ItalyConsumer goods2022 – ongoing

Benetton Group

Personalized product recommendations on the online store

Benetton Group's online shoppers who used the AI recommendation panel converted at 6 times the rate of those who did not

WorkingVendor case studyVendor ↗

Problem

Benetton Group's content management system had a built-in recommendation tool, but the company wanted a more sophisticated one for online customers. It wanted to use browsing behavior to personalize suggestions across a global customer base.

Approach

Benetton Group trained models in Google Cloud's Recommendations AI on 90 to 120 days of data per country from Google Analytics and BigQuery. Countries were grouped by language to build customized models. After a proof of concept with one model, the company rolled out the solution within three months. The models fill panels such as Frequently Bought Together on product pages and Recommended for you on the checkout page. The company A/B tested a redesign of the home page, product pages and cart around the panels, and released the panels in its top five markets.

Results

6 timesconversion rate of users who clicked the Recommendations AI panel compared with users who did not, per Google CloudVendor case study[1]

7%higher average sales through the recommendations panel than sales that did not come through it, per Google CloudVendor case study[1]

almost 3 timeslonger time on the site for customers who engaged with the panel than for those who did not, per Google CloudVendor case study[1]

Takeaway

The comparison is between shoppers who clicked the recommendations panel and those who did not, so it does not isolate the effect of the AI. Benetton Group was extending the panels to its other markets.

Sources

Benetton Group: Sharing bespoke recommendations with AI and cloud-based analytics
Google Cloud6 May 2023Vendor publication

Source summaries

[1]Google Cloud6 May 2023

Google Cloud's customer story describes how Benetton Group built personalized product recommendations for its online store with Recommendations AI and BigQuery data. It reports that users who clicked the recommendations panel converted at 6 times the rate of those who did not. It also says those sales were 7% higher on average and those customers stayed almost 3 times longer on the site. The page carries no date; the Internet Archive first saved it on May 6, 2023.

Full details are in the original sources.

Added 4 Nov 2026 · Updated 4 Nov 2026