Setting up Days to Pay Models
This topic details how to create a model to minimize the Days to Pay field from the Financial Matching dataset. Days to Pay is the number of days the customer takes to pay on average, based on their previous payments. This model analyzes all the different data related to this metric in your Financial Matching dataset and highlights the factors that influence it.
Requirements
Before you create or reconfigure a Business Analytics app to use Einstein Discovery models, ensure that you:
- Have the necessary licenses and permissions to generate an Analytics app, these requirements are listed in Configuring Financial Management Analytics.
- Have the license for Einstein Discovery enabled in your org. For more information, see the Salesforce Help.
- Have assigned the license for CRM Analytics Plus. For more information, see the Salesforce Help.
- Have assigned the CRM Analytics Plus Admin permission set.
For users who only need to view the Einstein Discovery models, ensure they have the following requirements:
- Have assigned the CRM Analytics Plus User permission set.
- Have assigned the licenses and permissions detailed in Configuring Financial Management Analytics.
Limitations
There might be limitations on the minimum amount of data required to generate meaningful predictions. For more information, see the Salesforce Help.
Creating a Model
This section details the specific options required to create a model to minimize the Days to Pay field from the Financial Matching dataset. For general information about creating a model, see the Salesforce Help.
Model Configuration
Select the following values for each field.
Field | Value |
---|---|
I Want to Predict | Days to Pay |
So I Can | Minimize |
App | Select the Financial Analytics app where you want to store your model. |
Model Column Configuration
Select the following option.
Option | Description |
---|---|
Manual | It allows you to manually select and configure the columns that you want to include in your model. |
Variables
The fields listed below are automatically included when you create a model, but you can select any fields that you consider relevant to generate accurate predictions.
Field |
---|
AccountBaseDate1 |
AccountBaseDate2 |
AccountBaseDate3 |
AccountBaseDate4 |
AccountBillingCity |
AccountBillingCountry |
AccountBillingMethod |
AccountBillingState |
AccountCollectionsOnHold |
AccountCollectionsOnHoldReason |
AccountCreditAgency |
AccountCreditRating |
AccountCreditStatus |
AccountDimension1Name |
AccountDimension2Name |
AccountDimension3Name |
AccountDimension4Name |
AccountIndustry |
AccountName |
AccountOwnerName |
AccountPaymentMethod |
AccountingBookName |
BankAccountBankName |
Dimension1Name |
Dimension2Name |
Dimension3Name |
Dimension4Name |
FfaCompanyName |
ProductName |
ValueAccountCreditLimit |
ValueAccountDaysOffset1 |
ValueAccountDaysOffset2 |
ValueAccountDaysOffset3 |
ValueAccountDaysOffset4 |
ValueAccountDiscount1 |
ValueAccountDiscount2 |
ValueAccountDiscount3 |
ValueAccountDiscount4 |
General Settings
We recommend that you select the following values for each field.
Field | Value |
---|---|
Algorithm | GLM |
Validation | Automatic Cross Validation |
Deploying a Model
Select one of the following deployment options.
Option | Description |
---|---|
Deploy as a new model within a new prediction definition | Select this option if you create a prediction definition for the first time. |
Update an existing model | Select this option if you have modified your model and you want to obtain a new deployed model. |
For each step, select the following options.
Step | Option |
---|---|
Connect to an Object | Deploy without connecting to a Salesforce Object |
Segment Data | Don't Segment |
Select Actionable Variables |
The following fields are automatically included when you deploy a model, but you can select any fields that you consider relevant to obtain accurate improvements:
|
Customize Predictions |
Don't Customize |
Days To Pay Models for Intelligence Analytics
If you want to use your model for your Intelligence Analytics app, you can select the prediction definition created for your model in the configuration wizard for Intelligence Analytics. For a list of questions included in the configuration wizard, see Configuration Wizard Questions.