Next best recommendations
Next best recommendations allow you to automatically send customers optimal actions and offers they are most likely to accept.
Next best recommendations explained
Next best recommendations leverage Oracle Unity data and machine learning to determine the most relevant recommendations for customers. The Oracle Unity machine learning model calculates scores for actions and offers based on customers' interests.
There are two types of next best recommendations: Next best actions, which score actions, and Next best offers, which score offers.
Next best actions
Next best actions determine actions that should be proposed to each customer at a particular time. Actions aren't necessarily tied to a specific product or service.
Examples of next best actions include the following:
-
Send a reminder to bring their car to a dealer for a service check up.
-
Email a customer and request that they sign up for a newsletter.
-
Send a birthday message.
Next best offers
Next best offers determine the most relevant offers to send to customers. Offers have the specific details on how the product or service will be offered, such as the price, payment terms, and applicable discounts.
Examples of next best offers include the following:
-
Purchase the new model of a smartphone for $1300 and receive a mail-in rebate of $200.
-
Purchase a television for $1200 by making 12 monthly payments of $100.
Setting up the data for Next best recommendations
Before creating Next best recommendations, you will need to follow the steps for Setting up data for Next best recommendations. After setting up the data, you can start creating Next best actions and Next best offers.
Creating Next best recommendations
Creating a Next best action or Next best offer includes doing the following.
-
Select evaluation criteria: You will select the segment that determines the criteria for customers receiving the actions/offers. For example, you select a segment for a Next best offer that will send offers to customers if it is their birthday and they've made more than one purchase in the past year.
-
Configure actions/offers: You have two options: leverage the machine learning model to automatically rank and select actions/offers that are sent to customers, or manually select the actions/offers you want to send.
-
If you leverage the machine learning model, you will need to select from a list of catalogs, which contain the scope of actions/offers that can be selected from and sent to the customers identified in the evaluation criteria. The machine learning model will calculate a score for each action/offer to every customer in the segment. The score will be calculated from 0 to 1, with 1 being 100% probability that the customer will be interested in the action/offer. The scores will determine the most relevant actions/offers to send to each customer.
-
If you do a manual selection, you will need to select the individual actions/offers you want to include.
-
-
Select destination: You will need to select the where the data will be exported to. This will determine the marketing orchestration platform that will activate the actions/offers. You have the option of creating the following destinations.
-
Secure FTP: Create a Secure FTP destination.
-
Responsys: Create a Responsys destination.
-
CX Sales: Create a CX Sales destination for Next best recommendations.
-
-
Configure schedule and notification settings: You will need to configure the schedule and notification settings for the job associated with the Next best action or Next best offer.
Next steps
Setting up data for Next best recommendations
Managing next best recommendations