Google Performance Max — strategia 2026 z AI assets
Performance Max (PMax) to AI-driven Google Ads campaign — jeden wpis, AI serwuje reklamy across all Google surfaces (Search, Display, YouTube, Discover, Maps, Gmail). W 2026 PMax stał się 60-70% Google Ads spend dla większości firm. Strategy 2026 wymaga rozumienia AI signals — search themes, audience signals, video assets — żeby AI nauczyło się Twoich celów.
TL;DR — Performance Max 2026
| Element | Detale |
|---|---|
| Co to jest | AI-driven campaign across all Google surfaces |
| Channels | Search + Display + YouTube + Discover + Gmail + Maps |
| Min budget | $50/dzień ($1500/mies.) dla learning |
| Optymal budget | $100-500/dzień |
| Asset types | Text + Image + Video + Audience signals |
| Search themes | Up to 50 per asset group (2026 increase) |
| Bidding | Maximize Conversions / Conversion Value (lub tCPA/tROAS) |
| Learning phase | 2-3 tygodnie + 30-50 conversions |
Dlaczego Performance Max dominuje 2026?
Reasons:
- AI improvements — Google's models more sophisticated each year
- Multi-surface — single campaign covers wszystko
- Asset variety — leverages images + video + text simultaneously
- Time efficiency — less manual optimization needed
- Better cross-channel attribution — AI understands user journeys
- Pressure from Meta/TikTok — Google improving competitiveness
Trade-off: less control. Power users miss query-level visibility, exact placement targeting.
PMax vs Search campaigns
When PMax wins
- Broad reach for ecommerce
- New product launch (cast wide net)
- Brand awareness + conversion combined
- Limited ad management time
- Multi-channel exposure (video + search + display)
When Search wins
- High-intent branded keywords
- Limited budget (<$50/dzień)
- Need query-level control
- Exact placement targeting required
- Tight CPA on specific terms
Best practice: Use both. PMax 60-70% budget, Search 30-40% dla branded + high-intent.
Asset groups — fundament PMax
Co to jest asset group?
PMax campaign zawiera asset groups — kontekstualne grupy assetów (text + image + video + audience signal) targetujące specific theme.
Example: Klient sprzedaje Software jako Service. Asset groups:
- Group 1: Email marketing software
- Group 2: Marketing automation
- Group 3: CRM software
- Group 4: Lead generation tools
Każda group ma własne:
- Headlines (30 chars)
- Long headlines (90 chars)
- Descriptions (90 chars)
- Images (landscape + portrait + square)
- Videos
- Search themes
- Audience signals
AI mixes assetów w real-time per user.
Asset group structure
Best practice ~4-6 asset groups per campaign. Zbyt many = AI splits learning, zbyt few = za szeroki targeting.
Search themes — kluczowa zmiana 2026
Co to są search themes?
Wskazówki dla AI: na jakie queries pokazać Twoje ads. Format: phrases, nie exact keywords.
Limit zwiększony do 50 per asset group (2026)
Wcześniej: 25 search themes. W 2026: 50 themes — większa elastyczność.
Examples (good search themes)
- „lead generation software"
- „marketing automation tools"
- „best CRM for small business"
- „email marketing platform"
Tips
- Specific > generic („lead generation software" > „marketing")
- Match user intent — to nie keyword research, to intent expression
- Mix broad + specific (some „software", some „CRM software for SaaS")
- Brand themes dla branded campaigns
Audience signals
Hint dla AI o swojej target audience. NIE strict targeting (PMax expanduje), ale AI uses jako starting point.
Types:
- Customer match — list email/phone existing customers
- Custom segments — keywords + URLs (Google AI builds audience similar do those who searched/visited)
- Detailed demographics — age, gender, parental status
- Interests + affinities — typowo broad
Best practice: Customer match + custom segments based na keywords + URL z Twojej witryny + competitor sites.
Bidding strategies
Maximize Conversions
Best dla:
- Cold accounts (brak history)
- Lead generation
- Pierwsze 2-4 tygodnie
Spends entire budget chasing conversions. Może give szeroki CPA range.
Maximize Conversion Value
Best dla:
- Ecommerce (different product values)
- Z conversion value tracking setup
AI bids more aggressively dla high-value conversions.
Target CPA
Set max CPA. AI bids do ile można i mantain target.
Use case: Lead gen z known LTV. Don't set CPA za niskie zbyt early — limituje delivery.
Target ROAS
Set ROI target (e.g. 300% = $3 revenue per $1 spend).
Use case: Ecommerce z conversion value. Don't set ROAS za wysokie — limituje volume.
