Auto-Target uses advanced machine learning (Random Forest models) to evaluate each visitor's profile attributes and predict which of several experiences will maximize the chosen success metric for that individual. Unlike A/B tests that find one winner for everyone, Auto-Target finds the best experience per visitor.
When to Use
You have multiple distinct experiences
Your audience is diverse with different preferences
You want 1:1 personalization without manual rules
You have enough traffic for ML to learn (1,000+ visitors/variant)
Key Characteristics
Random Forest ML model (Adobe Sensei)
Considers 40+ visitor attributes
Continuously learns and adapts
Exploration vs. exploitation balance
How It Works
1
Marketer creates multiple experiences
2
ML observes visitor attributes
3
Random Forest predicts best match
4
Visitor sees optimal experience
5
Model refines with each interaction
📐 Experience Composers
Adobe Target offers two ways to create Auto-Target experiences. Use the VEC for visual modifications, or the Form-Based composer for named-location content delivery. The ML model works with both.
🎨 Visual Experience Composer (VEC)
WYSIWYG editor — visually create experiences for ML matching
Target identifies elements via CSS selectors
Uses renderDecisions: true
SDK auto-applies DOM modifications
ML selects best visual variant per visitor
Best for: visual layout/copy changes with ML optimization
The hero section below contains elements targetable via the Visual Experience Composer. In an Auto-Target activity, the ML model selects the best visual variation of these elements for each individual visitor.
Waiting for VEC propositions…
Personalized by ML
Your Perfect Experience Awaits
Machine learning analyzes your profile to deliver the optimal experience from our curated collection.
💡 Create an Auto-Target activity in Adobe Target using the VEC. Create multiple visual experiences (e.g. Premium, Value, Trending layouts) and let ML automatically assign the best one per visitor via renderDecisions: true.
📝 Form-Based Simulation — ML Visitor-to-Experience Matching
Generate visitors with different profiles and watch the ML model score each experience. The model uses traits like device, age, spend level, and interests to predict the best match.