Drop dynamic recs into any email.
A code-level playbook for email recommendation widgets — generate per-recipient product and content picks, embeddable in any ESP via simple code.
Start Now →1How to Guide: Open-Time Email Recommendations
This guide explains how to generate an HTML snippet to embed open-time email recommendations into any third-party email provider. To begin, create a campaign of the type Get Code for Emails: Product Recommendations or Get Code for Emails: Content Recommendations.
Prerequisites
For this to work, you must have the following set up in Personyze:
- Catalog Feed: A Product or Content feed must be synced.
- Tracking Code: The Personyze tracker must be active on your website.
- Interactions: You must be tracking events (View, Add to Cart, Purchase) so the AI has data to work with.
- User Identification: Configure Personyze to capture the email or your own user ID during web sessions (via login, URL parameters, or cookies) so their onsite history is linked to the email recipient.
Step-by-Step Summary
- Select Wizard: Choose the Product or Content wizard for 3rd party emails.
- Design: Style your recommendation cells (Fonts, Colors, Badges).
- Algorithm: Select your logic (e.g., "Cross-Sells for Items in Cart") and a fallback for recipients Personyze does not know.
- Get Code: Copy the HTML snippet and replace the placeholder with your ESP's merge tag.
- Analyze: Track Views, Clicks, and Revenue in the Personyze dashboard.
2Design & Algorithm Selection
First, design the look of your recommendation block using the WYSIWYG editor or AI tools.
Design & Layout
- Templates: Choose a template in the design step, then change the layout, the card and the title.
- Styling: Adjust fonts, button colors, prices, and image sizes to match your brand.
- Dynamic Badges: Overlay automatic labels to drive clicks:
- Corner badges: the discount or sale percent, a price drop, the stock
- Stacked badges: New, Popular, Best Seller, time-based offers
Algorithm Logic
Define the logic that determines *which* products appear.
- Primary Algorithm: Choose based on goal (e.g., "Cross-Sells for Items in Cart", "Personalized Recommendations", "New in Stock").
- Fallback: Personalization is best-effort. If the merge tag is empty, the address is unknown to Personyze, or the recipient is brand new, the block serves your fallback items — or, with Fill empty cells on, your most popular ones. Never a broken image.
3User Recognition Methods
To personalize the content, you need to tell Personyze WHO is opening the email. You configure this in the "Get Code" step.
Option A: Email or your own user ID (Personalized)
The best option for "Recommended for You" logic. You will insert your ESP's merge tag (e.g., *|EMAIL|* for Mailchimp) into the Personyze code.
- Requirement: The user's email or user ID must also be known to Personyze via Site Grab, CRM Sync, or prior login.
Option B: Product ID (Contextual)
Good for transactional emails. "You bought Item X (pass ID), here are accessories for X." No user history needed.
Option C: Non-Personal (Generic)
If no ID is passed, Personyze uses the Fallback logic (e.g., Most Popular) for everyone.
4Get & Embed the Code
This step generates the HTML snippet you will paste into your email template.
1. The Code Structure
You will receive an HTML block (usually a <table> structure) containing <img> and <a> tags. These link to Personyze servers.
2. Populate Dynamic Variables
The code will contain placeholders like {INSERT EMAIL HERE}. You must replace this string with the dynamic Merge Tag from your Email Service Provider (ESP) so that every recipient gets a unique link.
Mailchimp: Replace with
*|EMAIL|*Klaviyo: Replace with
{{ email }}Salesforce: Replace with
%%emailaddr%%
3. Tracking Parameters
In the template editor, you can wrap every product link with unique UTM parameters (e.g., utm_source=personyze_email) to track clicks in Google Analytics.
5QA & Preview
Before launching, verify exactly what specific users will see.
Simulation
- In the block’s design step, open Preview as — or open the snippet’s image link with a test email in place of the merge tag.
- Enter a real email or user ID from your database.
- Personyze will generate the recommendation images exactly as that user would see them based on their history.
Example Check: Does a user who bought "Cat Food" see "Cat Toys" recommendations?
6Performance & Analytics
Measure the impact of your email widgets in the Personyze Dashboard.
Content Performance Metrics
- Views: Total impressions (times the email images were loaded).
- Clicks: Total clicks on recommended articles.
- CTR (Click-Through Rate): The percentage of views that resulted in a click.
- Goal Conversion Rate: The percentage of sessions where any defined goal (e.g., Newsletter Signup, PDF Download) was completed after the click.
Commerce Performance Metrics
- Added to Cart: Recipients who added items to their cart after clicking.
- Purchased: Recipients who completed a purchase on your site after clicking, with the revenue from it.
7Example: E-Commerce Journey
A complete walkthrough of how a visitor's site activity translates into a personalized email purchase.
Step 1: Visitor Activity
Visitor John Doe (john@example.com) lands on your site via a Google Ad for "Running Shoes".
- View: He views the "Nike Air Zoom" product page.
- Action: He adds "Nike Air Zoom" to his cart but abandons the site without purchasing.
- Identification: Personyze identifies John because he logged in previously OR clicked a newsletter link in the past (Email captured via Site Grab).