Recommended progression
- Week 1-3: Maximize Conversions (no target) — collect data
- Week 4+: Move to Target CPA / ROAS once 30-50 conversions
- Tuning: Adjust target based on actual performance
Asset best practices 2026
Text assets
5+ Headlines (30 chars):
- Brand name
- Key benefit (np. „Save 30% on email")
- Feature (np. „Free 14-day trial")
- Audience match (np. „For small businesses")
- CTA-style (np. „Get demo today")
5+ Long headlines (90 chars):
- Combine value props
- Specific use cases
- Differentiators
5+ Descriptions (90 chars):
- Benefit-focused
- Social proof („Trusted by 5000+ companies")
- CTA included
Image assets
- Landscape (1.91:1) — 1200x628px minimum
- Portrait (4:5) — 960x1200px minimum
- Square (1:1) — 1200x1200px minimum
- Logo — square + landscape versions
Best practices:
- High quality (no pixelation)
- Real product / people (not stock photos when możliwe)
- Brand consistent
- Text overlay <20% obraz area
- Multiple variations dla A/B AI testing
Video assets — kluczowe 2026
Google preferuje video assets w 2026 — campaigns z video performują 25-40% lepiej.
Requirements:
- Min 10 seconds
- Recommended 60+ seconds for performance
- Multiple formats: vertical (9:16), horizontal (16:9), square (1:1)
- Captions (sound-off viewing)
Video types:
- Product demo
- Customer testimonial
- Brand story
- How-it-works explainer
Bez video assets, Google auto-generates z text + images — quality lower than custom videos.
„Ads Using Video" segment (new 2026)
Google rolling out new reporting segment that breaks PMax performance by ads with video vs without. Insight: which formats perform best.
Use to validate video investment. Typical: video drives 30-50% more ROAS.
Demand Gen integration
Demand Gen campaigns (formerly Discovery) overlap z PMax. 2026 strategy:
- PMax — performance + reach
- Demand Gen — top-of-funnel demand creation
- Search — bottom-of-funnel intent capture
Hybrid: Demand Gen build awareness → PMax + Search convert.
Common mistakes
- Tight CPA targets too early — AI not enough data, limits delivery
- Too few assets — AI nie ma materiału do testowania
- Mediocre assets — quality matters dla AI optimization
- No video — performance left on table
- Single asset group — wszystko w jednym, no theming
- Ignoring search themes — tylko default
- Over-narrow audience signals — limits AI exploration
- Stopping campaign too early — learning phase 2-3 tyg.
- Brak conversion tracking — AI optymalizuje na nothing
- PMax replacing Search entirely — losing branded query control
Setup workflow
Day 1: Conversion tracking
- GA4 conversions → Google Ads (or direct Google Ads tag)
- Enhanced Conversions enabled (server-side tracking helps)
- Test conversions firing
Day 2-3: Campaign structure
- Create PMax campaign
- 4-6 asset groups (themed)
- Search themes (10-30 per group)
- Audience signals (customer match + custom segments)
Day 4-7: Asset creation
- 5-10 headlines per group
- 5-10 long headlines per group
- 5-10 descriptions per group
- 5-10 images (mix landscape, portrait, square)
- 1-3 videos per group
Day 8-14: Launch + monitor
- Maximize Conversions bidding
- Daily monitoring (delivery, errors)
- Add more assets if AI requests
Week 2-3: Learning phase
- Don't optimize yet
- Let AI gather data
- 30-50 conversions threshold
Week 4+: Optimization
- Switch to Target CPA / ROAS
- Pause underperforming asset groups
- Add new asset groups based on insights
- Test new asset variations
Mierzenie sukcesu
KPIs
- CPA / ROAS primary metrics
- Conversion rate by asset group
- Asset performance ratings (PMax shows „Low / Good / Best")
- Search themes performance (which themes drive conversions)
- Cross-channel insights (which surfaces convert best)
Reports
- Insights tab — AI recommendations
- Asset performance — which creatives win
- Search categories — actual queries triggered (limited visibility)
- Audience insights — which segments convert
Budget recommendations
Small business ($1500-3000/mies.)
- 1 PMax campaign
- 4 asset groups
- Test for 60 days minimum
Mid-market ($5000-15000/mies.)
- 2-3 PMax campaigns (e.g. acquisition vs retargeting)
- 4-6 asset groups each
- Multiple bidding strategies tested
Enterprise ($20000+/mies.)
- 5+ PMax campaigns
- Segmented by product line / audience / geo
- Dedicated PMax specialist
Podsumowanie
Performance Max 2026:
- Min $1500/mies. budget for AI learning
- 4-6 asset groups per campaign (themed)
- 50 search themes per group (use them all)
- Video assets — kluczowe dla performance
- Audience signals — customer match + custom segments
- Maximize Conversions → Target CPA after 30-50 conversions
- 2-3 tygodnie learning phase before optimizing
PMax wymaga AI thinking — nie precise control. Founders chcący micromanage często frustrate. Zaufaj AI z dobrą foundation (assets + signals) — w długim okresie wins więcej niż manual.
Konsultacja Google Ads + PMax setup — pomożemy zaplanować campaign structure + assets dla Twojej firmy.