You are here:
Set Up Potential to Become a Major Donor Predictions in Nonprofit
Set up the Potential to Become a Major Donor prediction to discover the donors who are most likely to give a major gift. Identifying potential major gift donors helps gifts officers focus their limited time on the right prospects and grow your major gifts pipeline.
| REQUIRED EDITIONS |
|---|
| Available in: Lightning Experience |
| Available in Enterprise, Unlimited, Developer and Agentforce 1 editions with the Agentforce for Nonprofit add-on license where Fundraising and Data 360 are enabled. |
| User Permissions Needed | |
|---|---|
| To set up, train, and manage prediction models: |
|
| To get predictions from a trained model: |
|
Before you set up the prediction, complete this preparation.
- Set up Data 360 and install the Fundraising data kit, so that the required data model objects (DMOs) are available. If the Fundraising data kit is installed, redeploy it so you have the latest updates.
- In Setup, turn on Scoring Framework and AI Accelerator.
- Assign the permission sets listed previously to yourself and to the fundraisers who view scores.
- Decide your major gift threshold and the time frame over which gifts count toward it.
- Confirm that your fundraising data is reasonably complete for the donors you want to score.
Set Up and Train the Model
The setup wizard runs on the AI Models. Each step renders from the template's configuration, so follow the guided pages in order.
- From the App Selector Grid, search for and select AI Models, or go to the AI Models tab in Data 360.
- In AI Models, start a new prediction.
- On the Select Type page, go to Create from Template and select Potential to become a major donor.
- On the Define Configuration page, review the template and continue.
- On the Select Data Space page, select the Data Space where you installed the Fundraising data kit.
- On the Select Subject (Grain) page, for Prediction Granularity Level (Grain), select Account, so that the model predicts at the donor level.
-
On the Select Model Training Duration page, select a Training
Time Period. The model uses data from this time period to
find patterns.
Note The most recent 12 months of data is reserved to validate the model's accuracy. Training uses historical data from before this time period.
- To limit which donors the model includes, on the Configure Filters for Account page, add filters.
-
On the Select and Configure Prediction Criteria page, define what you're
predicting.
- Confirm that the prediction is based on paid gift transactions’ current amount.
- For Define the Major Donor threshold, enter the minimum gift amount that qualifies a donor as a major donor, in your default currency.
- For Gift Transaction Roll-up Period, select to use either a 12-month roll-up period or a single gift transaction.
-
On the Select Prediction Input Variables page, select the input variables
(predictive features) to include.
Select a minimum of three up to a maximum of 49 variables.
-
On the Select an Algorithm for Your Model page, select an algorithm.
Initially, try XGBoost or CatBoost.
- On the Enable Asset Saving page, turn on asset saving to retain your configuration.
- On the Review and Train page, review your selections and start training.
Activate and Use the Scores
After training completes, activate the model in AI Models, formerly Einstein Studio Model Builder. Then make the scores available to fundraisers.
- Add the AI Accelerator card web component to the Person Account record page in the Lightning App Builder.
- Use the prediction DMO to build reports and dashboards in Data 360.
