Shopify Translate & Adapt: Why Your Product Data Comes First

A product description that says "High quality product" in English becomes "Hochwertiges Produkt" in German. The translation is technically correct. The description is commercially useless in both languages.

This is the problem most merchants discover after they have already set up Shopify Translate & Adapt, added a second language, and published their multilingual store. The translations are accurate. The content is still flat. The instinct is to look for a better translation tool. The actual problem is upstream.

Shopify Translate & Adapt translates strings. It does not supply what those strings lack. If you want a multilingual store that converts, the work happens before you open the translation app.

What Shopify Translate & Adapt actually does#

Shopify Translate & Adapt is a free Shopify app that connects to Shopify Markets to deliver localised content to shoppers in different languages. It translates the following content fields on each product:

• Product title
• Body HTML (the product description)
• SEO title and meta description
• Product option names and values (Size, Colour, Material)
• Custom metafield values, where the metafield is configured as translatable in Shopify

What it does not translate: image alt text (that is a separate process), prices (Markets handles regional pricing), inventory data, and SKU codes.

According to Shopify's Translate & Adapt documentation, the app supports over 130 languages. Every one of those 130 language options will translate whatever string you give it with the same accuracy. A specific, benefit-rich English description translates into a specific, benefit-rich German description. A vague three-word description translates into a vague three-word German description.

The field coverage is comprehensive. The translation fidelity is reliable. The quality ceiling is your source content. For reference on how Google handles multilingual content at the indexing level, Google's multilingual sites documentation explains why content quality in each language matters independently for search ranking.

Why description quality sets the translation ceiling#

Consider a fashion merchant importing a 300-product spring catalogue from a European wholesale supplier. The supplier's descriptions are written for trade buyers, not end consumers. A typical entry looks like this: "100% cotton jersey. MOQ 12. Available Q2. Colourways: sand, dusk, slate."

Shopify Translate & Adapt will render that description in French, German, and Spanish with complete accuracy. "100% coton jersey. MOQ 12. Disponible T2. Coloris : sable, crépuscule, ardoise." The translation is correct. A French-speaking shopper reads it and has no more reason to buy than an English-speaking shopper.

The best possible translation of thin content is thin translated content.

“The merchants who get the best multilingual stores spend their effort upstream: enriching product data before translation, not after.

Rich source content has specific properties that translate well. A description like "Lightweight trail running shoe with a Vibram outsole, 4mm drop, and waterproof upper. Designed for technical terrain where weight and grip both matter." gives a translator (human or AI) clear parsing targets. Each attribute is named. Each benefit is stated. The translation preserves all of it.

Compare that to "Comfortable, high-quality trail shoe." The translator produces an accurate version of a sentence that was already insufficient. Translating it does not fix the insufficiency.

This is not a problem with Translate & Adapt. It is a data quality problem. And data quality decisions happen at import time, not translation time. For a deeper look at the five product data fields that affect performance across all channels, including international ones, see product data quality across all five key fields.

What Translate & Adapt actually translates, field by field#

Understanding which fields get translated helps you decide where to invest your data quality effort before the translation step runs.

Product title: translated directly. A structured title like "Men's Trail Running Shoe: Lightweight, Waterproof, Size 8 to 14" gives the translation engine named attributes to work with. A title like "Running Shoe 2024 V2 Premium" gives it four tokens with no semantic structure. The Title Optimizer in Importier generates structured, marketplace-ready titles during the import wizard. Google Merchant Centre titles cap at 150 characters, eBay at 80, and Amazon at 200.

Product description: the body HTML is translated in full. This is where the quality gap is most visible. A 300-word consumer-ready English description produces a 300-word consumer-ready translated description. A 12-word trade entry produces a 12-word translated trade entry.

SEO title and meta description: both translated. A meta description like "Shop our trail running shoe range. Free delivery over $75." translates accurately but remains as generic in German as it is in English. A meta description with a specific feature claim ("Vibram outsole, 4mm drop, waterproof upper, available in sizes 8 to 14") translates with the same specificity.

Product option names and values: "Size", "Colour", "Material" and their values are translated. This generally works well because option values tend to be short and structured.

Custom metafield values: translatable where the metafield is configured as translatable in Shopify. Free-text metafield values like "100% recycled polyester, OEKO-TEX certified" are translated in full. Quality again depends on what the source string contains.

What does not get translated: image alt text requires separate handling, prices are managed by Shopify Markets per region, and SKU codes are not content fields.

