levels. Tagging overstocked items allows you to build a dedicated clearance campaign specifically for liquidating that stock quickly. Filtering your feed at the campaign level through these labels forces the algorithm to distribute spend exactly where your business needs it most.

Your Next Steps

Performance Max is undoubtedly the most powerful automated buying tool Google has ever created. Advertisers must guide the automation to align with their specific business goals. Automation does not mean you surrender control of your budget.

Mastering these 10 levers empowers you to reclaim control over the machine learning algorithm. By optimizing Conversion Goals, Bidding, NCA, Negatives, Brand Exclusions, URL Expansion, ACA, Locations, Scheduling, and Feed Segmentation, you transform PMax into a highly tuned, revenue-generating machine. Remember, the advertisers who win are the ones who know exactly how to manage the inputs!