OVERVIEW
Want to tune customer risk to your own business needs? You can do that in Decision Studio by adjusting the identity risk categories. When activating Identity Risk Intelligence, your account will receive Riskified’s pre-set identity risk categories (i.e., high risk identity). Merchants who are already using one of Riskified’s other products (in addition to Identity Risk Intelligence) have the ability to review and adjust identity risk categories within the Decision Studio of the Control Center.
TABLE OF CONTENTS
What are Identity Risk Categories and the policies that set them?
Policy Examples
View Policies
Policy Priority (Analysis Order)
Create, Edit, and Deactivate Policies
Policy Components
PREFER VIDEO?
Watch this quick walkthrough!
What are Identity Risk Categories and the policies that set them?
When using the Identity Risk Intelligence, Riskified provides an analysis of the identity. By default, every identity is assigned an Identity Risk Category. The options are:
- New Identity
- Low Risk Identity
- Medium Risk Identity
- High Risk Identity
For merchants already using other Riskified products
If your store is already using another Riskified product, every identity category has a corresponding policy rule in Decision Studio that you can view and adjust for your exact business needs.
For merchants only using Riskified's Identity Risk Intelligence
At this time, adjustment of Identity Risk Categories is only available to merchants who are also using other Riskified products and sending us their order data. If you want to share your order data with Riskified in order to customize your risk category policies, reach out to our Customer Support team for more information. Decision Studio is part of Riskified's full product suite where you can adjust the thresholds and attributes behind each identity category. You can also see how your identities are distributed across categories.
Policy Examples
Below are some example use cases of when you may want to change categories:
- Adjust Existing Policy - If you want a stricter policy, you can lower a threshold. For example, your current policy flags a customer as high risk when their claim or chargeback rate is above 25%. Lower it to 15% to flag more customers as high risk.
- Create a New Policy - If your business has its own rules about how many accounts one identity can have, you can create a new policy based on the number of accounts. Customers who match the policy are assigned to the identity risk category you select. For example, if an identity has more than 4 accounts, you can assign it to the medium risk category.
- Deactivate a Policy - If a policy you created is no longer relevant or isn’t working as expected, you can deactivate it. Customers are no longer assigned to identity risk categories based on that policy.
View Policies
The policies that set identity risk categories can be viewed within Decision Studio the Control Center.
- Log into the Control Center.
- Go to Decision Studio > Policies > Identity Categories.
- You can use the filters and search bar if needed to narrow down your search for the relevant policy.
- Click on the policy to view it. Read more about the different parts of a policy under Policy Components.
Policy Priority (Analysis Order)
When evaluating an identity, policies are applied in a strict priority order (displayed from top to bottom in Decision Studio). As soon as an identity matches the criteria in a policy, the identity category is assigned and evaluation stops. Below is a flowchart that explains how policies are applied during analysis:
Create, Edit, and Deactivate Policies
Your team can also create new policies, edit existing ones, and deactivate directly in the Control Center’s Decision Studio.
Create a Policy
- Inside Decision Studio > Policies, click on “Create New Policy”.
- Select “Identity Categories”.
- A form will appear to enter all the details you need to create the policy. See Policy Components below for more information about policy criteria. Once you enter the necessary policy details, click “Continue.”
- You will be prompted to run a simulation. Riskified will use historical data to tell you how the new policy will affect your identity population so that you can see the implications before activating.
- If you are happy with the impact of the new policy in simulation, you can click “Publish.” The policy will be used in identity analysis as soon as you publish it.
Edit a Policy
- Go to the relevant policy (see View Policies if you need steps on how to find your policies). In the top right corner, you will see “Edit Policy,” which will allow you to change the policy. Once you edit the necessary policy details, click “Continue.”
- You will be prompted to run a simulation of the policy to see the impact of the change.
- If you are satisfied with the impact of the new policy in simulation, you can click “Publish.” The updated policy will be used in identity analysis moving forward.
Deactivate a Policy
- Go to the relevant policy (see View Policies if you need steps on how to find your policies). Click on “Edit Policy” in the top right corner.
- Change the status from “Active” to “Inactive.” Click “Continue.”
- You will be prompted to run a simulation of the change before publishing so that you can see the implications of deactivating the policy.
- Click “publish” to finalize the deactivation of the policy.
Note: You may reactivate a policy at any time using the same steps and moving the status from “Inactive” to “Active.”
Policy Components
Every policy for an identity category consists of 4 components:
- Policy Details
- Criteria
- Outcome
- Simulation
Below is a breakdown of how each component works and best practices for configuring policies.
1. Policy Details
This section is where you define the administrative information for a new policy. Every policy has a name and status.
- Policy Name: A short, representative name visible in the main policy table. We recommend using a concise, descriptive title such as “Flag high-risk for claim rate > 20%”
- Status: The current state of the policy. You can choose from one of three options:
- Active: The policy is live and actively applied during identity analysis. Policies are created with “Active” status by default. If you don’t want the policy to go live immediately, toggle the status to “Inactive” after creating it.
- Inactive: The policy is not live and is not applied during identity analysis.
2. Criteria
The Criteria section defines the conditions under which a policy will be triggered. All conditions are combined with an “and” clause, meaning a policy is only executed when every condition is met. If you want to create a policy with an “or” clause, then you will need to create two separate policies instead.
3. Outcome
The Outcome section defines the identity category that will be returned when a policy’s criteria are met. (Remember that you can only choose "high risk" or "medium risk" as a value since the "new identity" policy is preset and not changeable and anything that doesn't fall under your policies will be assigned as "low risk". Read more about this in Policy Priority).
4. Simulation
Simulation is a final step before a policy can be published. Riskified uses historical data to show you how the policy would have affected your identity population, so you can assess the impact before the policy is live. The simulation includes the following benefits:
- Select whether you want to run the simulation based on data from the past 7, 30, or 90 days
- View aggregate results of your policy change across all identities or isolate to just identities with claims or chargebacks
- See the before and after population changes in identity category analysis (how many identities will be classified as high risk, medium risk, low risk, and new identity)
- Dig deeper into a sample list of real customer examples where their identity risk will be changed due to the new policy
After simulation, you will be prompted to publish the policy. Keep in mind that as soon as you publish, the policy will be activated and Riskified will start to use this policy. The policy you create will be auto-prioritized (read more about policy priority).