Most brands build Performance Max campaigns inside Google Ads. They click "create asset group," dump in some headlines, upload whatever images they have, and hope the algorithm figures it out.
That's backwards.
The architecture decisions you make before you open Google Ads determine 80% of your campaign's performance. How many asset groups. Which products go where. What headlines and images each group gets. Which audience signals to load.
We've built or restructured 110+ Performance Max campaigns across 20+ brands. The pattern is consistent: brands that plan their asset groups on paper first and then build in Google Ads outperform brands that build on the fly.
This worksheet walks you through every decision. Fill it out completely before creating anything in your Google Ads account.
Before you decide how many asset groups to create, you need to know whether your account has enough conversion data to support them.
Performance Max is a machine learning system. Each asset group needs conversion signals to learn which headlines, images, and audiences perform best. Split your data too thin and the algorithm is guessing.
Your numbers:
| Metric | Your Value | Notes |
|---|---|---|
| Monthly PMax conversions (last 30 days) | _______ | Pull from Google Ads > Campaigns > PMax > Conversions |
| Monthly account conversions (all campaigns) | _______ | Total across all campaign types |
| Monthly PMax budget | $_______ | Current or planned |
| Average order value | $_______ | From your analytics |
The threshold table:
| Monthly PMax Conversions | Maximum Asset Groups | Why |
|---|---|---|
| 30-50 | 1-2 | Keep data concentrated. Splitting further starves each group. |
| 50-100 | 2-3 | Enough data to split meaningfully by category or intent. |
| 100-200 | 3-5 | Can segment by category, price tier, or margin tier. |
| 200+ | 5-8 | Full segmentation possible across multiple dimensions. |
Your recommendation: _______ asset groups maximum
If your monthly conversions are below 30, Performance Max probably isn't the right campaign type yet. The algorithm doesn't have enough data to make intelligent bidding decisions below that threshold. Standard Shopping and Search campaigns with manual bidding give you more control until you build the conversion history.
Minimum budget check: Each campaign needs at least $50/day to give the algorithm enough daily budget to explore and learn. If your total PMax budget can only support 2 campaigns at $50/day each, run 2 campaigns - not 3 underfunded ones.
Now decide how to split your products across asset groups. The segmentation criteria you choose determines your ability to write targeted headlines, assign relevant audience signals, and set appropriate bid strategies.
Segmentation options (in order of effectiveness):
| Criteria | Best For | Example |
|---|---|---|
| Product category | Catalogs with 3+ distinct categories | Pillows, Mattress Toppers, Duvets |
| Price tier | Wide price range ($20 product alongside $200 product) | Budget (<$50), Mid ($50-$150), Premium ($150+) |
| Margin tier | Variable margins across catalog | High margin (>60%), Standard (40-60%), Low (<40%) |
| Seasonality | Mix of evergreen and seasonal products | Evergreen, Summer, Winter, Holiday |
Your segmentation plan:
| Asset Group # | Name | Products Included | Product Count | Estimated Monthly Conversions |
|---|---|---|---|---|
| 1 | ________________ | ________________ | _______ | _______ |
| 2 | ________________ | ________________ | _______ | _______ |
| 3 | ________________ | ________________ | _______ | _______ |
| 4 | ________________ | ________________ | _______ | _______ |
Critical rule: Each product should appear in exactly one asset group within a campaign. Overlapping products forces them to compete against each other in the auction, driving up your cost per click.
Verify your split: Look at the "Estimated Monthly Conversions" column. If any asset group has fewer than 15 estimated monthly conversions, merge it with another group. A group with too few conversions will stay in LEARNING mode indefinitely.
Each asset group needs its own headline set. Google allows up to 15 short headlines (30 characters each) and 5 long headlines (90 characters each). The minimum is 3 short and 1 long.
Don't use the same headlines across asset groups. Each group has different products, different buyers, and different search intent. Generic headlines like "Shop Now" or "Best Products" get rated LOW by Google's performance labeling system. Accounts with generic headlines have 70-80% of assets rated LOW. Accounts with specific, targeted headlines have 60-70% rated GOOD or BEST.
The difference in practice: 0.3-0.5% CTR with generic headlines versus 0.8-1.2% CTR with specific headlines. On a $10K/month budget, that's 300 clicks versus 900 clicks for the same spend.
