How to translate your Shopify store with your brand voice and SEO strategy

There is a sentence we hear in almost every first conversation with a brand selling internationally on Shopify.

"We are already doing well in that market. Our SEO there is good."

It is usually true. It is also usually the reason the brand has stopped looking. Doing well in a market where your store is not properly translated means you are ranking despite your content, not because of it. The ceiling you have hit is not the market's. It is your own.

TL;DR

• Shopify's own data puts translation ahead of every other localization lever. Localized content is worth a 13% GMV increase, against 7% for local currency and 6% for local payment methods.

• Across Newtone clients, conversion uplift runs 11% to 24% after proper translation, and first-year organic traffic gains range from 15% to 120%.

• The range is that wide for a reason. Your upside depends on whether you already have presence in the market, and on how good your master site SEO was to begin with.

• There are four ways to translate a Shopify store and they are not interchangeable. Two of them actively damage your SEO.

• Overlay translation apps are the biggest trap. Fast to launch, but search engines do not index the translated content, and uninstalling the app takes your translations with it.

• The only approach that compounds writes real translated content into Shopify's own Translatable API, with brand voice and target-language keyword research built in.

"We are already good at SEO in this country"

Here is what that sentence usually means in practice.

The brand has an English storefront. It gets organic traffic from Germany, or Spain, or Japan. Someone checked analytics, saw the sessions, and concluded the market was working.

What that traffic actually represents is the subset of local shoppers who search in English, or who search for your brand name directly. It is demand you already had. It tells you nothing about the much larger volume of shoppers searching in their own language for the category, the product type, the use case or the problem you solve, and never encountering you at all.

You are not competing for those queries. You are not eligible for them. An untranslated page cannot rank for a query in a language it does not contain.

What translation is actually worth

Shopify has published its own figures on this, drawn from merchants using Shopify Markets. The useful part is that it measures three localization levers separately, which lets you see their relative weight.

Source: Shopify, Commerce without borders. Note the precise wording: Shopify reports a 13% lift in GMV from buyers who see localized content. It is frequently quoted elsewhere as a conversion rate figure, which is not quite what the source says.

Our own client data points the same way, measured as conversion rate on translated storefronts against the untranslated baseline.

Why the traffic range is so wide

A 15% to 120% range is not a hedge. It reflects two variables that genuinely decide the outcome, and you can assess both before you start.

The second variable is the one brands consistently underestimate, and it has an uncomfortable implication.

Translation multiplies your master site. It does not fix it.

If your English product pages carry six attributes and eighty words of category copy, translating them gives you the same thin content in nine more languages. You have multiplied the problem, paid for the privilege, and gained very little. We measured exactly this kind of catalogue thinness in Why 60% of your product catalog is invisible to AI search.

So the honest sequence is: fix the master site first, then translate. Depth on product pages, real editorial on category pages, clean internal linking, and content that answers the questions shoppers actually ask. That is the groundwork we set out in How to optimize your ecommerce store for AI search, and it is what makes the translated versions worth having. The same freshness problem applies to every language you publish, as we covered in Your category pages have an expiry date.

The four ways to translate a Shopify store

Once you decide to translate, there are four realistic routes. They differ enormously, and the differences are mostly invisible until you look at what search engines actually receive.

1. A human translation agency

The traditional route. An agency, professional linguists, a brief, a review cycle.

What works: quality is genuinely good. A skilled human translator will produce natural, accurate copy.

What does not: it is slow and expensive, and it scales linearly with your catalogue. Worse for your purposes here, agencies translate. They do not do SEO in the target language. A faithful translation of your English page title is not the phrase German shoppers actually type. You end up with beautiful copy that ranks for nothing, because nobody researched the local keyword set before writing.

2. Overlay translation apps

Apps such as Weglot sit in front of your storefront and translate it on the fly. Install, pick your languages, and the store appears translated within hours.

What works: speed. This is by far the fastest way to show a translated storefront, and for a quick market test it has real appeal.

What does not: your store is not actually translated. The content is rendered in the visitor's language at load time, while the underlying store remains in the source language. For SEO that distinction is everything. Search engines are not reliably indexing translated pages they were never served as real, stable, localized content, which means you do not rank in the target language. You get a storefront that looks international to visitors and looks monolingual to Google.

Quality is the second issue. Overlay translation is generic machine output with no tone of voice, no glossary, and no awareness of your brand. It reads like software, because it is.

The third issue is the one that catches brands out eighteen months later.

3. Shopify Translate & Adapt

Shopify's own free app. It writes translations into your store through Shopify's Translatable API, so the content genuinely exists in your storefront.

What works: the architecture is right. Because translations are stored natively, you get real localized URLs that search engines can crawl and index. This is a structurally better foundation than any overlay, and it costs nothing.

What does not: quality. The output is machine translation without brand voice, without a glossary, and without any SEO work in the target language. It is also awkward to operate at catalogue scale, and it offers no meaningful quality control. You end up with correctly architected pages carrying content no brand team is happy to publish.

