We recently pushed out two new features enabling teams to generate more value from real-time next-best-action and next-best-offer experiences. The first new feature, multiple decision ranges, empowers a broad set of new use cases to move users down marketing funnels towards conversions (we provide examples below of a next-best-offer E-commerce use case and subscription next-best-action experience). The second new feature makes it easier for teams to quickly experiment with and deploy real-time decisioning using Cortex. 

New Feature: Multiple Decisioning Ranges within a Single Machine Learning Pipeline

One of the unique abilities of Cortex is to convert machine learning outputs (which are typically “propensities” that have no real-world meaning) into real-world probabilities, which can have a real-world meaning (you can learn more about the difference between propensities and probabilities here). Decisioning enables a business to take actions on users based on their probability scores. Our customers asked for the ability to make unique decisions across multiple probability ranges. With this latest release, Cortex enables teams to easily add new decision ranges for any project.

Here are a couple use cases this new functionality enables –

Increasing E-Commerce Conversions with Coupons

E-commerce companies often increase user conversions through targeted offers and discounts. Let’s take an example of an E-commerce company incentivizing the purchase of shoes. In this case, we might start by building a machine learning pipeline to predict the likelihood of a user to purchase shoes. We can leverage this pipeline within a decisioning framework to promote the right discount or coupon to the right user. For instance, a user who is very likely to purchase might not need any discount, while a user who is currently not likely to purchase might require a larger discount.

Example in Cortex of a Next-Best-Offer Decision Project that optimizes the user experience for encouraging the purchase of shoes. Note: multiple decision ranges are now available within the UI. A user can specify as many decision ranges as they like. In addition, those ranges can be easily modified to enable fast and easy experimentation.

Subscription Service – Moving Users Down the Funnel

Any subscription business has a primary objective to increase the total subscribers. Typically, a marketing team for a subscription business spends a good deal of time considering how to move users down a funnel from an anonymous first-time user into a loyal subscriber. Machine learning decisioning can help. Here’s how. First, a subscription company might build a likelihood to subscribe pipeline. By adding this pipeline to a project, the business can specify decisions based on the likelihood of a user to convert. For instance, users who are very likely to convert to a subscriber (far down funnel) might be directly encouraged to convert. Users with a medium likelihood to convert, might first be encouraged to register. Users with a low likelihood to convert might have the option of trying to service out longer without any commitment. A Cortex user can specify as many funnel stages as they like within a project.

New Feature: Streamlined Deploy Pipeline Tab

We know that pushing machine learning experiences quickly into production is a huge win for our customers. We also have seen that the ability to quickly experiment with and iterate on machine learning experiences provides a lot of value. With this latest release, we’ve made it even easier to deploy pipelines into production.  

In Cortex, the ‘Export Predictions’ button has been updated to ‘Deploy Pipeline’

This includes the ability to both Export Predictions to a custom integration on a recurring basis, as well as directly within the UI by creating a Decision Project, as shown below.

Over the past few weeks we’ve pushed out a variety of features in Cortex, many focused on real-time decisioning. We have a lot more planned over the coming weeks and months. That being said, we always welcome feedback from our partners and customers. Please don’t hesitate to reach out – especially if you have a great next-best-action or next-best-offer experience you are working on!

Want to Learn More?

Schedule a demo and talk to a product specialist about how Vidora’s machine learning pipelines can speed up your ML deployment and ultimately save you money.