Do Predictive Scores work by default?
Predictive Scores are calculated within 24 hours after you enable synced accounts, create at least one Keyword Set, and click Save and Train Model on the Predictive Score Setup page.
To access the page, from the left navigation bar, go to Settings > Demandbase Wide-Settings > Predictive Score Setup.
How long does it take to train a Predictive Score model?
Training a Predictive Score model takes up to 24 hours.
What is considered a good Predictive Score?
Use the following scores to prioritize accounts:
- 50 or higher: Accounts worth prioritizing for additional attention.
- 95 or higher: Accounts that require immediate attention.
Can I revert to an earlier version of a Predictive Score model?
No. You cannot revert to an earlier version of a Pipeline Predict or Qualification Score model.
How often should I retrain Predictive Score models?
Retrain Predictive Score models at least once a year to incorporate the latest data and maintain accuracy.
Demandbase supports two retraining options:
- Auto Retrain (recommended): Automatically retrains models to maintain score accuracy.
- Manual Retrain: Allows you to retrain models manually. We recommend retraining Pipeline Predict Scores every three months and Qualification Scores every six months.
Important: Auto Retrain is enabled by default.
How do Predictive Scores improve over time?
Predictive Score models incorporate new signals from available data sources each night to keep account scores current.
How often are Pipeline Predict Scores updated?
Pipeline Predict Scores are updated nightly as new account activities become available. Scores can also decrease as existing activities become less recent.
Do Pipeline Predict Scores only use engagement data from Marketing Automation Systems (MAS)?
No. Pipeline Predict Scores use activities from supported data sources that can be assigned Engagement Points in Demandbase.
How does Pipeline Predict determine which personas and job titles are most relevant?
The Pipeline Predict model automatically identifies relevant personas and job titles based on available account data, engagement activity, and job levels, such as CMO, VP, Director, and Manager.
Are accounts imported through CSV files scored differently from CRM accounts?
No. Demandbase uses the same scoring methodology for all accounts, regardless of whether they originate from a CRM integration or CSV import.
What opportunity data is required to train a Predictive Score model using a CSV file?
The model requires the following information:
- Account Website/Domain: Identifies the account associated with the opportunity.
- Created Date: Identifies when the opportunity was created.
Important: CSV data imports may require additional fields to support other Demandbase features.