Personalization Types
Salesforce Personalization supports two types of personalization that you can deliver on a website: Manual Content and Recommendations. Understanding the difference between these two types is the first step in designing your personalization strategy.
Manual Content
Manual Content is the simplest type of personalization you can execute with Salesforce Personalization. It's ideal for getting started quickly and delivering targeted, rules-based experiences.
How It Works
With Manual Content personalization, a business user defines string values directly on a personalization decision — things like CTA text, header text, image URLs, and subheadings. Whatever values they type in are stored on the decision definition and returned in the decision response at run-time for qualified individuals.
The fields available for text entry are defined by the response template associated with the personalization point (see Response Templates).
Common Use Cases
- Banners with CTA — Targeted hero banners with custom headlines and call-to-action buttons
- Infobars — Promotional bars at the top or bottom of a page (shipping offers, sale announcements)
- Pop-ups — Overlay modals triggered by scroll, exit intent, or time on page
- Text replacement — Swap headlines, descriptions, or button text based on audience segments
Key Characteristics
| Characteristic | Detail |
|---|---|
| Data source | Content is stored on the decision definition itself — no Data Cloud DMO needed |
| Complexity | Low — no recommender, no item data graph required |
| Targeting | Rules-based, using profile data graph attributes, segments, CIs, or contextual rules |
| Engagement tracking | Tracked at the personalization point / decision level (not at the item level) |
| Best for | Quick wins early in your implementation lifecycle |
💡 Tip: Manual Content use cases are excellent candidates for your first personalization experiences. Since they don't require a recommender or item DMOs configured in Data Cloud, you can deliver value quickly while the more complex data infrastructure for recommendations is being built out.
Recommendations
Recommendations personalization is dynamic and ML-driven, capable of delivering 1:1 personalized item recommendations to each individual visitor.
How It Works
With Recommendations, a business user selects a recommender on a personalization decision. The recommender is built against a profile data graph and an item data graph. At run-time, the recommender evaluates the individual's profile data, engagement history, and the item catalog to determine which items to return.
Common Use Cases
- Product recommendations — "Recommended for you" carousels on homepage or product pages
- Article recommendations — "You might also like" sections on knowledge base or blog pages
- Cross-sell / upsell — Related product suggestions on cart or checkout pages
- Similar items — Products or articles similar to the one currently being viewed
Key Characteristics
| Characteristic | Detail |
|---|---|
| Data source | Item data must exist in a Data Cloud DMO (product name, price, image URL, etc.) |
| Complexity | Higher — requires a recommender, item data graph, and profile data graph |
| Targeting | Can be personalized down to the 1:1 level via ML (objective-based recommenders) |
| Engagement tracking | Tracked at the item level — each recommended item gets a personalizationContentId |
| Best for | Dynamic, data-driven experiences after your data infrastructure is established |
⚠️ Important: For recommendations to return content in a decision response, the item metadata must be available in a Data Cloud DMO. For example, to render product recommendations, you need product attributes like name, price, image URL, and purchase URL mapped to a DMO and included in an item data graph.
Choosing the Right Type
Use the decision tree below to determine which personalization type fits your use case:
Is the content you want to show stored in a Data Cloud DMO?
│
├── NO → Use Manual Content
│ - Define the content directly on the decision
│ - Great for banners, infobars, pop-ups, text changes
│
└── YES → Do you want ML-driven 1:1 personalization?
│
├── NO → Use Manual Content (with static rules)
│ - You can still target specific audiences
│ - Content is manually defined per decision
│
└── YES → Use Recommendations
- Configure a recommender (rules-based or objective-based)
- Build an item data graph for the item type
- Dynamic, personalized results per individual
Summary
| Manual Content | Recommendations | |
|---|---|---|
| Content defined by | Business user typing values on the decision | Recommender selecting items from an item data graph |
| Item data graph required? | No | Yes |
| Recommender required? | No | Yes |
| Personalization level | Segment-level (rules-based targeting) | 1:1 individual-level (with objective-based recommenders) |
| Implementation complexity | Low | Medium to High |
| Speed to value | Fast — can be set up in minutes | Slower — requires data infrastructure |
| Example | "Free shipping on orders over $50" infobar | "Recommended for You" product carousel |
Next: Recommenders — Learn how to configure the engines that power personalized recommendations.