The import-first, translate-second workflow#

The sequencing matters for two reasons. First, translation quality depends on source content quality, and source content decisions are made at import time. Second, changing the English source content after translations exist re-queues the translation and may overwrite manual translation edits you have made in specific languages.

Getting the content right in English before the translation layer runs is not just a quality decision. It is also a workflow efficiency decision.

1. 01
   Import supplier catalogue through the 14-step wizard and generate AI-enriched descriptions before publishing to Shopify
2. 02
   Run Title Optimizer to structure product titles with named attributes and correct format for each sales channel
3. 03
   Populate category metafields using an industry pack so taxonomy values are clean and well-formed
4. 04
   Review the complete product set in English (descriptions, titles, metafields, SEO fields) and confirm everything is consumer-ready before publishing
5. 05
   Open Shopify Translate & Adapt and run translation on your enriched, complete source content

Importier's AI Description Generator supports 18+ models across four pricing tiers, 7 description styles, and 156 expert personas across 43 industries. Running a 300-product wholesale catalogue through the import wizard and generating descriptions takes under an hour. Translating those descriptions with Translate & Adapt takes a similar amount of time. Fixing flat translations by rewriting source content after the fact and re-running translation takes much longer.

If you have already imported products with thin descriptions and translated them, you are not permanently stuck. Importing or regenerating descriptions for those products, then re-running Translate & Adapt, will update the translations. The generating AI-enriched descriptions before translation article covers the description generation process in detail.

The trade-off is that any manual translation edits you have made in specific languages will need to be reviewed after the source English content changes. For stores with a large volume of manual translation work, this is a strong argument for getting the English content right before starting translation at all.

Metafields: the often-forgotten translation layer#

Most merchants focus on translating product descriptions and titles. Metafields are the content layer that frequently gets overlooked, and it affects both translation quality and on-site filtering in non-English markets.

Shopify Translate & Adapt can translate custom metafield values when those metafields are configured as translatable in Shopify. The configuration step happens in your Shopify admin under content and metafields. Once a metafield is marked translatable, Translate & Adapt will include its value in the translation queue.

Category metafields (the structured taxonomy attributes that come from Shopify's Standard Product Taxonomy) behave differently. These are not free-text values. "Trail Running" as a product type value is a taxonomy term that Shopify manages. The translation behaviour for taxonomy values depends on whether Shopify has pre-translated that term for the target locale, which varies by category and language.

Custom free-text metafields, by contrast, translate exactly as you would expect. A "Materials" metafield containing "100% recycled polyester, OEKO-TEX certified" will be translated in full. A "Care Instructions" metafield containing "Machine wash cold, hang to dry" will produce accurate translations in any supported language.

The practical implication: metafields populated during import by Importier's specification auto-fill and industry pack tools contain complete, well-formed strings. "Magnification: 8x | Objective Lens Diameter: 42mm | Field of View: 110m at 1000m" is a clean string that translates without ambiguity. An empty metafield or a partial string ("8x42") leaves the translator with nothing to work with.

For international merchants who also need to consider compliance data (HS codes, IOSS VAT numbers, and EU de minimis requirements), see importing products for international compliance, which covers those requirements separately from translation.

What to do if you have already published thin descriptions#

If your store is already live with translated content and the English source descriptions are thin, the fix is the same as the prevention: enrich the English source content first.

Use Importier's Store Scanner to identify which products have short or missing descriptions across your catalogue. Generate new descriptions for those products using the AI Description Generator. Once the English content is updated in Shopify, return to Translate & Adapt and re-publish the translations for those products.

Any manual translation edits made in specific languages will need to be reviewed after the source English changes. For stores where manual translation investment is high, prioritise the products with the most traffic first rather than running a full catalogue re-translation in one pass.

The international tax compliance fields article is worth reading alongside this one if your multi-language store also sells to EU customers. The compliance data layer and the translation layer both need attention for international stores, and they are separate problems solved at different points in the workflow.

Three things to take from this#

• Shopify Translate & Adapt translates what you give it, accurately, in over 130 languages. The translation tool is not the source of flat multilingual content.
• Translation quality is bounded by source content quality. A specific, benefit-rich English description produces a specific, benefit-rich translation. A thin trade description produces a thin translation in every language.
• The highest-leverage point for multilingual content quality is the product import step, not the translation step. Enrich product data in English before running translation, and the translation layer works as intended.

Try Importier free at importier.app.