Data Requirements for Sales Cloud Einstein

Einstein Automated Contacts

Important Feature is scheduled for retirement on February 15, 2025. See Einstein Automated Contacts Retirement.

• You must have at least 30 business accounts.
• If you use person accounts, at least 50 percent of accounts must be business accounts.

Einstein Activity Capture

• You must have at least 30 accounts, contacts, leads, or opportunities.

Einstein Lead Scoring

These requirements apply to each segment of leads that you create during setup, including the All Leads (Default) segment.

• At least 1,000 leads must be created in the last 200 days.
• Of the leads created in the last 200 days, at least 120 must be converted to an account and contact.
• (Optional) Of the leads created in the last 6 months (180 days), at least 120 must be converted to an account and contact with an opportunity created at the conversion time.

When you score all leads together without creating segments, and you don’t have enough lead conversion data to build your own predictive model, Einstein uses a global model. The global model uses anonymous data from many Salesforce customers. When you accumulate enough lead data, Einstein builds a scoring model with your data and uses the model with the better results.

Einstein Opportunity Scoring

• You must have at least 200 closed won opportunities in the last 24 months, each with a lifespan of at least 2 days.
• You must have at least 200 closed lost opportunities in the last 24 months, each with a lifespan of at least 2 days.
• Opportunity history shows an average of one update to each closed opportunity.
• Use the standard opportunity Stage field because it’s used to calculate win rates, and win rates are used to generate scores. If you change the names of the opportunity stage picklist values, make sure that the values are mapped to the correct stage type: Open, Closed/Won, or Closed/Lost.
• If your win rate is extremely high or low, your scores could be skewed. For example, if your win rate is higher than 90 percent, you could get a large number of opportunities with scores higher than 90. To avoid skewed scores, make sure that opportunities are set to the correct closed stage. (A win rate is calculated by dividing the last two years of closed-won opportunities by all closed opportunities from that same period.)

If you don’t have enough opportunity data to build your own predictive model, Einstein uses a global model. The global model uses anonymous data from many Salesforce customers. When you accumulate enough opportunity data, Einstein builds a scoring model with your data and uses the model with the better results.

Einstein Forecasting

• Salesforce Forecasting must be enabled.
• You must work with opportunities in Salesforce for at least 12 months. Specifically, the opportunity history must show at least one update in each of the past 12 months.
• You must use a standard fiscal year. Standard fiscal years follow the Gregorian calendar, but can start on the first day of any month of the year.
• Forecasts must be measured by opportunity revenue. Predictions are generated for only the oldest activated opportunity revenue forecast type.
• Your forecast hierarchy must include at least one forecasting-enabled user who reports to a forecast manager.
• The Amount field should be populated in at least 80 percent of open opportunities.