Translation and Localization Services
Approved source content re-produced by in-market linguists, priced per target language, with machine-translation post-editing and localization QA as separate bounded tasks.
Trusted by industry leaders
Content Localization Tasks and Pricing
Every price below is per target language. The Language Style Guide and Glossary is paid once per language, before the first batch, which is why the second batch in that language costs less than the first: the foundation is already paid, not discounted.
Content Translation
Approved source copy carried into one target language by a native in-market linguist, with the agreed glossary, termbase, and do-not-translate list applied. Search-demand research per market, creative rewriting of the concept, and site internationalization work are excluded.
Deliverables
- Translated copy per target language, ready to publish
- Glossary and termbase applied across the batch
- Do-not-translate list honoured for product and legal terms
- Source-to-target segment map for later reuse
Volume Tiers
Machine Translation Post-Editing
A full post-edit of machine-translated copy to publishable quality, not a light readability pass. Suitable for high-volume, low-risk content where the source is clean and the target market is already understood.
Deliverables
- Full post-edit to publishable quality, not light MTPE
- Terminology and claims corrected against the glossary
- Edit-distance report showing what the engine got wrong
- Go/no-go note on whether MTPE suits that content type
Volume Tiers
Localization QA
An independent linguistic quality pass over already localized copy, in context, before or after publication. It reviews terminology, register, claims, and formatting against the agreed style guide; it does not rewrite the copy or re-research the market.
Deliverables
- Linguistic QA pass with a severity-classified issue log
- Terminology and style-guide conformance report
- In-context review of live or staged pages
- Sign-off record per language
Volume Tiers
Language Style Guide & Glossary
The one-time foundation for each target language: how your brand speaks in it, which terms are fixed, and what is never translated. Paid once per language, before the first batch, and reused by every later batch in that language.
Deliverables
- Per-language glossary and termbase
- Tone and register rules for that market
- Adaptation rules for claims, units, dates, and names
- Do-not-translate list for product and legal terms
Volume Tiers
Localization Keeps the Meaning; Transcreation Keeps the Effect
A faithful translation can still fail in the target market, because it carries the source market's search logic and persuasion logic with it.
Translated copy inherits source-market search logic
The phrasing that ranks in English is rarely the phrasing buyers type in German or Polish. Word-for-word translation moves the sentence but keeps the English query behind it, so the page reads correctly and is found by nobody. Demand research per market is a separate job, sold on the Cross-Market Setup scope.
Translated copy inherits source-market persuasion logic
Proof points, risk language, and directness travel badly. A claim that reassures a US buyer can read as overselling in Germany and as vague in Poland. An in-market linguist adjusts register, evidence, and emphasis while the offer itself stays intact.
Machine output is a draft, not a deliverable
Modern engines produce fluent sentences and confident mistakes: inverted claims, invented product names, and terminology that drifts across a page set. That is why post-editing is priced here as its own task with its own evidence, rather than hidden inside a lower per-word rate.
Where localization ends and transcreation begins
Localization keeps the meaning: the same message, produced natively for another market. Transcreation keeps the effect and lets the wording change completely, which is a different unit of work with a different price. That work lives on the Transcreation scope of work, not here.
Four Purchasable Tasks, One Localization System
Every task has a defined unit, a fixed price per target language, and a named specialist.
Content translation
Approved source copy produced natively in the target language by a linguist who works in that market, with your glossary and do-not-translate list applied from the first sentence.
- + Native in-market linguist per target language
- + Glossary and termbase applied across the batch
- + Register and claims adapted to the market
- + Segment map kept for reuse in later batches
Machine translation post-editing
Full post-editing of machine output to publishable quality, with an honest note on whether the content type was a good candidate in the first place.
- + Full post-edit, not a light readability pass
- + Claims and terminology corrected against the glossary
- + Edit-distance evidence per batch
- + Go/no-go recommendation per content type
Localization QA
An independent quality pass over copy that is already localized, whether we produced it or you did, reviewed in context rather than in a spreadsheet.
- + Severity-classified issue log per language
- + Style-guide and terminology conformance report
- + In-context review on live or staged pages
- + Sign-off record you can show internally
Language style guide and glossary
The foundation that makes every later batch cheaper and more consistent: how the brand sounds in that language, and which words never move.
- + Glossary and termbase per language
- + Tone and register rules for the market
- + Rules for claims, units, dates, and names
- + Do-not-translate list agreed with legal
Raw Machine Translation vs Post-Editing vs Native Localization
Three ways to get text into another language. They differ in what survives the trip, not just in price.
|
Capability
|
Raw Machine Translation
Engine output, published as-is
|
MT Post-Editing
Engine output, fully post-edited
|
Native Localization
Written natively by an in-market linguist
|
|---|---|---|---|
| Search behaviour in the target language | Carries the source-language phrasing straight through. The page exists in the market but matches almost nothing buyers actually type. | Terminology is corrected, but the underlying phrasing still follows the source text. Works when the market and its queries are already understood. | Wording is chosen for how that market reads and searches, against the agreed glossary and the demand map you supply. |
| Brand voice | Flat and uniform. Every brand sounds like the same engine in every language. | Voice is repaired sentence by sentence and holds inside a batch, but does not originate from a defined register. | Voice is defined once per language in the style guide, then applied consistently across every later batch. |
| Regulated and claims copy | Unsafe. Engines invert negations, soften absolutes, and invent product names without signalling any of it. | Claims are checked against the glossary and corrected, with the changes visible in the edit-distance report. | Claims, units, dates, and legal terms follow explicit adaptation rules agreed before production starts. |
| Terminology consistency | Drifts across a page set. The same product feature gets three different names on three pages. | Held to the glossary inside the batch being post-edited. | Held to the glossary and termbase across every batch in that language, including future ones. |
| QA evidence | None. You find out from customers. | Edit-distance report plus a go/no-go note on the content type. | Severity-classified issue log, conformance report, and a sign-off record per language. |
| Entry cost per target language | Tool subscription only, with the review cost moved to your own team. | MT post-editing from $400 for 10,000 words. | Content translation from $600 for 5,000 words. |
Pick the method per content type, not per company
High-volume support and catalog copy is often a sound post-editing candidate. Money pages, claims copy, and anything a regulator or a buyer reads closely are not. We price both so the choice can be made content type by content type.
