Es tut uns leid, aber diese Seite wurde leider noch nicht übersetzt. Doch keine Sorge, das Emarsys Lokalisierungsteam arbeitet eifrig daran, unsere gesamte Dokumentation in Ihrer Sprache bereitzustellen.
Before you start to design and implement your recommendation widgets, you should read these pages to get a better understanding of what Predict can do for you and your business.
Contents:
Introduction
In this article we describe what we regard as the absolute minimum best practices that you should follow to take proper advantage of the automated revenue generation Predict can provide.
The Email Recommender widgets can be used in virtually any email. However, this product can be highly effective when used in conjunction with the Automation Center, since you can then control when and how to send the message, as well as what content to send.
Where to place the widgets
Our experience has shown us that where and how you place your widgets can have a major impact on their efficacy, so we want to you get it right first time. The use cases are based around a typical buying process where the visitor browses items, decides whether to add them to the shopping cart, and ends the session either by purchasing them or abandoning them in the cart.
The image below illustrates how these examples can be followed in order to optimize and automate revenue generation along the buying process.
1. Using Recommendations In Abandoned Cart Campaigns
When a cart is abandoned, this widget displays the products that were left behind. Depending on the number of products the widget can show, any remaining placeholders show complementary recommendations.
The abandoned cart campaign is an Automation Center campaign that uses the “predict last abandoned date” behavior field to send an email campaign following a cart abandonment. In the email campaign we use Predict’s Abandoned Cart email widget, which is created either directly in the campaign editor (for block-based emails) or using Predict’s Email Widget Designer (for the HTML editor) .
This Abandoned Cart widget will list:
- The products left in the cart.
- Other relevant product recommendations to help the up- and cross-sales.
Typical use cases:
- Revenue recovery by bringing the customer back to complete the purchase.
- Fully automated – no need to think about it once it is active.
- Offer dynamic incentives based on the value of the cart.
Please note that Emarsys offers an improved, faster solution for sending abandoned cart recommendations, called Cart Recovery; please contact Emarsys Support for details.
In this example we use the Abandoned Cart widget to include two sets of three items in a standard cart abandonment email. You could also use this campaign to offer further incentives such as discounts or free shipping. Suggested subject line: “Your LifestyleLabels Cart was left behind. Don’t miss out on these items.”

How to create this program
In the Automation Center, select the Create Program From Blueprint option.
In the blueprints list, select Cart abandonment from the Website behaviour category. Then add the segments and filters described below, and of course the email you want to send.
In this program we have included a simple A/B testing step to let you measure the success of using this simple campaign (90%) against not using it (10%). The results will be visible on the Program Summary page.
To avoid sending too many emails to serial browsers, we would suggest limiting participation to once every 30 days for this program; if incentives are included, you might even want to make this a one-time only campaign, or increase the participation period to 180 days.
Required segments and filters
- Last abandoned date was in the last 2 days
This filter is intended to check whether the data field change reflects an abandoned cart which occurred today or yesterday (in case of time zone variations).

2. Using Recommendations In Post-Purchase Campaigns
Similar to the Also_bought website recommendation widget, this is based on the last product the customer purchased and displays complementary products or products that are likely to be purchased as next purchase.
Typical use cases:
- Help convert first-time buyers into active buyers.
- Increase customer lifetime value.
- Re-purchase reminders (i.e. when the customer is likely to run out and need a refill).
In this example we have followed up on a purchase with a thank you message and some additional recommendations. For this we have used the Post-Purchase widget to display two sets of three recommendations that might also interest the customer.
Suggested subject line: "Based on what you bought, you’re going to love this."

2.1. How to create this program
In the Automation Center, select the Create Program From Blueprint option.
Within the blueprints list, this program is called Post-purchase and can be found in the Website behavior category. Then add the segments and filters described below, and of course the email you want to send out.
This campaign also uses a recurring filter, and again includes an A/B to measure its efficacy. The follow-up email is sent seven days after the purchase.
2.2. Required segments and filters
- Last purchase date was in the last 2 days
This filter is intended to check whether the data field change reflects a purchase which occurred today or yesterday (in case of time zone variations).

- Last purchase date was in the last 4 days
Although the Wait node says ‘3 days’, we also need to include the first day as well as today, so the filter must include one extra day.

3. Using Recommendations in Abandoned Browse Campaigns
The classic ‘recommended specially for you’ widget, based on browse history and the wisdom of the crowd. First-time visitors receive the best-selling products instead.
Typical use cases:
- Drive website traffic and conversions by dramatically improving click-through-rates from emails.
- Enhance the experience by letting customers discover new products they never knew they wanted.
- Reduce clicks to purchase: from email straight to product page and then to check-out.
For use in:
- Newsletters
- Browse abandonment follow-ups: ‘personally selected for you’
- Welcome emails
- Re-engagement
In this example we use the Mail Personal widget to include two sets of three recommendations in an email sent after a visitor has browsed your web shop without adding any items to the cart or making a purchase.
Suggested subject line: “You will love what LifestyleLabels has to offer you. Don’t miss out on these items.”

3.1. How to create this program
In the Automation Center, select the Create Program From Blueprint option.
Within the blueprints list, this program is called Abandoned browse and can be found in the Website behavior category. Then add the segments and filters described below, and of course the email you want to send out.
This program identifies all eligible visitors, waits for one day, and then sends the most up-to-date recommendations for each one. The two-day delay is designed to reduce the chance of coming across as overly aggressive. We would suggest a participation cycle of once every 14 days (or more) for this program.
3.2. Required segments and filters
- Last session time spent > 180 seconds
This filter includes only visitors who really spent time on your website (more than three minutes).

- Exclude: Last abandoned date was in the last 3 days
This filter ensures that the same contact will not also receive your abandoned cart program.
- Exclude: Last purchase date was in the last 3 days
This will ensure that you don’t include customers who did return and complete their purchase.
- Last purchase date was in the last 4 days
Although the Wait node says ‘3 days’, we also need to include the first day as well as today, so the filter must include one extra day.