4. Bespoke AI trained on your brand, writing through the Translatable API

The fourth option takes Translate & Adapt's architecture and fixes the content problem.

A model trained on your brand guidelines, tone of voice and glossary, with target-language keyword research built into generation rather than bolted on afterwards, and a quality check before anything syncs. Content is written back through the same Shopify Translatable API that Translate & Adapt uses, so you inherit the same native, indexable, Shopify-managed infrastructure.

That combination is the only one that gives you all three things at once: content that sounds like your brand, content optimised for how people actually search in that language, and content that lives in your store rather than in a vendor's overlay. It is the difference between a tool and infrastructure, which we argued in Why retail content teams are moving from AI tools to AI infrastructure.

Newtone is one of these solutions, trusted by Victoria Beckham, Isabel Marant, Adanola and Meshki among others.

The four options compared

What search engines actually see

The single most expensive misunderstanding in Shopify translation is assuming that a storefront which looks translated to a human is translated as far as Google is concerned.

This matters more every quarter, because the same content now has to satisfy AI assistants as well as classic search. Models cite content they can retrieve, in the language of the question. A storefront that only exists in English is invisible to a shopper asking an AI assistant a question in Italian, as we set out in Generative engine optimization, a new visibility frontier for enterprise retail and in Google just settled the GEO debate.

Translation is not the same as localisation

A literal translation preserves your sentences. It does not preserve your commercial performance.

The phrase a German shopper types is frequently not the German translation of the phrase a British shopper types. Category names differ. Sizing language differs. The occasion vocabulary differs. Seasonal demand differs, sometimes inverting entirely between hemispheres. A translator working from your English page without a local keyword set will produce accurate copy aimed at queries nobody makes.

This is also where brand voice is usually lost. Tone is the first casualty of machine translation and the hardest thing for an agency to hold consistent across nine languages and four freelancers. We looked at the commercial cost of getting this wrong in How brand-quality translation drives international revenue, and at what voice consistency actually requires in The brand is no longer in the room.

What to require before you commit

1. Confirm the content lands in your store. Ask directly whether translations are written through Shopify's Translatable API or rendered by the vendor. If it is the latter, you are renting, not building.

2. Ask what happens if you leave. If the answer is that your translated storefront disappears, price that into the subscription.

3. Require target-language keyword research. Not translated keywords. Researched ones, from the local market.

4. Require a glossary and tone of voice. Product names, materials, claims and brand terms should be locked before a word is generated.

5. Audit the master site first. Translating thin content multiplies a weakness. Fix depth and structure before you scale it across markets.

6. Measure per market, not in aggregate. Set a baseline for organic traffic and conversion rate per market before launch, or you will not be able to prove what the translation did.

Frequently asked questions

Does translating a Shopify store actually improve SEO?

Yes, provided the translation is stored natively rather than overlaid. Native translations create indexable localized URLs with content in the target language, which is what makes you eligible to rank for local queries. Overlay translation, rendered at load time, does not reliably produce that, so rankings in the target language generally do not follow.

Is Shopify Translate & Adapt good enough on its own?

Architecturally yes, editorially no. It uses the right mechanism, writing into Shopify's Translatable API, so the SEO foundation is sound and it is free. The limitation is content quality: machine translation with no brand voice, no glossary and no target-language keyword optimisation, which makes it hard to publish at the standard most brands require.

Will I lose my translations if I uninstall a translation app?

With an overlay app, yes. The translated content lives in the app rather than in your Shopify store, so removing the app removes the translated storefront. With any solution writing through Shopify's Translatable API, including Translate & Adapt, the content stays in your store.

Can AI translation maintain brand voice across languages?

Generic machine translation cannot, because it has no model of your brand. A system trained on your tone of voice, style guidelines and approved glossary can, because voice is applied at generation rather than corrected afterwards. The practical test is whether the vendor can show you consistent output across several languages from the same guidelines.

Should I translate my whole catalogue or start with one market?

Start with one or two markets and a defined product scope, with a clean per-market baseline for organic traffic and conversion rate. The upside varies enough by market and category that a scoped test tells you far more than a model will.

How long before translated content ranks?

Expect a discovery and indexing period first, then gradual ranking gains as the localized pages accumulate signals. First-year outcomes across Newtone clients range from 15% to 120% organic traffic growth, driven mostly by existing market presence and the quality of the master site being translated.

Where Newtone fits

Newtone is content engineering infrastructure for retail brands. For Shopify specifically, we generate translated and localized storefront content trained on your brand guidelines, tone of voice and glossary, with target-language SEO and GEO optimisation built into generation, quality checks before anything publishes, and sync back through Shopify's native Translatable API.

That means you get agency-grade voice, keyword coverage no agency provides, and the same native Shopify architecture Translate & Adapt uses, at a scale and cadence neither a human team nor a free app can reach. Brands including Victoria Beckham, Isabel Marant, Adanola and Meshki use it to run their international storefronts. The wider approach is described in The coherent catalog, and the multi-market version of the same problem in How FMCG brands grow retailer sales with AI content infrastructure.