Step 2: The Email Trigger
Two hours later, your ESP (e.g., Mailchimp) triggers an automated "Cart Abandonment" campaign to john@example.com.
Step 3: Open-Time Rendering
John opens the email. The embedded Personyze code sends a request to the server with his ID.
1. Main Item: "Nike Air Zoom" (The exact item left in cart).
2. Cross-Sell 1: "Nike Running Socks" (Algorithm: Accessories for Cart Items).
3. Cross-Sell 2: "Running Shorts" (Algorithm: Bought Together).
Step 4: Performance Metric
John clicks the "Running Socks" image in the email.
- Metric: "Purchased" counts John, and his order’s revenue is credited to the email.
8Example: Publisher Journey
How a news or content site drives engagement through personalized newsletters.
Step 1: Visitor Activity
Visitor Sarah Smith visits your Tech News portal.
- Context: She reads 3 articles about "Artificial Intelligence" and 1 about "Cybersecurity".
- Identification: Personyze tracks this interest profile under her user ID (captured when she signed up for the free newsletter).
Step 2: The Email Campaign
On Friday, you send your standard "Weekly Tech Digest" blast to all subscribers.
Step 3: Open-Time Rendering
Sarah opens the email on Saturday morning.
Instead of generic "Top News", Personyze renders a "Recommended for You" block:
1. "New AI Regulations Released" (Algorithm: Content Recommended for You — Published Since Last Visit).
2. "Top 5 AI Tools for 2025" (Algorithm: Content Recommended for You).
Step 4: Performance Metric
Sarah clicks the "AI Regulations" article.
- Metric: CTR for the campaign increases.
- Metric: "Goal Conversion" triggers if she shares the article or upgrades to Premium.
Playbook questions, answered.
Common questions about dynamic email recommendation widgets. Anything else, our team is one message away.
How do open-time recommendations work technically?
The code you embed is essentially HTML <img> tags. When the recipient opens the email, their mail client requests these images from Personyze servers. We generate the image in real-time based on the user's data and current inventory, ensuring the content is fresh at the moment of open, not when sent.
Does this work with my Email Service Provider (ESP)?
Yes. The generated code is standard HTML that works with virtually any ESP, including Mailchimp, Klaviyo, Salesforce Marketing Cloud, HubSpot, and others. You simply need to replace the placeholder variables with your ESP's specific Merge Tags (e.g., *|EMAIL|*).
Can I use JavaScript inside the email?
No. Email clients (Gmail, Outlook, etc.) do not support JavaScript for security reasons. That is why Personyze uses dynamic images to deliver personalization, bypassing the need for scripts.
Why is a "Fallback" algorithm required?
Personalization is best-effort: a recipient can open on a new device, the merge tag can be empty, or the address can be new to Personyze. Then the block serves your fallback items — or, with Fill empty cells on, your most popular ones — so the email never shows a broken image.
What if I don't have the user's email or user ID?
Use a product ID instead for non-personal picks: related items for whoever opens it, such as what is bought together with the product in the email. With no identifier at all, the block serves your fallback items.
Can I show "Items Left in Cart" in the email?
Yes. Choose the Left in Cart Items algorithm — or Left in Cart Items Now Discounted, or Cross-Sells for Items in Cart — and the block renders that recipient’s own cart when the email is opened.
Do I need to upload my product catalog?
Yes. Personyze needs your product or article catalog — from a connected store, a feed or a file — to know which images, titles and prices to show.
Does the content update if I open the email days later?
Yes. Since the image is generated at the moment of request, opening an email on Monday might show "5 items in stock," and opening the same email on Friday might show "Sold Out" or a completely different product recommendation if the original is out of stock.
Can I use this for Articles and Content?
Yes. Create a Get Code for Emails: Content Recommendations campaign and recommend articles with algorithms such as Content Recommended for You or Content Recommended for You — Published Since Last Visit.
How do I create urgency in the email?
Badges are drawn into the image when the email is opened: a stock badge shows how many are left, and a corner badge shows the discount or a price drop.
How do I track the performance of these emails?
Under Campaign performance → Email Get Code: how often the blocks were viewed (opens), the clicks, and what happened on your site after them — added to cart and purchased.
Can I preview what a specific user will see?
Yes. In the block’s design step, Preview as picks a real recipient and shows the products they would get, with the algorithm behind each one. You can also open the snippet’s image link with a test email in place of the merge tag.
Can I customize the look of the recommendations?
Yes. The design step has templates, layout, fonts, colours, prices and badges, and HTML Source (advanced) for full control — so the block matches your email’s style.
Is it possible to use social proof in emails?
Yes. Stacked badges mark products as New, Popular or a Best Seller, and corner badges show the discount, a price drop or the stock — all decided when the email is opened.
What is required to set this up?
You need three things: 1) A Catalog Feed synced to Personyze, 2) The Personyze Tracking Code on your website to capture user behavior, and 3) A method to identify users (e.g., capturing email during login or signup) to link their web history to their email address.
Personalize every email.
You have the full playbook — now embed AI-driven recommendation widgets into every email you send.