Connected Recommendations overview & configuration

Connected Recommendations allow you to deliver data-driven product and content suggestions within your Customer Engagement experiences.

By linking Customer Engagement with Product Discovery's recommendation engine, you can use recommendation algorithms directly inside Design Studio modules. This allows you to present tailored recommendations to users based on dynamic data such as SKUs, filters, or customer data, all configured seamlessly within Customer Engagement.

How it works

Connected Recommendations use Product Discovery's recommendation models directly inside Customer Engagement's Design Studio.

The recommendation flow works as follows:

1. A module in Design Studio is configured with metadata (for example product rules, SKU input, or user ID)
2. When the email is preprocessed, Customer Engagement sends this metadata to Product Discovery
3. Product Discovery processes the request using its recommendation algorithms
4. A set of optimized recommendations is returned
5. Customer Engagement renders the recommendations directly in the experience

If a discoveryKey is included, purchased products can be excluded, both online and in-store (if applicable). Read more about these concepts within Connected recommendations:

• Top selling products
• Contextual products

Setup & requirements

Here is what you need to set up.

discoveryKey

A discoveryKey is required to use Personal Products. Without it, "LikelyToBuy" and "Based on Favorite" will fall back to top sellers. "Recently Viewed" requires a discoveryKey and will not be displayed if one is unavailable. When a discoveryKey is present, it also enables purchased products, both online and in-store, to be excluded from recommendations across all modules.

Read more: discoveryKey implementation

SKU and variantKey alignment

For Contextual Products to work correctly, SKU values in your Customer Engagement product feed must match variantKey values in Product Discovery. Mismatches will prevent contextual triggers from resolving and recommendations will not be served.

Read more: SKU and variantKey alignment

Marketing groups

Before you can use Connected Recommendations, you must configure the connection between Product Discovery markets and Customer Engagement marketing groups. This is done once in Customer Engagement and establishes the persistent connection between the products.

See detailed guide: Configure marketing groups for Connected Recommendations

Email module

Each Connected Recommendation data source includes a ready-to-use module in the starter library, making it easy to get started. You can use these modules as they are, but in most cases you will want to adapt them to match your brand’s design, layout, and visual identity.

See more details about modules: Working with modules

Longer delivery period

Connected Recommendations uses live data and processes each contact individually. As a result, the sendout will be distributed over a longer period.

Image requirements and limitations

Image styling in email

Connected Recommendations do not support CSS-based image manipulation inside emails. If you previously used Product Discovery email recommendations and applied CSS to the product template to adjust image ratio, cropping, or scaling, that approach will not work with Connected Recommendations.

Email clients have limited CSS support. This means:

• Product images must already have the correct and consistent aspect ratio
• Images must be cropped exactly as they should appear in the email
• Image formatting must be handled in the source feed or during asset generation

Image formats and sizes

Connected Recommendations use the Product Discovery feed to fetch metadata, including image URLs.

For the broadest compatibility across email clients, we recommend using JPEG or PNG images. Formats such as WebP, AVIF, and SVG are supported by modern web browsers but may not be supported by all email clients. This is a limitation of the receiving email client, not Voyado.

If your Product Discovery feed uses one of these formats, contact Voyado so we can review your setup and recommend a compatible approach for email.

The required image size depends on the module layout. In the standard three-column layout, product images are rendered at approximately 186 px wide. Source images should therefore be at least 186 px wide, but we recommend using images 2–3 times the rendered size (approximately 370–560 px) to ensure sharp rendering on high-density displays. Larger source images are scaled down when rendered.

There is no corresponding height requirement. The image height can vary, as the email content simply expands vertically to accommodate the image.

Email recommendations comparison

Voyado offers three approaches to email product recommendations. Connected Recommendations is the most integrated option, purpose-built for customers running both Customer Engagement and Product Discovery, with native Design Studio authoring, Product Discovery's recommendation engine, and support for dynamic trigger-based recommendations. The table below outlines how the three approaches compare.