How to Optimize Omniscore for Subscription Transactions

Kount Omniscore Safety Ratings improve when evaluated transactions contain the same data elements, as with most Machine Learning models. Website IDs can be used to segment different types of transactions with varying amounts of data because they each get their own contained model. For example, a subscription transaction does not typically include Device Data Collector information because the transaction is not created by the customer’s computer or mobile device, it is instead created at regular intervals by the merchant. For optimal predictive safety ratings from the models, subscription transactions should be sent to Kount with a different Website ID than what is used for the initial transactions. Guidance for creating a Website ID in the Agent Web Console is available on our support site. For merchants with direct integrations, the corresponding Risk Inquiry Service (RIS) field for the Website ID is SITE. Additional information on RIS Submission Data can be found in our technical documentation.

Kount’s Omniscore Safety Ratings are mapped to transactions with a known distribution using supervised and unsupervised machine learning. A breakdown of the Safety Rating distribution is shown in the following table:

Mapping of similar transactions to the same Website ID doesn’t have to be complex. Similar transactions from different business lines - or websites - can be sent to the same Kount Website ID. It can then be differentiated for rule creation or reporting within the Website ID using User Defined Fields (UDFs). Alternatively, you can send all recurring monthly transactions from different business lines to the same Website ID.