Cross-border e-commerce: how translation agencies can capture the market¶

A German knife seller on Amazon.de decided to expand into France and Spain. He ran his listings through Google Translate, launched PPC campaigns - and watched his conversion rate drop to 0.3% instead of the usual 4.2%. What happened? French buyers saw “couteau de cuisine en acier au carbone” instead of the natural “couteau forge en acier carbone” - Google translated literally, ignoring how French people actually search for knives. The Spanish listing was even worse: it kept measurements in “pulgadas” instead of centimeters, because nobody thought to convert units. One translation mistake = thousands of euros in ad spend wasted. And there are millions of sellers just like him. For a translation agency, this isn’t just a niche - it’s a market with obvious pain and real money behind it.

The cross-border e-commerce market: why this matters right now¶

The global cross-border e-commerce market sits at roughly $4.3 trillion in 2025 and grows at a CAGR of 18-23%, according to Precedence Research. By 2032, it’s projected to double. This is one of the fastest-growing segments of the global economy.

What does this mean for a translation agency? Every seller expanding into a new market needs translation of at minimum: product titles, descriptions, search keywords, reviews, FAQ, return policies, email campaigns. Multiply that by 5-10 languages and thousands of SKUs - and you start to see why e-commerce localization has become its own specialization within the translation industry.

But the most interesting part is buyer behavior data.

The CSA Research “Can’t Read, Won’t Buy” study, conducted with 8,709 consumers across 29 countries, found:

76% of online shoppers prefer to buy products with information in their own language. 40% will never buy from websites in other languages.

Think about what that means for a seller targeting Germany, France, or Spain. There’s a direct, measurable link between translation and revenue.

There’s another side to this. According to the ChannelEngine Marketplace Shopping Behavior Report, 58% of buyers purchase from marketplaces outside their own country in 2026. But 46% avoid cross-border marketplaces due to trust issues - returns, shipping, product authenticity. Language remains one of the primary trust barriers, and localized content directly reduces that anxiety.

The translation market overall is growing toward $65 billion, and e-commerce is one of the sectors where demand for human translation holds steady even as AI advances, because the stakes are high: one mistake = lost sales.

What exactly gets translated in e-commerce (and why Google Translate doesn’t cut it)¶

When a store owner hears “translation,” they think “product description, that’s it.” In reality, e-commerce localization covers far more. Here’s the full scope of what needs translation and adaptation:

Product content: - Product titles (SEO-optimized for local search terms) - Descriptions - both full and short versions - Bullet points / key features - Sizes, weights, materials (with unit conversion) - Image alt-text

SEO metadata: - Meta titles and meta descriptions - URL slugs - PPC keywords for ad campaigns - Structured data / schema markup

Legal and compliance content: - Terms of service and privacy policies - Return and exchange policies - Product labeling and markings (CE requirements for the EU) - Disclaimers and warnings

Customer communications: - Email campaigns (order confirmation, shipping updates, marketing) - FAQ and knowledge base - Support response templates - Chatbot scripts

Marketing: - Ad copy (Google Ads, Meta Ads) - Landing pages - Social media posts - Marketing materials for foreign markets

Why doesn’t Google Translate or raw ChatGPT output work here? Three reasons:

1. SEO keywords don’t translate literally. “Running shoes” in the US becomes “Laufschuhe” in Germany, but Germans also search for “Joggingschuhe” and “Sportschuhe.” A literal translation covers only one of three search queries. You need separate keyword research for each market.

2. Units, sizes, currencies. A US clothing size S is 36 or 44 in Europe, depending on the country. If you just translate the text without adapting size charts, you’re guaranteeing returns.

3. Cultural nuances and writing style. German buyers expect technical details and precise specifications. French buyers want a beautiful story and lifestyle framing. Spanish buyers respond to emotional copy with price and discount emphasis. The same product needs different copywriting for each market, not just a word-for-word swap.

As BigCommerce notes in their multilingual e-commerce guide:

Product titles and descriptions need to reflect how customers search in the local language, which often differs significantly from a literal translation.

That’s the difference between translation and localization. And it’s what clients pay a premium for.