Compare Tasks and PricesLanguage Coverage Without a Linguist per Language on Payroll
One system, one glossary per language, one QA standard. Language pairs, target markets, and content types are agreed before the first batch.
Multilingual
One content programme, many target languages
Native in-market linguists per language, working from one approved source and one agreed glossary. Language capability pages list who works in each language and what they produce.
International
One offer landing in several markets
Register, proof points, and claims adapted per country rather than per language alone, because the same language behaves differently across markets.
Global
One standard across every market you enter
Glossaries, style guides, and QA records are held centrally, so market five is produced to the same standard as market one and nothing has to be rebuilt.
From Source Text to a Signed-Off Target Language
Four steps. The foundation is paid once per language; everything after it is priced by volume.
Brief
Agree the source content, the target languages, and the content types. If you also need to know what each market searches for, International Keyword Research is bought on the Cross-Market Setup scope of work, not here.
- Source files, volumes, and content types
- Target languages and target markets
- Existing terminology, claims, and legal constraints
Foundation
Build the style guide, glossary, termbase, and do-not-translate list for each target language. Paid once per language, before the first batch, and reused by every batch after it.
- Glossary and termbase per language
- Tone, register, and claims rules
- Do-not-translate list agreed with your team
Produce
Translate natively, or post-edit machine output where the content type justifies it. The method is chosen per content type and stated in the order, never assumed.
- Native production by an in-market linguist
- Full post-editing where MTPE is the right buy
- Glossary applied across the whole batch
Verify
An independent localization QA pass in context, with a severity-classified issue log and a sign-off record per language.
- In-context review on live or staged pages
- Severity-classified issue log
- Conformance report and sign-off record
Content Marketing Scopes of Work
Content localization works best alongside these complementary scopes.
What Clients Say About Our Localization Work
"We stopped managing five separate vendors and started getting one coherent system. The productized model made it easier to scope work, track progress, and justify spend internally."
"Running marketing in four languages used to mean four agencies with four processes. Onlinuum gave us one operating model across all of them without losing local quality."
Common Questions About Translation and Localization
Straight answers about scope, pricing, and delivery.
How much does content localization cost per word?
We price by volume tier per target language rather than quoting a single per-word rate. Content translation is $600 for 5,000 words, $2,000 for 20,000, and $4,000 for 50,000; 100,000+ is custom-scoped. Machine translation post-editing starts at $400 for 10,000 words, and localization QA at $300 for 10,000 words.
What is the difference between translation and localization?
Translation moves the sentence into another language. Localization produces the same message natively for another market: terminology, register, claims, units, dates, and examples are adapted so the copy behaves the way the source does at home. Everything we sell here is localization; translation is the part of it that happens to be visible.
What is the difference between localization and transcreation?
Localization keeps the meaning and produces it natively in the target language. Transcreation keeps the effect and allows the wording, the image, and sometimes the argument to change completely. That is a different unit of work with a different price, and it is sold on the Transcreation scope of work rather than on this page.
What is localization QA?
An independent linguistic quality pass over copy that is already localized, reviewed in context on live or staged pages rather than in a spreadsheet. You get a severity-classified issue log, a terminology and style-guide conformance report, and a sign-off record per language. It works on copy we produced and on copy you produced elsewhere.
When is machine translation post-editing the right buy?
When the volume is high, the source is clean, the market is already understood, and the content carries low commercial or regulatory risk: support articles, catalog descriptions, internal documentation. It is the wrong buy for money pages, claims copy, and anything a regulator reads. Every order includes a go/no-go note on that judgement.
Why is there no charge for a base language?
Because the language is the unit of work. There is no base-language line item and no percentage surcharge for additional languages: you pay for each target language you actually ship. The second language is genuinely cheaper than the first when its style guide and glossary already exist, since that foundation is paid once per language rather than per batch.
How long does localization take?
Content translation at the S tier (5,000 words, one target language) is delivered in 7-10 business days. M tier runs 2-3 weeks and L tier 3-5 weeks. Post-editing is faster at the same volume, and localization QA is typically 3-5 business days. Multiple languages run in parallel, so total turnaround does not multiply by language count.
Which languages do you cover?
The languages listed on our language capability pages are covered by named in-market linguists. Other languages are available through vetted partners under the same glossary, style guide, and QA standard. Tell us the pairs you need and we confirm coverage and turnaround before the order is placed.
What do you not offer?
We do not sell certified, sworn, or notarized translation, and we do not handle legal documents, patents, medical records, or interpreting. This is marketing and product content production. If you need a certified document, a specialist LSP is the right vendor and we will say so rather than take the order.
Start With One Language and 5,000 Words
One target language, one content batch, one glossary that every later batch reuses. Content translation from $600, with post-editing and localization QA priced separately so you can choose the method per content type.
Delivery: 7-10 business days for 5,000 words per target language. One-time engagement.