For each asset group, write headlines using these six types:
| Type | When to Use | Example |
|---|---|---|
| Bold Call-Out | Benefit-first, cold traffic | "Monday to Friday, One Charge" |
| Statistic | Product has a standout metric | "40 Hours. One Charge." |
| Comparison | Audience has tried alternatives | "6 Pillows in 2 Years. This One Stayed." |
| Customer Quote | Strong review language available | "Why Didn't I Buy This Sooner" |
| Question | Problem-based angles | "Still Waking Up With Neck Pain?" |
| Offer/Price | Price is the value proposition | "40% Off This Week Only" |
Asset Group 1: ________________
| # | Headline (30 chars max) | Type | Char Count |
|---|---|---|---|
| 1 | |||
| 2 | |||
| 3 | |||
| 4 | |||
| 5 | |||
| 6 | |||
| 7 | |||
| 8 | |||
| 9 | |||
| 10 | |||
| 11 | |||
| 12 | |||
| 13 | |||
| 14 | |||
| 15 |
Asset Group 2: ________________
| # | Headline (30 chars max) | Type | Char Count |
|---|---|---|---|
| 1 | |||
| 2 | |||
| 3 | |||
| 4 | |||
| 5 | |||
| 6 | |||
| 7 | |||
| 8 | |||
| 9 | |||
| 10 | |||
| 11 | |||
| 12 | |||
| 13 | |||
| 14 | |||
| 15 |
(Repeat for each additional asset group)
Write 5 long headlines per asset group. Long headlines appear in Display and YouTube placements where you have more space.
Asset Group 1: ________________
| # | Long Headline (90 chars max) | Char Count |
|---|---|---|
| 1 | ||
| 2 | ||
| 3 | ||
| 4 | ||
| 5 |
(Repeat for each additional asset group)
Before you finalize, run each headline through these three tests:
Visualizable: Does the headline create a mental image in under half a second? "6 pillows in 2 years" creates an image. "Superior sleep technology" does not.
Falsifiable: Can the claim be proven or disproven? "2,847 five-star reviews" is falsifiable. "The best pillow ever made" is not.
Unique: Could only your brand say this? "The only retinol that absorbs in 30 seconds" is unique. "High-quality skincare products" could be any brand.
If a headline fails any of these three tests, replace it. Headlines that pass all three consistently earn GOOD or BEST performance labels from Google.
Banned phrases (auto-fail): "best in class," "premium quality," "innovative solution," "next-level," "cutting-edge," "world-class," "state-of-the-art," "industry-leading."
Write 5 descriptions per asset group. 90 characters max each, target 70-85 characters. Each description should take a different angle.
| # | Angle | Description Framework |
|---|---|---|
| 1 | Proof-Led | [Social proof] + [Mechanism hint] + [CTA] |
| 2 | Mechanism-Led | [How it works] + [Differentiator] + [Timeline] |
| 3 | Outcome-Led | [Transformation] + [Specificity] + [CTA] |
| 4 | Social Proof-Led | [Customer language] + [Result] + [Trust signal] |
| 5 | Offer-Led | [Price/deal] + [Risk reversal] + [Ease signal] |
Asset Group 1: ________________
| # | Description (90 chars max) | Angle | Char Count |
|---|---|---|---|
| 1 | Proof-Led | ||
| 2 | Mechanism-Led | ||
| 3 | Outcome-Led | ||
| 4 | Social Proof-Led | ||
| 5 | Offer-Led |
(Repeat for each additional asset group)
Each asset group supports up to 20 images across three aspect ratios.
Required dimensions:
| Aspect Ratio | Dimensions | Count Per Group | Placement |
|---|---|---|---|
| Landscape (1.91:1) | 1200 x 628 px | Up to 8 | Display, YouTube, Discover |
| Square (1:1) | 1200 x 1200 px | Up to 8 | Shopping, Display, Social |
| Portrait (4:5) | 960 x 1200 px | Up to 4 | Mobile placements |
Minimum per group: 3 images (1 landscape, 1 square, 1 portrait). Recommended: fill all 20 slots.
