Risk Score

The D•One Risk Score aims to predict an applicant’s credit risk, using only Open Banking transaction data. The score is an integer between 0–1000, with 0 being the highest risk and 1000 being the lowest risk.

How the D•One Risk Score Predicts Credit Risk

Leveraging transaction data sourced from Open Banking connections, the D•One Risk Score predicts the likelihood of a user missing two or more payments at month 9. The D•One Risk Score can be used to predict credit risk across both credit card and personal loan products.

Integration and Testing

The D•One Risk Score can be tested in the D•One Sandbox using test user data from the D•One Test Bank.

For detailed API specifications, see the Risk Score API endpoint.

An example successful response can be viewed below:

{
  "userId": "6e5c4a92-c3e9-42dd-827e-9de12975ccba",
  "reportDateTime": "2025-05-02T09:01:34Z",
  "categorisationVersion": "43.0.0",
  "modelVersion": "1.0.0",
  "riskScore": 813
}

Integration Guidelines for Marketplace Journey

When integrating with a marketplace who has partnered with D•One such as ClearScore the D•One Risk Score will be available when calling the Risk Score API endpoint. For further details on how to integrate with the marketplace please view the guide here.

Integration Guidelines for Direct Connection Journey

When integrating the D•One direct connection journey, it’s best to wait until transaction data is available before calling the D•One endpoints, including the D•One Risk Score endpoint. You can check if a user has data for a specific account by calling the accounts endpoint and reviewing the lastTransactionRefreshDate field. If this field is present, transactions are available and a risk score generation will have been triggered.

For more information, see Data Availability.

Error Responses

The D•One Risk Score requires a minimum amount of data in the user’s account for a score to be generated. Below is an example error response when minimum data requirements are not met:

{
  "outcome": "CRITERIA_NOT_MET_MONTHLY_AVG_DEBIT_12_MONTHS"
}

A complete list of possible criteria failures is provided below:

Criteria	Description	Failure Outcome
Transaction volume	Average of 15+ debit transactions/month over 12 full months	`CRITERIA_NOT_MET_MONTHLY_AVG_DEBIT_12_MONTHS`
Recent activity	At least 1 transaction in the 12 days before scoring	`CRITERIA_NOT_MET_TRANSACTIONS_12_DAYS_BEFORE`
Historical activity	At least 1 transaction in the same month one year prior	`CRITERIA_NOT_MET_TRANSACTIONS_1_YEAR_BEFORE`
Income visibility	Average of 0.5+ income transactions/month over 12 full months	`CRITERIA_NOT_MET_MONTHLY_AVG_INCOME_12_MONTHS`

Criteria checks are performed in a fail-fast approach in the order shown above.

Risk Score updates

As the D•One Score requires categorised transaction data to predict credit risk, the model underpinning the D•One Risk Score is linked to categorisation versions. The Risk Score launches with model version 1.0.0, which utilises categorisation version 43.0.0.

D•One aims to release major updates to the transaction categorisation service at least annually. A corresponding risk model update will also be issued.

We encourage partners to update to the latest version of the D•One Risk score as soon as possible. However, we appreciate that updates to credit risk decisioning requires consideration. Therefore, D•One will not force partners to change score versions once live.