Personalized product recommendations are a powerful tool for driving sales, improving conversion rates, and increasing order value. In Zaius, entire sections of emails can be dynamically generated to make campaigns more relevant and appealing to customers.
There are four default options available at this time:
Automated recommendations based on each user's behavior compared to other similar users. Zaius Recommendations uses a recommendation engine powered by user-user collaborative filtering with matrix factorization. This algorithm seeks out commonalities among customers based on their patterns of product interactions to predict a customer's response to another product. The machine learning algorithm runs nightly using event data from the previous 180 days. Products with the highest predicted response are recommended.
If a particular customer does not have enough behavioral data to make a prediction, Zaius will use default recommendations instead. Those default recommendations are defined by the best selling products from the last week. As Zaius collects additional behavioral data, the default recommendations will be replaced with more specific ones.
When determining product recommendations, Zaius utilizes the behavior of each customer individually to determine dynamic, personalized recommendations. Given the importance of behavioral data for Zaius Recommendations, this feature is ideal for:
- Campaigns for long-time customers with a history of interaction.
- Relatively stable product catalogs and consistently visited or purchased.
Zaius Recommendations are not ideal for:
- Campaigns for customers without much activity (e.g., New users).
- Campaigns for customers with outdated activity (e.g., Winbacks > 90 days).
- Highly-diverse product catalogs where many products receive little interaction or sales and a comparatively few products are blockbusters attracting much of the activity.
- Seasonal product catalogs where recommendations based on past customer interactions might not be applicable.
Zaius Recommendations will be useful for some customers for some campaigns. However, if there is a solid understanding of your customers, or you want to control your customer's product journey more tightly, Zaius Recommendations may not be a good fit. Instead, consider using Best Sellers, Most Viewed, Revenue Generators, or creating custom recommendations by working with your Customer Success Manager.
Best Selling Products
The products that customers purchase the most. These are "recommendations" that are based on the behavior of all customers and are given as a summary. These recommendations are not personalized on the customer level but are useful in specific nudges/winbacks campaigns.
The products that customers view the most. These are "recommendations" that are based on the behavior of all customers and are given as a summary. These recommendations are not personalized on the customer level but are useful in specific nudges/winbacks campaigns.
The products that currently generate the most revenue for the business. These are "recommendations" that are based on the behavior of all customers and are given as a summary. These recommendations are not personalized on the customer level but are useful in specific nudges/winbacks campaigns.
Alternative to recommendations
If the default recommendations don't fit your situation, it's also possible to refine and define the results of a dynamic grid through product behaviors.
Add recommendations to an email
To add a content section based on recommendations to an email:
- In your Zaius account, navigate to Campaigns.
- Click the name of the campaign you'd like to edit or create a new campaign.
- Locate the email touchpoint that you'd like to personalize and select the Edit icon.
- Drag a Dynamic Grid into your email from the Elements > Layouts section of the editor's sidebar.
- Select the desired default recommendation or custom behavior from the Source dropdown menu.
- Select the Save button to confirm the placement of the content.