How Primark scales content localization across 18 locales

Eighteen locales. Thirteen new articles a month. And one small team that doesn’t speak 18 languages.

That’s the content localization challenge Primark, the international fashion retailer with over 400 stores worldwide, handles with Amplience CMS, part of the Amplience Content Operations Platform. Content is created once in English, translated through automated workflows, adapted for each market and kept up to date across every locale, without the team growing to match.

Key takeaways

1. Primark, the international fashion retailer, runs content localization for primark.com across 18 locales on Amplience CMS, part of the Amplience Content Operations Platform.

2. Content is created once in English, and workflow statuses in Amplience CMS automatically send it to Primark’s translation agency and drop the translations back into every localized version.

3. A mass update workflow carries image and product changes across all locales in one action, without paying to translate the same copy again.

4. In a proof of concept with Amplience Workforce, Primark created 70% more product descriptions a day and 150% more SEO articles a month, and saved 9 hours a month on alt text.

5. Article output has grown from three a month to 13, while a content audit is merging duplicate articles so Primark stops competing with itself in search.

Why content localization at Primark kept getting harder

Primark keeps expanding, but its digital design team is still small. The team is led day to day by Jordan Longhurst, Primark’s Digital Design Lead. New markets keep launching, with Hungary and Slovenia among the most recent, so the same team now manages content for 18 locales. Everything starts in English, in the UK master locale, and has to be translated and adapted before it can go live anywhere else. Every new locale means more content to translate, check and keep up to date.

Not every locale needs the same content at the same time. Primark sorts its locales into three tiers. The UK is tier one. Tier two covers mostly European markets with a large store presence. Tier three covers markets with only one or two stores, which may not get every collection. The UK’s content is updated at least once a week and tier three’s roughly once a month, while major campaigns go out to every locale at once.

Even within a tier, content often has to change by market. Some products aren’t sold everywhere, influencer campaigns that work in one country may mean nothing in another, and the assets Primark can use sometimes differ from one market to another.

The volume of editorial content keeps growing too. Five years ago, the team published three articles a month. Now it’s 13, covering everything from new and sustainable collections to Primark’s adaptive clothing range and styling guides that answer common searches, like what to wear to a spring wedding.

Ecommerce localization for each Primark market

The same workflows let the team choose which content goes to which market. Primark doesn’t sell much skincare in the US, for example, so skincare content isn’t sent there.

Campaigns can vary by market too. Primark recently launched a collection created with a Spanish influencer who isn’t as well known outside Spain. Every region got a general campaign shoot showing the collection, while Spain also got a shoot featuring the influencer herself.

Mass updates across 18 locales without paying to translate twice

Products sell out and carousels need refreshing. If every small update had to go back through translation, Primark’s translation costs would keep climbing. So the team built a mass update workflow. One change in one locale carries across to every other locale automatically.

It only updates linked content, such as images from Amplience DAM and product information. It never touches copy, so there’s no risk of English text appearing in another language’s page.

AI alt text in Amplience DAM that knows the Primark brand

Every image on primark.com needs alt text, for accessibility and for SEO. Writing it by hand took about two minutes per image, with at least nine images per article and more on every banner in every region.

Primark tested AI-generated content with Amplience Workforce in a two-month proof of concept, starting with a workshop day for around 30 people. Month one covered product descriptions, alt text and SEO articles in English, and month two tested AI translation of the same content. The team scored every output out of five across six areas, from brand consistency and product accuracy to how often they had to regenerate it, and reviewed the scores with Amplience every week.

Because Amplience isn’t tied to a single AI model, Primark could compare several, and it tests new ones as they launch before switching. ChatGPT worked best for informational content like size guides. Claude became the team’s favorite for alt text and product descriptions, because it described the clothes in far more detail, from print and pattern to cut and style, and sounded more like Primark’s own writers.

Licensed ranges, such as Disney and Marvel, come with strict rules from each brand on how products and characters can be described. Primark set up a separate brand in Amplience Workforce for each license, so all of its rules sit in one place. That meant far less editing for the team at the end.

Now, alt text is generated automatically as images are uploaded to Amplience DAM, pulled straight into content and translated as part of the same workflow.

A content audit for fewer, stronger Primark articles

More output doesn’t always mean more pages. Primark used to publish a new article every time it reshot a familiar product, like its Christmas onesie. Over time, those articles started competing with each other in search. The team is now running a content audit across thousands of pages, merging duplicates and updating existing pages instead of creating new ones. Because content in Amplience CMS is structured and reusable, the team can refresh what’s already there rather than starting from scratch each time.

What’s changed for Primark with Amplience

For Jordan, the point of all this automation is to take repetitive tasks off the team. Primark’s designers can now focus on design and on building new workflows with Amplience, instead of writing copy they were never hired to write.

The writers have come round to AI too. Jayne Bibby, Primark’s Senior Digital Content Manager, who oversees the editorial and digital design teams, says the proof of concept eased early fears about AI taking jobs and showed the team how much it still needs them.