Best Practices for Setting Up and Using Predictive Scores

Demandbase predictive scores help you prioritize accounts for marketing campaigns and sales interactions. To calculate these scores, Demandbase applies machine learning with data from all available sources, including your CRM, Marketing Automation System (MAS), and CSV imports.

See the following best practices to get the most out of your predictive scores.

Pipeline Predict Scores

1. Retrain scores every time you update your engagement minutes or intent keyword configurations. We recommend enabling the Auto Retrain option. See Understanding Pipeline Predict Scores.
2. Recently opened opportunities (past 12 months) are recommended but the date range can be expanded if you have longer sales cycles.
3. Use a minimum of at least 50 distinct accounts with qualified opportunities.
   Important:
   • The model only looks at accounts that have preceding engagement prior to opportunity creation. 
   • For growth (renewal, upsell, or cross-sell) opportunity scores, make sure to include existing Customers when training the model. For example, use the selector Journey Stage = Customer.
4. Train the model with as much engagement data as possible. Instead of using a specific subset of data to train the model, exclude the data you don’t want, allowing the model to be trained on a larger dataset.

Qualification Scores

1. Retrain scores every time you update your technographics or intent keyword configurations. We recommend enabling the Auto Retrain option. See Retrain Qualification Scores.
2. Accounts that recently became a customer (past 12 months) are recommended but the date range can be expanded if you have longer sales cycles.
3. Train the model with as much engagement data as possible. Instead of using a specific subset of data to train the model, exclude the data you don’t want, allowing the model to be trained on a larger dataset.
4. Use a minimum of 50 distinct accounts. We recommend 100 or more accounts.