Definition

Performance Max (PMax) is a Google Ads campaign type that uses machine learning to serve ads across all Google inventory — Search, Display, YouTube, Gmail, Discover, and Maps — from a single campaign with automated bidding and creative assembly.

Detailed Explanation

You provide assets (headlines, descriptions, images, videos, URLs) and conversion goals; Google’s AI combines them and allocates budget to highest-performing placements. PMax replaced Smart Shopping in 2022 and is now default for many e-commerce advertisers.

Key levers: asset groups, audience signals, listing groups (for retail), URL expansion on/off, and brand exclusions. Reporting granularity is limited compared to standard Search campaigns — use asset group and placement insights.

Nepal Context

Nepali e-commerce brands with product feeds can run PMax for Shopping visibility. Ensure Merchant Center is linked and product data includes NPR prices. Start with tROAS goals based on break-even margin; review the search terms report and add campaign-level negative keywords where needed (both were added to PMax in 2025), and watch for brand query leakage. Most service SMEs should wait until Search conversions are real — PMax go/no-go for Nepal. Protect brand: brand Search.

Practical Examples

  1. Single asset group with 5 headlines, 5 descriptions, 10 images, 1 video — minimum for good ad strength.
  2. Add customer list as audience signal for remarketing-weighted PMax.
  3. Exclude brand terms via campaign-level brand restrictions to protect branded Search campaigns.

How to Check It

  1. Review the Insights page and search terms report for irrelevant or branded queries.
  2. Check Asset group reporting for assets rated Low and replace them.
  3. Compare PMax conversions with your backend orders or leads, since PMax can claim credit for brand and remarketing sales.

Key Takeaways

  • PMax = automated cross-channel campaigns; less manual control, more ML dependency.
  • Strong assets and conversion tracking are prerequisites.
  • Supplement with standard Search for high-intent keyword control.

Common Mistakes

  1. Launching PMax with weak conversion volume — needs ~30 conversions/month for stable learning.
  2. No brand exclusions — PMax cannibalizes cheaper branded Search clicks.
  3. Identical assets across all asset groups — limits ML optimization signals.