What This Does
Your ESP is good at sending. What it can’t easily do is decide, per recipient, which articles to show, and keep that decision fresh between the moment you hit send and the moment someone actually opens the email.
Open-time content recommendations solve that. You embed one HTML snippet in your campaign template. Every time a recipient opens the email, the snippet calls Personyze live, identifies that person, and renders the articles they’re most likely to read right then, based on what they’ve been reading on your site.
This is the same open-time mechanism used for product recommendations in 3rd-party emails. If you’ve set that up, this will feel familiar: the difference is the catalog behind it (articles instead of products) and the algorithms available.
In the panel: New campaign → Get Code for Emails: Content Recommendations. The wizard has eight steps: Catalog, Content events, Recommendation, Look, Code, QA, Automation and Performance.
Why Open-Time Matters More for Content
Editorial content ages fast. A newsletter built on Monday and opened on Thursday can recommend an article the reader already saw, or miss the piece published Tuesday that’s now the obvious pick.
Because the snippet resolves at open time, the recommendations reflect:
- what the recipient has read since the email was sent;
- content published after the send;
- what’s trending across your audience right now.
Prerequisites
- An article catalog in Personyze: your articles, guides or videos, each with at least an ID, a title and a URL, and an image if your template shows one. It lives in Settings → Recommendation setup → Article Catalog. See Setting the Product / Content Feed and Uploading Feeds & Data Files.
- Article tracking on your site, so Personyze knows what each visitor reads. Set it up in Settings → Recommendation setup → Article Tracking.
- A way to identify the recipient: the email address (or your own user ID) filled into the snippet by your ESP’s merge tag. It only personalizes if the same address or ID is on the reader’s Personyze profile, captured on your site (a form, a login or the JavaScript API) or sent through the server API.
How to Set It Up
- Catalog and Content events. Confirm the article catalog is loaded and that reads are arriving. If content recommendations already run on your site, both are in place; if not, each step offers a guided setup.
- Recommendation. Choose the algorithm. It is the same picker as for on-site widgets, but the Page row offers what the reader did instead of a page of your site: use General for newsletters and Articles they read to follow what they read. The algorithm you choose is written into the template you pick next, on Look, and What this block recommends beside it counts what your catalog and settings leave to send. Strong choices for email are Content Recommended for You, Content Recommended for You — Published Since Last Visit, Trending Now, Most Popular and Recently Published. What each one picks is on Recommendation Algorithms.
- Look. The step shows the newest templates as pictures; click one and the block starts from it, or choose Browse template gallery for the rest — for example Article Grid, Article Rows or Article Strip. Then adjust colours, text and layout. Each article is rendered as an image.
- Code. Choose the identifier (Email, Internal User ID, or Content ID for picks that are the same for everyone), then your email platform. Its merge tag is written into the snippet for you. Copy the snippet.
- Paste it into your ESP template, in a Code, HTML or Custom Code block.
- QA. Preview the exact images a reader would get: look up a recipient, anchor to an article, or choose New recipient to see what an unknown reader gets.
Merge Tags by ESP
The snippet needs the recipient’s email address filled in at send time. The Code step knows the tag for 13 platforms, including Mailchimp, Klaviyo, HubSpot, ActiveCampaign, Brevo, Salesforce Marketing Cloud, Braze, SendGrid and Iterable. For any other platform, choose Custom and enter its tag. The full tag list and a platform-by-platform walkthrough are on the product recommendations embed page. The tags are identical here, since it’s the same snippet mechanism.
Fallbacks
Some recipients won’t have enough reading history to personalize against, such as a brand-new subscriber or someone who only ever opens email. In the Recommendation step’s Output card you have two controls for them:
- Fallback algorithms → Add fallback: algorithms that run in order when the main one returns too few articles. For example, Most Popular or Recently Published behind Content Recommended for You.
- Fill empty cells: when this is on, a short or empty result is topped up with your most popular articles, then the rest of the catalog. When it is off, an empty result is sent as blank space rather than a broken image.
Check the result in the QA step with New recipient → Preview the fallback. More on Fallbacks.
Performance
Personyze does not send the email, so the Performance step starts at Viewed (a recipient’s email client showed the widget). Then comes Engaged (the reader clicked a recommended article and reached your site) and Converted (the reader went on to reach your content goal in the same session). Each stage is a % of the one before. Every card has a Users … link to the people behind the number. What each figure means: Email, SMS, web push and 3rd-party emails.
Related
- Recommendations Outside Your Site — Where to Start: every off-site surface, side by side.
- Open-Time Email Recommendations: the same mechanism for products, with the full ESP-by-ESP walkthrough.
- Recommendation Algorithms: content recommendations
- JSON API — Content Recommendations, if you’d rather render the picks yourself.
- Uploading Feeds & Data Files