Key platforms and what sellers need¶

Not all e-commerce platforms are equal when it comes to translation demand. Here’s where the biggest opportunities sit:

Amazon (European marketplaces)¶

Amazon is the primary driver of e-commerce translation demand. A seller with one account can sell on Amazon.de, .fr, .es, .it, .nl, .se, .pl - and each marketplace requires separate native-language listings. Amazon offers auto-translation, but the quality is weak enough that serious sellers don’t even consider it.

What needs translating for Amazon: title (200 characters), bullet points (5 items), A+ content (enhanced description with images), backend keywords, FAQ, customer Q&A responses.

Margin Business is one example of an agency specializing in Amazon listings. They serve 2,500+ e-commerce businesses and have translated 13,000+ product listings for European marketplaces.

Shopify¶

Shopify is the second-largest platform. It has a built-in Translate & Adapt tool (free but limited to 2 languages for auto-translation) and a plugin ecosystem: Weglot (110 languages), Transcy (147+ languages with AI powered by GPT-4, DeepL, Gemini), LangShop. These plugins solve the technical side (language display, switcher), but the content still needs quality translation.

A typical request from a Shopify seller: “I have 500 SKUs, need translation into 3 languages for European expansion.” That’s 1,500 product descriptions + metadata + legal texts + email templates.

Etsy¶

Etsy is a niche for handmade and vintage goods. The specifics: descriptions tend to be longer and more storytelling-focused than on Amazon. Sellers are often small businesses or solopreneurs with limited budgets, but they suffer the most from bad translation because their product sells on the strength of the text.

WooCommerce, Magento, BigCommerce¶

Custom online stores. Translation here covers the entire site: from catalog to checkout. These are the most complex projects by volume, but also the most profitable.

Platform	Typical project volume	SEO complexity	Client budget
Amazon EU	50-500 listings x 5-7 languages	high (backend keywords)	medium-high
Shopify	100-2,000 SKUs x 2-5 languages	medium	medium
Etsy	20-200 listings x 1-3 languages	low	low-medium
WooCommerce/Magento	full site x 2-10 languages	very high	high

How much can a translation agency earn from this¶

Here’s where it gets interesting. E-commerce localization costs more than standard translation because it includes SEO research, cultural adaptation, and often ongoing support.

According to Smartling and Alconost, typical rates in 2026:

Service type	Rate	Notes
Standard translation	$0.08-0.25/word	no SEO, no adaptation
E-commerce localization	$0.15-0.35/word	with SEO metadata and adaptation
Full site localization	$5,000-25,000+/language	for content-rich e-commerce
MTPE (AI post-editing)	65-75% of human rate	growing demand
SEO keyword research	$200-800/language	separate service

Localization costs 30-50% more than standard translation. And it makes sense - you’re not just translating words, you’re adapting content for a specific market with local search terms factored in.

Let’s run a concrete example. A mid-size Shopify store with 300 SKUs wants to expand into 3 new markets (Germany, France, Spain). Each SKU is roughly 150 words (title + description + bullet points + metadata). Plus general site content (FAQ, policies, emails) - another 10,000 words.

• Product content: 300 SKUs x 150 words x 3 languages = 135,000 words
• General content: 10,000 words x 3 languages = 30,000 words
• Total: 165,000 words
• At $0.20/word = $33,000
• Plus SEO keyword research: 3 languages x $500 = $1,500
• Grand total: ~$34,500 for a single project

And that’s one client. After that comes ongoing maintenance: new products, seasonal descriptions, A/B testing copy, legal text updates. That’s a retainer worth $1,000-5,000 per month.

Compare that with a typical document translation order (passport, diploma) at $30-50 per page. E-commerce is an entirely different scale.

How to build an e-commerce localization service from scratch¶

If you already run a translation agency and want to add e-commerce as a service line, here’s a step-by-step approach.

Step 1: Define your language pairs and markets¶

You don’t need to offer 20 languages from day one. Start with the 2-3 most profitable pairs for e-commerce:

• English to German - the largest e-commerce market in Europe ($113 billion in 2025)
• English to French - second largest ($79 billion)
• English to Spanish - third in Europe plus all of Latin America

If you have translators for these pairs, you can start right away.

