Introduction to Einstein Content Selection

The Einstein Content Selection tool selects engaging content (images) based on past popular content and the rules you define.



As the subscriber views the email (not when the email was sent), the tool takes into account changes to attributes and actions taken by others in the pool. It works in real-time, scanning content in your content “pool” available at the exact moment the subscriber views the email.


How Einstein Content Selection Works?

  1. Create a content pool, including asset classes, attributes, profile attributes, and fallback assets (we’ll cover an overview in the next section).
    If you already have assets in Content Builder, you will need to read them to Einstein Content Selection manually.


  1. Profile attributes will be provided by a data extension.
  2. Configure Attribute Rules: This connects asset attributes with profile attributes (i.e. content and subscribers, which might look different but mean the same thing in reality).
  3. Einstein will make its choices for what content to serve based on exclusion, fatigue, and variety rules.
  4. Create an email using Content Builder and add a Einstein Content Selection block. Einstein will know not to show the same image in more than one Content Selection block referencing images from the same asset class.
  5. Using your rules, subscriber attributes, and what’s been popular in the past, Einstein searches your content pool for the content most likely to be clicked on.
  6. As Einstein tracks subscriber interactions with each piece of content, he learns what works well, when, and with which audience.


Einstein Content Selection Setup

Asset Classes: Act like a category for assets, grouping them. This will help with making your analysis, more targeted to a specific category of assets. An asset can only belong to one asset class.

Asset Attributes: Metadata provides information about an asset (e.g. a related product that you sell).

Profile Attributes: Language, location, etc., about your subscribers. In Einstein Content Selection, you select a “consumer profile” data extension.

Fallback Assets: If there are no matching assets, these assets will be used.


Einstein Content Selection Setup

These rules ensure Einstein has business context (i.e. what actually happens) so that it can make the best decisions:

Attribute Rules;
Exclusion Rules;
Fatigue Rules;
Variety Rules;

Attribute Rules: Map asset attributes to subscriber profile attributes (i.e. connect the dots between content and subscribers). The “Must Match” checkbox means that the asset is only for a targeted audience, in other words, only subscribers that have that attribute. This prevents Einstein from straying in its experimentation.

Exclusion Rules: What content should not be displayed to which subscribers? Basically, Einstein must abide by hard rules that can’t be inferred from other data Einstein is using.

Fatigue Rules: Ideally, you don’t want to repeat content that isn’t sticking in emails to a specific subscriber.

  • X Days After Last Selection: Einstein last selected the content for the subscriber).
  • X days After Last Click: When the subscriber last clicked on the content).
  • Selection maximum: The number of times a subscriber can view an asset.


Einstein Content Selection + More Einstein

With Marketing Cloud Einstein, Einstein Content Selection can be incredibly powerful, especially when combined with:

☁️ Einstein Copy Insights: With text analytics and natural language processing, you can analyze email subject lines to uncover insights (such as the impact of phrases, tones, and punctuation) that can be used to improve open rates.

☁️ Einstein Content Tagging: Utilizes Google Vision to analyze and process images in your Marketing Cloud account (allowing you to view asset performance by content tag).




* indicates required

6 Ways to Use Workbench in Salesforce

Salesforce Workbench is a Free, Online device that can be utilized for Data Management activities such as Bulk Record Updates, Mass Creation, and Mass Deletion. If you’ve ever used Data

 3,519 total views

Read More »