Platform
PricingSign inBook a demo
← All articles
PersonalizationJuly 12, 2026

Multilingual Personalization: Content, Recommendations, and Messaging in Every Language

P
Personyze TeamPersonalization experts
Multilingual Personalization: Content, Recommendations, and Messaging in Every Language

A visitor from Milan, a customer in Berlin, a shopper on vacation abroad — a global site has to speak each of their languages and reflect each of their markets. Real multilingual personalization goes past translation: the content, product and content recommendations, promotions, and messaging all adapt to the visitor’s language and location.

Here’s how to do it in Personyze — how it detects language and location, how to vary content and messaging, how to serve recommendations from the right localized feed, and how it handles the tricky case of a customer who’s traveling.

How Personyze knows a visitor’s language and location

Everything starts with two signals: language and location.

For language, Personyze can read your site’s own language selector (from the URL or a page element), a cookie value that stores the chosen language, or the browser language as a fallback — and you can combine them.

For location, it uses IP geolocation (country, region, city) or, for identified visitors, CRM / account data (easy to integrate). Language and location can be combined into a single condition, so you can target “Italian speakers in Switzerland” as easily as “anyone in Italy.”

How Personyze detects language and location
Language from your site selector a cookie or the browser location from IP or CRM combined into one resolved locale

Content, promotions, and messaging by language

With language and location resolved, you can drive content variations across the site. In Personyze you set up cases like “if the visitor’s language is Italian, show this hero, this banner, this message” — then a different variant for German, French, and so on.

Because location is a signal too, you can go further with combinations: Italian in Italy sees one promotion; Italian in Switzerland sees another, with different currency, shipping, or offer. The same website personalization mechanism localizes headlines, CTAs, images, popups, and promotional banners — not just translated copy, but the right message for the right market.

One template, many messages: conditional content

You don’t need a separate campaign per language. In Personyze, a single template can speak to every visitor differently through cases: you set up a list of conditions, each with its own content, and Personyze shows the first one that matches — with a default for everyone who doesn’t. That lets one template swap the headline, copy, or CTA per segment, in real time, with nothing extra to maintain.

This is what it looks like in Personyze — you add each case, choose its condition, and type the content, right in the visual editor:

Building audience based content variations in the Personyze visual editor
Building audience based variations in Personyzes visual editor no code required

And if you’d rather do it programmatically, the same cases look like this:

${case}
${when Locate('Italian', language)}Offerte pensate per te
${when Locate('German', language)}Für dich ausgewählte Angebote
${when Locate('French', language)}Des offres pensées pour vous
${else}Deals picked for you
${end}
One template rendering a different headline per language
Personyze evaluates each condition in order and outputs the first match so one template renders the right headline for each visitor

Personyze checks each condition in order and outputs the first match, so an Italian speaker sees one headline, a German speaker another, and everyone else gets the ${else} default.

And it isn’t limited to language — swap language for any attribute you track (country, plan, returning vs. new, page topic, or industry) and the same few lines cover as many variations as you need:

${case}
${when Locate('Ecommerce', industry)}Turn more browsers into buyers
${when Locate('SaaS', industry)}Turn more trials into paying users
${when Locate('Publisher', industry)}Keep readers reading, longer
${else}Personalize every visit
${end}

That’s relevance at the field level: one asset, driven by your own visitor data, rendering the right message for each segment.

Multilingual product and content recommendations

Recommendations are where localization gets interesting, because the items themselves need localized titles, descriptions, and often prices. Personyze handles this through your feeds, in one of two ways:

  • One feed with language columns. Add extra columns to your existing feed for each language’s values — localized titles and descriptions alongside the defaults.
  • Multiple feeds, one per language or market. Keep a separate feed per locale (an Italian feed, a German feed, and so on), each with its own titles, descriptions, prices, and availability.
Localized recommendations from the right feed
Localize with language columns or per market feeds then use the widgets feed filter to select the right one by language or location

Then, when you configure a recommendation widget or JSON recommendations, you use the feed filter (data source) setting to pick the right feed — by language, by location, or by a combination. An Italian visitor gets Italian titles and euro pricing; a German-market customer gets the German feed. This works for both product recommendations and content recommendations, so you can recommend articles in the reader’s language too. (For how the underlying algorithms choose items, see our guide to recommendation algorithms.)

The traveling customer: when CRM location wins

Location by IP is right most of the time — but not when someone is traveling. Picture a German customer browsing from a hotel in Italy. By IP they look Italian; by account they’re German.

Traveling identified customer CRM location wins
For an identified visitor CRM market Germany overrides IP Italy so the feed language and pricing stay consistent with their home market

For identified visitors, Personyze can let CRM data decide the market. So that German customer keeps the German feed — German language, EUR-DE pricing, German offers — even while abroad, consistent with their billing and catalog. For anonymous visitors, IP location decides. It’s a small rule that prevents a logged-in customer from suddenly seeing the wrong currency or catalog just because they’re on vacation.

Delivered on-site and in email

Every piece of this — localized content, recommendations, promotions — runs across channels. On the site it’s widgets, dynamic content, and popups; in the inbox it’s open-time email that renders the right language and offers at the moment of open; and for headless or app stacks, it’s the JSON/API. One set of language and location rules, applied everywhere.

Setting it up: the short version

  • Tell Personyze how to read language (site selector, cookie, or browser) and location (IP or CRM).
  • Build content variations by language, location, or a combination.
  • Localize your feeds — add language columns, or create one feed per market.
  • In each recommendation widget or JSON call, set the feed filter to select the right feed by language or location.
  • For identified visitors, let CRM location override IP where it matters.

Multilingual personalization: FAQ

How does Personyze detect a visitor’s language?

It can read your site’s own language selector (from the URL or a page element), a cookie value that stores the chosen language, or the browser language as a fallback – and you can combine these signals.

How do multilingual recommendations work in Personyze?

You localize your feed either by adding language columns to a single feed or by creating one feed per language/market. When you set up a recommendation widget or JSON recommendations, the feed filter selects the right feed by language, location, or a combination.

Can I combine language and location targeting?

Yes. Language and location are both signals, so you can target combinations – for example Italian speakers in Switzerland versus Italian speakers in Italy – each with their own content, currency, and offers.

What happens when an identified customer is traveling abroad?

For identified visitors, Personyze can let CRM/account data decide the market, so a German customer browsing from Italy still gets the German feed, language, and pricing. For anonymous visitors, IP location decides.

Does this work for content recommendations, not just products?

Yes. The same feed and targeting approach applies to content recommendations, so you can recommend articles and content in the reader’s language, not only localized products.

Does multilingual personalization work in email?

Yes. Open-time email renders recommendations, content, and offers in the visitor’s language at the moment they open the email, using the same language and location rules as the site.

How do I show different messages to different segments from one template?

Use Personyze’s built-in ${case} / ${when} logic inside a template. You list conditions with ${when Locate(‘value’, attribute)} and a message for each, plus an ${else} default, and Personyze outputs the first match. One template then renders the right headline, copy, or CTA for each segment – by language, country, plan, industry, or any attribute you track.

Get started

Personalize for every language and market from one platform. Book a demo or view plans.

Related reading

Let's talk

Book a demo with a personalization expert

30 minutes with a personalization expert. Bring your stack, your goals, your skepticism. We'll show you what changes when every visit feels like the only one.