Step 2: Build an e-commerce-specific portfolio¶

A generic translation portfolio won’t convince an e-commerce client. You need:

• Sample translated Amazon listings (you can create demo versions)
• Before/after with metrics (conversion rates before and after localization)
• Case studies with specific numbers (“translated 200 SKUs into German, conversion rose from 1.8% to 3.1%”)
• Knowledge of platform requirements (Amazon character limits, Shopify theme structure)

Step 3: Add SEO keyword research as a service¶

This is what separates an e-commerce agency from a generic translation bureau. For each market you need to:

• Find local search terms (Ahrefs, SEMrush, Google Keyword Planner)
• Check search volume for each keyword
• Adapt titles and descriptions to target those keywords

You can offer this as an add-on at $200-800 per language or bundle it into your localization price. If you want to go deeper, we’ve written about multilingual SEO as a system.

Step 4: Choose a pricing model¶

Three main models:

• Per-word ($0.15-0.35) - most common, easy for clients to understand
• Per-SKU ($5-25 per SKU depending on complexity) - works well for Amazon projects
• Flat project rate ($3,000-15,000 per project) - for large launches

For retainer clients, a monthly subscription with a fixed word volume and ongoing support works best.

Step 5: Find your first clients¶

Where to find e-commerce sellers who need translation:

• Amazon Seller Forums - sellers regularly ask about translation services
• Shopify Partner Program - become a partner and get referrals
• LinkedIn - search for “Amazon FBA seller,” “e-commerce manager,” “international expansion manager”
• E-commerce conferences - Amazon Accelerate, Shopify Unite, IRCE
• Upwork / Fiverr - as a starting channel to build reputation (but watch your pricing)

Tools and workflow¶

E-commerce translation differs from document translation because you’re dealing with massive volumes of similar content. You can’t translate 500 listings manually in Word - you need the right tools.

CAT tools¶

Standard CAT tools (Trados, MemoQ, Smartcat) work for e-commerce, but with some nuances:

• Translation Memory (TM) - critically important. If every knife in the catalog has a description with “stainless steel, ergonomic handle,” TM stores that fragment and translates it automatically for all subsequent SKUs. This reduces work volume by 30-60%
• Termbase - a unified glossary for the brand. “Premium quality” = “erstklassige Qualitat” (not “hochste Qualitat,” not “Premium-Qualitat”). Consistency across the entire catalog matters

Specialized e-commerce platforms¶

• Phrase (formerly Memsource) - integrates with Shopify, WooCommerce, Contentful. Automatic content import from CMS, translation, export back
• Smartling - serious TMS with native e-commerce integrations
• Lokalise - popular for SaaS and e-commerce, API-first approach
• Weglot - frontend solution for Shopify/WordPress, translates on the fly

AI in the workflow¶

A realistic 2026 workflow looks like this:

1. AI produces the first draft (DeepL, GPT-4, Claude) - saves 40-60% of time
2. Translator post-edits - fixes nuances, adapts cultural elements, checks SEO keywords
3. QA check - automated (tags, numbers, consistency) + manual (style, tone of voice)
4. SEO specialist - verifies keywords, metadata, alt-text

This is the MTPE (Machine Translation Post-Editing) model, and it’s already becoming standard in e-commerce translation. AI is changing pricing, but it doesn’t replace humans - because mistakes in e-commerce translation directly impact sales.

5 mistakes that cost e-commerce clients money¶

These mistakes repeat so frequently that every agency should know them - both as a sales argument and as a quality control checklist.

1. Skipping local keyword research. A seller translates “baby stroller” as “Kinderwagen” (correct), but doesn’t know that Germans also search for “Buggy,” “Sportwagen,” and “Kombikinderwagen.” Without these keywords, the listing is invisible to 60% of potential buyers.

2. Unconverted units of measurement. Inches instead of centimeters, pounds instead of kilograms, Fahrenheit instead of Celsius. Buyers won’t convert on their own - they’ll just close the page.

3. Using the same Spanish for every Spanish-speaking market. Spanish for Spain, Mexico, and Argentina is three different variants of the language. “Computadora” (Mexico) vs “ordenador” (Spain). Amazon.es and Amazon.com.mx need different texts.

