When someone opens an email from you, you have precious little time to engage him before he decides what to do with that message. Either he will click through and visit your site, or he will dismiss the message. Machine learning helps guarantee that what you put in front of each person is relevant to that individual based on his or her affinities and intent. Displaying that content dynamically at open time ensures that your message remains relevant regardless of when your recipient chooses to view it.

While historically, email has been limited to static content or content personalized at send time, a recent Evergage blog post illustrated why providing real-time email content should be important to all email marketers.

In this article, I’ll cover how to do this for one sample use case — injecting open-time dynamic product recommendations into a retail marketing email.

1. Set Your Recommendations Recipe

The first step is to create your recommendations “recipe” that fits your requirements. Your recipe can be as simple an idea as, “recommend the most popular items by view time in the last seven days.”

Or, you can create a more sophisticated recipe that utilizes advanced machine learning to provide true one-to-one recommendations based on an individual’s behavior, taking into account that individual’s affinities for certain brands or product categories. An example of a more sophisticated recipe idea might be, “make personalized recommendations based on an individual’s favorite brands and styles, but exclude any items that individual has purchased in the past, currently has in their cart, or that are below a certain price.”  

The recipe you create here will determine the items that your customers see when they open your email.

2. Design Your Item Templates

An item template is a block of HTML that Evergage uses to generate product images for your entire catalog that can then be inserted into an email. These templates allow you to tailor your recommendations so they harmonize with the style of your email and site.

For example, here’s an item template that’s driving the visuals in an email and on the website.


3. Design Your Email

Using the email campaign editor, you can craft a campaign that uses your item template to issue dynamic recommendations at email open time, because the image in the item template can be regenerated on the fly.  

The example below shows a row of four recommended items using a collaborative filtering recipe to provide personalized recommendations that are boosted by the individual’s gender affinity. This means that users who have predominantly shopped for menswear will receive recommendations for menswear, and the same is true for shoppers who have predominantly shopped for women’s clothes. Shoppers who have not demonstrated a strong preference either way will receive a mix of recommendations.


When this campaign is published, the campaign editor generates a block of HTML that can be easily placed into your email’s layout.  

4. Send Email and Deliver 1:1 Personalization Dynamically at Open Time

With this approach, each recipient of your email gets a personalized experience that is always relevant, regardless of when they open the email.  

To continue the example, when a customer who has shopped primarily for women’s clothing opens this email campaign, she will see an experience that recommends women’s items.


The dynamic elements in this email are the four product recommendations. If a male-affined customer opens the very same email, he receives a totally different experience, seeing recommendations for menswear.


These two unique experiences are driven by exactly the same HTML.  

But here is where things really get magical. Let’s say the Merino V-Neck Pullover Sweater above suddenly sells out, and perhaps the Chelsea Tee is marked down from $75.00 to $50.00. Personalization decisions that occur at send time are unable to adjust to these changing circumstances. But because dynamic open-time emails make personalization decisions at open time, the email always contains up-to-date content. The same customer, perhaps opening the email later in the day, will receive an up-to-date experience.


Each time an email is opened, Evergage issues recommendations based on the current state of the catalog, so the message now reflects the markdown on the Chelsea Tee and the now out-of-stock sweater is no longer offered as a promoted item.

5. Reuse Campaigns

It is tempting to create a new email campaign each time a new batch of emails needs to be sent. This is, after all, a habit that many email marketing platforms teach us. But this isn’t always necessary. If your goal is to send a weekly marketing email to your customers with recommended offerings, it is possible to reuse the same email campaign. This means that nothing needs to change in your email’s HTML layout — you can simply resend the same campaign over and over and Evergage will take care of recommending personalized and timely products each time. If needed, the recommendations recipe can be tweaked on the fly.

Just Scratching the Surface

This blog post has illustrated a fairly straightforward retail example, but dynamic open time emails driven by machine learning work equally well for other content types. Content providers can tailor recommendations for articles and blogs just as easily as for products. B2B marketers can tailor their email’s hero image or other content in their messages to provide unique experiences to customers. You can even A/B test different dynamic elements against each other. For example, you may wish to try testing your email’s existing message against one that contains dynamic recommendations to measure the lift that can be achieved by adding them.  

Evergage for Email allows you to accomplish all of the steps outlined in this post. To learn more about how Evergage can help you dynamically change the content in your emails, request a demo today or discuss it with your Customer Success Manager!