Einstein Engagement Scoring scores contacts in Marketing Cloud by likelihood to engage — open emails, click links, make purchases. Lets marketers target high-engagement segments and avoid sending to low-engagement contacts who are only adding to deliverability cost. Marketing Cloud Engagement feature; configured in MC, not standard Salesforce CRM.
- Open Marketing Cloud → Einstein → Einstein Engagement Scoring
Different from CRM Setup — Marketing Cloud has its own admin UI.
- Tick Enable Engagement Scoring
Activates analysis on contacts in Marketing Cloud Engagement (formerly ExactTarget).
- Configure scoring dimensions
Email engagement (opens, clicks) / Purchase likelihood / Web browsing patterns. Multi-pick.
- Wait for training
Marketing Cloud builds the model. Typical training time depends on Sender Authentication Package (SAP) data volume.
- Use scores in Journey Builder / Email Studio segmentation
Filter sends by engagement score — "send only to contacts with score > 70."
- Monitor send-quality metrics
Better targeting via scoring should reduce bounce rates and improve sender reputation.
Open / Click / Purchase / Browse.
Per-segment cutoffs.
MC retrains on schedule.
- Engagement Scoring requires Marketing Cloud licensing (separate from Sales / Service Cloud). It's not in standard CRM — admins look for it in MC, not Setup.
- Scores degrade for inactive contacts. Reactivating a long-dormant contact list won't have great Einstein scores until they re-engage and the model picks up new signals.
- Scoring trained on email engagement may not predict purchase well. Each scoring dimension is a separate model — cross-using scores between use cases produces unreliable results.