4. Machine-translating legal texts. Return policies, warranties, privacy policies - these are legal documents. A translation error in German return conditions can violate Widerrufsrecht (the right of withdrawal) and lead to fines.

5. Forgetting alt-text and meta-descriptions. The seller translates the product description but leaves the meta-description in English. Google shows English snippets to German buyers - CTR drops 40-60%.

As one seller shared on the Amazon Seller Forum:

I spent $3K on PPC in Germany last month with auto-translated listings. My ACOS was 85%. Then I hired a native German speaker to redo all the listings with proper keywords. Next month ACOS dropped to 32%. The translation paid for itself in 3 days.

This experience is typical. Investment in quality translation pays back through higher conversion and lower ad spend.

Trends shaping 2026-2028¶

A few trends worth factoring in as you build an e-commerce translation service:

AI + Human workflow becomes standard. Pure human translation for e-commerce catalogs with thousands of SKUs is economically impractical. Pure machine translation gives low quality. The AI + human hybrid (MTPE) is the sweet spot: 40-60% faster than pure human, 30-50% cheaper, quality at 95%+ of human level.

Continuous localization. E-commerce isn’t a one-time project. New products, seasonal promotions, A/B testing copy - content changes weekly. Agencies that offer continuous localization with integration into the client’s CMS land long-term retainers instead of one-off orders.

Per-market personalization. The same product can have a different emphasis in different countries. A memory foam pillow in Germany sells through technical specs (“viskoelastischer Schaum, orthopadisch getestet”), in France through lifestyle (“un sommeil reparateur, comme dans un hotel 5 etoiles”), in Spain through price and discounts (“50% descuento, envio gratis”). This isn’t translation anymore - it’s transcreation, and it commands higher rates.

Video content is growing. Amazon A+ Content, Shopify product videos, TikTok Shop - all of this needs subtitles, scripts, voiceover. An agency that combines text localization with video translation covers more of the client’s needs.

FAQ¶

How much does e-commerce catalog translation cost?¶

It depends on volume and language pair. The typical rate for e-commerce localization is $0.15-0.35 per word. For a catalog with 200 SKUs (roughly 30,000 words), that’s $4,500-10,500 per language. Large catalogs (1,000+ SKUs) usually get a 15-25% discount thanks to Translation Memory.

Is it profitable to specialize only in e-commerce?¶

Yes, if you have at least 3-5 stable clients. An average e-commerce client brings $10,000-50,000 per year (initial translation + monthly updates). Five such clients means $50,000-250,000 in annual revenue from a single service line. Plus e-commerce clients grow - they expand their catalogs, enter new markets, need more languages.

Do you need specialized translators for e-commerce?¶

Ideally, you want translators with copywriting or marketing experience. E-commerce text isn’t legal translation - what matters here is selling with words, understanding SEO principles, and knowing how marketplaces work. A standard translator will produce grammatically correct output but miss SEO keywords and won’t adapt the style for the target market.

How do you compete with cheap Fiverr freelancers?¶

Price isn’t the only argument. A freelancer will translate the text, but won’t do keyword research, doesn’t know Amazon’s listing requirements, can’t maintain consistency across 500 SKUs, and doesn’t have TM or termbase infrastructure. Offer the full package: translation + SEO + QA + consistency + ongoing support. Clients pay 2-3x more for that.

Which language pairs are most profitable for e-commerce?¶

English to German (largest EU market), English to French (France + Canada + Africa), English to Spanish (Spain + all of Latin America), English to Japanese (third-largest e-commerce market globally). Asian languages (Japanese, Korean, Mandarin) deliver the highest margins due to complexity.

Will AI replace translators in e-commerce?¶

It won’t replace them, but it will change the role. AI already produces first drafts, but without human post-editing, the quality isn’t sufficient for serious e-commerce. AI mistakes in product descriptions directly affect conversion rates and returns. The translator’s role is shifting from “translate from scratch” to “edit the AI draft + do SEO adaptation.” It’s faster, but it requires a different skill set.

How do you start with a small 2-3 person agency?¶

Start with one language pair and one platform (for example, English to German for Amazon). Create 2-3 demo listings for your portfolio, get active on Amazon Seller Forums and LinkedIn, offer a free audit of a potential client’s first 5 listings. Once you land the first paying client, scale from there.