Overview

Recommendations V2 Analytics helps you understand how shoppers interact with products shown through recommendation widgets.

Two events are captured for recommendation experiences:

Together, these events can be used to understand recommendation visibility and shopper engagement.

Recommendation Impression

Event: experience_impression

The experience_impression event records the products displayed in a recommendation experience.

The event includes the list of recommended product IDs through pids_list, along with the recommendation widget and page type where the experience was displayed.

An impression tells you:

• How often recommendation experiences are being shown.
• Which recommendation widget is being displayed.
• Which products were presented to shoppers.
• On which page type the recommendation experience appeared.
• Which shopper visit/session the experience belongs to.

Key fields

The source notes that _uf, ver, url, and referrer can be empty strings for mobile apps.

Example

A shopper opens a product page and sees a "You May Also Like" recommendation widget containing 12 products.

The experience_impression event records those 12 product IDs in pids_list, along with information such as the widget (experience_widget) and page type (experience_pagetype).

This allows marketers to associate the products that were actually presented with subsequent shopper interactions.

Recommendation Click

Event: click

The click event records when a shopper clicks a product displayed in a recommendation experience. The event identifies the clicked product through pid and also records its position through pr.

A click tells you:

• Which recommended product attracted the shopper's attention.
• Which recommendation widget generated the click.
• Where the clicked product was positioned in the recommendation list.
• Which page type contained the recommendation.
• Which shopper visit/session generated the interaction.

Key fields

The source also notes that _uf, ver, url, and referrer can be empty strings for mobile apps.

How the Two Events Work Together

The most useful way to analyse recommendation engagement is to look at the two events together.

This creates the basic relationship between recommendation exposure and shopper engagement.

Example

A recommendation widget displays 10 products.

• experience_impression → records the 10 displayed product IDs.
• Shopper clicks the 3rd product.
• click → records that product's pid and pr = 3.

This allows the interaction to be analysed in the context of the recommendation widget and the product's position.

Marketer-Focused Analysis

The following views can be built from these events.

Recommendation Exposure

Question: How frequently are recommendation experiences being shown?

Use:

• experience_impression
• experience_widget
• experience_pagetype
• visitId

This helps compare recommendation exposure across widgets and page types.

---

Recommendation Engagement

Question: Which recommendation experiences generate shopper interactions?

Use:

• experience_impression
• click
• experience_widget
• experience_pagetype

A marketer can compare the volume of impressions with clicks to understand relative engagement.

Engagement rate / CTR calculation:
If your analytics layer calculates CTR, it can be expressed as:

Recommendation CTR = Recommendation Clicks / Recommendation Impressions × 100

The event documentation supplied here defines the underlying impression and click events, but does not specify an official CTR calculation or attribution window.

---

Product-Level Recommendation Performance

Question: Which recommended products receive the most engagement?

Use:

• pids_list from experience_impression
• pid from click

This allows recommended products to be analysed based on whether they were displayed and whether shoppers clicked them.

---

Widget Performance

Question: Which recommendation placement performs better?

Use:

• experience_widget
• experience_impression
• click

For example, marketers could compare different recommendation widgets by the number of recommendation impressions and clicks they generate.

---

Page-Type Performance

Question: Where do recommendations generate the most engagement?

Use:

• experience_pagetype
• experience_impression
• click

This can help compare recommendation performance across different page contexts, such as product pages or other supported recommendation placements.

---

Position-Level Performance

Question: Does the position of a recommendation influence clicks?

Use:

• pr
• pid
• experience_widget

The pr field identifies the position of the clicked product in the recommendation experience.

This can be used to analyse whether products appearing earlier in a recommendation list receive more clicks.

Important Fields at a Glance

The underlying event definitions and field availability are provided in the Recommendations event specification.

Troubleshoot and FAQ

Q. How can I tell if recommendations are being shown to shoppers?

A. Look at the experience_impression event. This event records when a recommendation experience is displayed to a shopper.

Q. How can I tell if shoppers are engaging with recommendations?

A. Look at the click event. This event records when a shopper clicks a product from a recommendation experience.

Q. How can I compare the performance of different recommendation widgets?

A. Break down recommendation impressions and clicks by experience_widget. This helps you compare how different recommendation placements or widgets perform.

Q. How can I compare recommendation performance across different page types?

A. Break down impressions and clicks by experience_pagetype. This helps you understand which page types generate better recommendation engagement.

Q. How can I identify which recommended products attract the most attention?

A. Use the pid from the click event and compare it with the products presented in pids_list in the experience_impression event.

Q. How can I understand whether product position affects clicks?

A. Use the pr field from the click event. It represents the position of the clicked product within the recommendation experience and can be used to analyse click performance by product position.

Q. Can I connect recommendation engagement to a shopper's visit?

A. Yes. Use visitId to associate recommendation interactions with a shopper visit. You can also use visit_type to distinguish between new and repeat visits.