Byredo

BeautyFashionRetail

Byredo ships 13x more often, and passes every speed check Google runs.

The global fragrance house moved off a monolithic Magento frontend that had not changed since 2018. The speed numbers on this page come from Google’s record of real visits to byredo.com, so anyone can check them.

Request a demo See the numbers

Industry: Fragrance, 40+ countries
Backend: Magento
CMS: Contentful
Implementation: Vaimo, with Alokai on review
Live since: 2025, measured June 2026

13x

more releases, from one a month to one a day

0.01

layout shift on mobile, where Google allows 0.10

All 3

Core Web Vitals in Google’s "Good" tier, mobile and desktop

Zero

maintenance pages: the store stays open through every release

Releases went from monthly to daily, and every speed measure improved.

August 2024 on the old Magento setup against June 2026 on Alokai: same brand, same catalogue, same shoppers, measured the same way.

June 2026 · Alokai

13x

more releases, up to one a day

85.8%

of mobile visits loaded fast enough for Google

83.9%

of mobile visits had no part of the page jump

The frontend runs separately on Alokai Cloud. The team can release daily when the brand needs it, the store stays open while they do, and all three Core Web Vitals pass on both phones and desktops.

Visits Google rates as good

Real Chrome visits

Mobile · loading speed85.8%

Mobile · response to taps86.8%

Mobile · page staying still83.9%

Desktop · loading speed92.2%

Desktop · response to clicks95.1%

Desktop · page staying still81.2%

Mobile is 63.6% of Byredo’s traffic, desktop 36.4%. These numbers cover every page on the site, not one page picked to look good.

One release a month, and a frontend nobody had changed since 2018

Byredo is a fragrance house in the Puig group, selling in more than 40 countries. Its storefront ran on one standard Magento application that had done the job for years, then started holding the team back. Three problems showed up at once.

A month from idea to live
Every change went through the full cycle of spec, build, test and release. A big release meant a maintenance page and a store closed to buyers.
One vendor decided the stack
Search, payments and content came bundled with Magento. Even when a better tool existed, using it meant a rebuild.
A frontend frozen since 2018
The design stayed the same for years because even a small change needed a developer. The team could not touch the details the brand is built on.

“We stayed quite static for a very long time in terms of web design. Each time we wanted to change something, we had to go through the full cycle. It was anything but efficient.”
Arnaud Fritz, CIO Beauty & Fashion

Byredo replaced the frontend, kept Magento, and chose the rest itself.

Byredo looked at going mobile-first and at rebuilding on Magento’s own frontend, then chose to run the frontend separately on Alokai. Vaimo built it, with Alokai’s team reviewing the code and connecting the parts.

Frontend platform

Alokai

Runs the storefront shoppers use, connects the systems below it, and ships releases without touching Magento.

Commerce backend
Magento
Kept
Kept exactly as it was: catalogue, prices, orders, checkout.
Content
Contentful
The team edits pages without waiting for a release.
Search
Klevu
Chosen for search alone, and replaceable later.
Payments
Adyen
Chosen on merit, not inherited with the shop software.

“Alokai resonates with us for being more than just a toolbox. It helps us aggregate all components, bring them together. And this gives us the agility without feeling any constraints.”
Arnaud Fritz, CIO Beauty & Fashion

Byredo stays well inside every limit Google sets

Each bar shows what Byredo measured against the limit Google allows. The shorter the bar, the more room to spare.

Loading speed (Largest Contentful Paint)

1,943 ms of 2,500 ms allowed

22% faster than Google allows

Response to taps (Interaction to Next Paint)

133 ms of 200 ms allowed

33% quicker than the limit

Page staying still (Cumulative Layout Shift)

0.01 of 0.10 allowed

Almost no movement, even on image-heavy pages

First content on screen (First Contentful Paint)

1,738 ms of 1,800 ms allowed

Passing, with the least room to spare

Where these come from: Google records how real Chrome users experience a site and publishes the result. These figures cover every page on byredo.com rather than one page tuned for a test, and they use the same measurement Google uses to decide whether a site passes.

“In GSC, all our pages are classified as the fastest. Meaning that in SEO, we can’t be any better. We are at the top.”
Arnaud Fritz, CIO Beauty & Fashion

The engineering gains showed up in the business numbers

Releases got faster and so did the storefront. Both of those change what shoppers do next.

For the engineering team

Shipping stopped being the bottleneck

1 month → 1 week: Lead time on a frontend change, with 2 to 3 releases a week when the business asks for them.
133 ms / 74 ms: Time to respond to a tap or a click, both well inside Google’s limit.
Zero maintenance pages: Frontend releases no longer close the store, however big the release is.

Bring your own numbers to the call

For the business

Faster pages have a measured payoff

+8.4% conversion: For every 100 ms saved on mobile, in Google and Deloitte’s retail data.
+9.2% order value: For every 100 ms saved, in the same data.
32.7% → 83.9%: Of mobile shoppers now see a page that stays still, so fewer accidental taps in checkout.

See it on your catalogue