Performance Max (PMax) is Google’s fully automated campaign type spanning Search, Display, YouTube, Gmail, Discover, and Maps. This guide covers when to use it, how to structure asset groups, and how to maintain control in an AI-driven campaign.

→ Performance Max dictionary · Google Ads Glossary · Smart Bidding

Factor Performance Max Search Campaign
Control Low (ML decides placement) High (keywords, match types)
Placements All Google inventory Search results only
Best for E-commerce feeds, full-funnel High-intent keywords
Min. conversions ~30/month recommended ~15/month for tCPA
Reporting Limited granularity Full keyword/search term data

Rule: Run Search for keywords you know convert. Use PMax for incremental reach and Shopping visibility.

Performance Max Campaigns Guide (2026): Setup, Assets & Optimization

Campaign setup checklist

1. Prerequisites

  • Conversion tracking verified (Purchase with value)
  • Google Merchant Center linked (for retail)
  • 30+ conversions/month in account (or expect long learning phase)
  • Brand Search campaign running separately

2. Campaign settings

  • Goal: Sales or Leads
  • Bidding: Start with Maximize Conversion Value → switch to tROAS after 30 conversions
  • Budget: ≥ 3× your target CPA daily (minimum for learning)
  • Location: Nepal only, or expand if shipping allows
  • Brand exclusions: Add your brand terms

3. Asset groups (minimum for “Excellent” ad strength)

Asset type Minimum Best practice
Headlines 5 10–15 varied angles
Long headlines 1 5
Descriptions 4 5
Images 5 10+ (1.91:1, 1:1, 4:5)
Logos 1 2
Videos 0 (optional) 1+ (even simple slideshow)
Final URL 1 Match landing page to audience

Nepal tip: Include NPR pricing, “Free delivery in Kathmandu”, and wallet payment icons (eSewa/Khalti) in descriptions.

Audience signals (not targeting — hints to ML)

Add as signals, not restrictions:

  1. Custom segments: URLs of competitor sites, search terms your customers use
  2. Your data: Customer list (email upload), website visitors, purchasers
  3. Demographics: Start broad; let ML optimize

Signals improve cold-start; they do not limit reach like traditional targeting.

Bidding strategy progression

Week 1–2: Maximize Conversion Value (gather data)
Week 3+:  tROAS at break-even + 20% buffer
Month 2+: Tighten tROAS if volume stable

Break-even tROAS = 1 ÷ Gross Margin. Example: 40% margin → break-even = 250% → target tROAS 300%.

→ ROAS Calculator

Retail / Shopping PMax

  1. Link Merchant Center with complete product feed
  2. Enable Listing Groups in asset group — segment by brand, category, or custom label
  3. Use feed-only asset group for pure Shopping if creative is weak
  4. Exclude out-of-stock via Merchant Center rules

Reporting and optimization

Report Location Action
Asset performance Asset group → Assets Replace “Low” rated assets
Search terms insight Insights → Search terms Add negatives to brand Search, not PMax
Placement report Limited Monitor Display/YouTube spend share
Location report Locations Adjust bids for Kathmandu vs. other cities

Weekly routine:

  1. Check asset ratings — replace poor performers
  2. Review search terms insight — feed negatives to Search campaigns
  3. Compare PMax ROAS vs. Search ROAS — reallocate budget if Search outperforms on branded terms

Brand cannibalization prevention

PMax will bid on your brand name unless excluded:

  1. Campaign settings → Brand exclusions → add brand list
  2. Run Branded Search campaign with exact match [your brand]
  3. Monitor Search terms insight for brand leakage weekly

When NOT to use PMax

  • New account with <15 conversions/month
  • No conversion value tracking
  • Need keyword-level control for B2B niche terms
  • Very small budget (< NPR 500/day) — insufficient learning data

Nepal e-commerce example structure

Account
├── Search: Brand (exact match)
├── Search: Non-brand high-intent ("buy [product] Nepal")
├── PMax: All products (tROAS 350%)
│   └── Asset group: Bestsellers (custom label)
│   └── Asset group: New arrivals
└── Display Remarketing (optional, if PMax budget limited)