Don't upload 20 versions of the same product-on-white photo. Mix your image types to serve different stages of the buyer's attention.
| Image Type | Count Per Group | Purpose |
|---|---|---|
| Product-on-white | 4-6 | Shopping-style placements, clean comparison |
| Lifestyle/context | 6-8 | Emotional engagement, "see yourself using it" |
| Social proof overlay | 3-4 | Star ratings, review quotes, customer count |
| Problem visualization | 2-3 | Show the pain your product solves |
| Before/After | 2-3 | Transformation proof |
Asset Group 1: ________________
| # | Image Description | Type | Aspect Ratio | Have It? |
|---|---|---|---|---|
| 1 | Product-on-white | Landscape | Y / N | |
| 2 | Product-on-white | Square | Y / N | |
| 3 | Product-on-white | Portrait | Y / N | |
| 4 | Lifestyle | Landscape | Y / N | |
| 5 | Lifestyle | Square | Y / N | |
| 6 | Lifestyle | Landscape | Y / N | |
| 7 | Lifestyle | Square | Y / N | |
| 8 | Lifestyle | Portrait | Y / N | |
| 9 | Social proof | Landscape | Y / N | |
| 10 | Social proof | Square | Y / N | |
| 11 | Social proof | Landscape | Y / N | |
| 12 | Problem viz | Square | Y / N | |
| 13 | Problem viz | Landscape | Y / N | |
| 14 | Before/After | Square | Y / N | |
| 15 | Before/After | Landscape | Y / N | |
| 16 | Lifestyle | Square | Y / N | |
| 17 | Lifestyle | Portrait | Y / N | |
| 18 | Product-on-white | Square | Y / N | |
| 19 | Product-on-white | Landscape | Y / N | |
| 20 | Social proof | Square | Y / N |
(Repeat for each additional asset group)
Image quality rules:
Audience signals are suggestions to Google's algorithm, not hard targeting. Google uses them as starting hints, then expands through machine learning. But the quality of your starting signals significantly affects how quickly the algorithm finds your best audiences.
Signal types to include:
| Signal Type | What to Add | Where to Find It |
|---|---|---|
| Custom segments (URLs) | Competitor website URLs (5-10) | List your direct competitors' homepages and product pages |
| Custom segments (search terms) | Queries your customers actually search | Google Ads search terms report, Google Search Console |
| In-market audiences | Google's purchase intent categories | Google Ads > Audiences > Browse > In-market |
| Customer lists | Past purchasers, email subscribers | Export from your email platform or CRM |
Asset Group 1: ________________
Competitor URLs for custom segments:
Search terms for custom segments:
In-market audiences (select 3-5 relevant categories):
Customer lists to upload:
(Repeat for each additional asset group)
Search themes tell Google what types of queries your asset group should appear for. Load 10-30 per asset group. More than 30 dilutes the signal. Fewer than 10 leaves money on the table.
Theme tiers by intent level:
| Tier | Intent | Budget Share | Example Themes |
|---|---|---|---|
| Tier 1 (HIGH) | Transactional | 50% | "[product] buy," "[product] price," "[product] shop" |
| Tier 2 (MEDIUM) | Commercial investigation | 30% | "best [category]," "[product] review," "[competitor] alternative" |
| Tier 3 (OPPORTUNITY) | Long-tail intent | 20% | "where to find [product]," "[product] in stock," "[product] delivery" |
Asset Group 1: ________________
| # | Search Theme | Tier |
|---|---|---|
| 1 | ||
| 2 | ||
| 3 | ||
| 4 | ||
| 5 | ||
| 6 | ||
| 7 | ||
| 8 | ||
| 9 | ||
| 10 | ||
| 11 | ||
| 12 | ||
| 13 | ||
| 14 | ||
| 15 |
(Repeat for each additional asset group)
Deduplication rule: If you're running multiple campaigns (Feed-Only + Demand Gen + Remarketing), never load the same search theme in two campaigns. Feed-Only gets Tier 1 themes. Demand Gen gets Tier 2-3 themes. Remarketing gets minimal themes - audience signals do the targeting.
Before you build anything in Google Ads, verify these items are complete.
This worksheet handles the planning phase. The complete system includes:
The Google Ads AI Agentic System ($4,997+) covers the full architecture, creative, launch, and optimization protocols.
This worksheet is extracted from the Google Ads AI Agentic System, built on 110+ Performance Max campaigns across 20+ brands spending $3K-$500K/month.