Translation: What it is, how it works, types, & examples

Translation is the process of converting written or spoken content from one language into another while preserving meaning, tone, and intent.

Good translation isn't word-for-word substitution, since holding onto the original meaning almost always means changing the words that carry it.

A strong translation keeps the message, the emotional register, and the purpose of the original text, even when the sentence structure and word choice change completely. What it preserves matters more than what it changes.

For global businesses, translation supports product documentation, marketing content, user interfaces, customer support, legal material, and more. The right approach depends on what you’re translating, who will see it, and what happens if the translation gets something wrong.

This article covers what translation is, how it differs from localization and transcreation, how the translation process works, and how to match the right method to your content before you scope a program.

Translation vs localization vs transcreation

Translation, localization, and transcreation sit on a spectrum of adaptation.

Translation converts meaning from one language to another while preserving the original message and intent.

Localization adapts the broader experience for a specific market, including language, formatting, cultural references, and design or functionality.

Transcreation takes the greatest creative freedom, rewriting content for emotional and cultural effect while preserving original intent.

A software company translating its terms of service from English into German needs precision above all else. The legal meaning must remain intact, so the translator focuses on accuracy and terminology.

Localization goes further. A company localizing an app for Japan may adapt date formats, payment methods, cultural references, and interface layouts in addition to translating the text.

Transcreation is used when the words themselves are part of the creative idea. A tagline or advertising campaign may need to be rewritten rather than translated literally so that it creates the same emotional response in another market.

The decision rule is simple: the closer content is to pure information, the less adaptation it generally needs. The closer it is to persuasion, the more creative adaptation likely matters.

What is the translation process?

A typical translation workflow has four stages, whether the content is a support article, product interface, legal document, or marketing campaign.

1. Prepare the source content

Before translation begins, the source content should be checked for consistency, terminology, formatting, and context. Existing glossaries, style guides, and translation memory can help translators, AI translation, or machine translation work from established terminology rather than starting from zero.

Good source content makes quality translation easier. Ambiguous language, inconsistent terminology, and unclear context create problems downstream.

2. Translate

The content is translated using a human translator, machine translation, AI translation, or a combination of methods.

The appropriate method depends on factors such as content type, volume, visibility, risk, turnaround time, and the level of nuance required.

3. Review and edit

The translation is reviewed for accuracy, fluency, terminology, tone, and cultural appropriateness.

Human review can catch problems that automated systems may miss, including phrasing that is technically correct but unnatural, terminology that doesn't fit the brand, and meaning that has been lost during translation.

4. Publish and validate in context

The translation is published in its final environment and checked in context.

A sentence that looks correct in a translation editor can behave differently on a website, in an app, or in a printed document. Text can overflow a button, change the visual balance of a page, or lose meaning when surrounding content changes.

Contextual validation helps catch those issues before they reach the audience.

What does a translator do?

A translator does much more than replace words from one language with words from another.

The job is to make the message feel natural to the target audience while preserving what the original author intended.

A translator’s role can involve:

• Adapting idioms, cultural references, and informal language that do not transfer literally.
• Choosing terminology that fits the subject matter and audience.
• Adjusting phrasing so the translation reads naturally rather than like a translation.
• Preserving tone, whether the source is technical, conversational, formal, or persuasive.
• Applying subject-matter expertise in areas such as legal, technical, medical, or financial content.
• Reviewing translations in context using computer-assisted translation (CAT) tools and other technology.

Professional translators may also work from glossaries, style guides, translation memories, and other linguistic assets to maintain consistency across large volumes of content.

What makes a translation good?

A translation can be grammatically correct and still be a poor translation.

Quality depends on more than whether individual words are accurate. A strong translation balances several factors:

• Accuracy: The translation communicates the meaning of the source without omissions, additions, or distortions.
• Fluency: The content reads naturally in the target language.
• Cultural relevance: References, idioms, and conventions make sense to the intended audience.
• Terminology consistency: Important product, technical, and brand terms are used consistently.
• Brand voice: The translation sounds like the company, not like a generic translation.

For large multilingual programs, teams often use linguistic quality assurance (LQA) to evaluate translations against defined quality criteria. Glossaries and style guides provide translators and AI systems with the terminology and rules needed to maintain consistency across languages.

Types of translation services

Translation needs vary by industry, content type, and audience. Common types include:

• Literary translation: Books, poetry, and other creative works where voice, style, and meaning all matter.
• Technical translation: Manuals, specifications, engineering documents, and other technical content where precision is critical.
• Legal translation: Contracts, filings, regulatory documents, and other legal material where terminology and accuracy can carry significant consequences.
• Medical translation: Clinical documents, patient materials, pharmaceutical content, and other healthcare information where accuracy can affect patient safety.
• Financial translation: Reports, disclosures, financial communications, and compliance material for regulated markets.
• Certified translation: Official document translation accompanied by a certification or statement of accuracy when required by a government agency, institution, or other authority.
• Business translation: Internal communications, corporate documents, and day-to-day business content.
• Multimedia translation: Subtitles, captions, voiceovers, and other audiovisual content.
• Website and software localization: Digital content and products adapted for specific languages, regions, and user expectations.

Each category calls for a different balance of speed, cost, subject-matter expertise, and precision.

What is the best translation method?

There is no single translation method that works for every type of content.

A product update, legal filing, support article, and global advertising campaign may all need to be translated, but they do not carry the same level of visibility or risk.

The most useful question is not simply “Should we use human or AI translation?” It is “What level of translation is appropriate for this content?”

Human translation

Professional linguists are particularly valuable when nuance, subject-matter expertise, or creative judgment matters.

Human translation is well suited to high-visibility or high-risk content such as legal material, regulated content, brand campaigns, and important customer-facing communications.

Machine and AI translation

Machine translation (MT) uses automated systems to convert content between languages at high speed and scale. Modern AI translation can combine neural machine translation, large language models, and other AI technologies to improve output across different content types.

Automated translation is particularly useful for high-volume, lower-risk content where the primary goal is fast comprehension and broad coverage.

AI translation with human review

A hybrid approach combines automated translation with human review. The AI system produces the initial translation, and a linguist reviews and edits the output before publication.

This approach can provide more speed and scale than fully human translation while adding a layer of human judgment where accuracy, tone, and natural language matter.

Smartling's AI Human Translation (AIHT) is a pre-configured workflow that integrates AI translation with human expertise into a single workflow, layering AI-driven efficiency into every step.

How should you choose a translation method?

Match the method to the content rather than applying the same workflow to everything.

These are starting points, not fixed rules. A support article containing safety instructions may deserve more review than a routine marketing page, while a high-volume marketing archive may be appropriate for automated translation.

The right translation program can therefore use different methods for different content tiers instead of forcing every piece of content through the same workflow.

Translation, localization, and internationalization

Translation is one part of preparing content for global audiences.

Internationalization prepares a product or system to support multiple languages and regions. It can include technical decisions such as character encoding, text expansion, date and number formats, and interface design.

Localization then adapts that product or experience for a particular market.

A simple way to think about the relationship is:

Internationalization prepares the product. Translation changes the language. Localization adapts the experience. Transcreation adapts the creative idea.

For software and digital products, considering internationalization early can make later translation and localization faster and less expensive.

Match the method to the content

The goal of translation is not to change as few words as possible. It is to preserve what matters.

For some content, that means preserving precise legal meaning. For other content, it means making an interface feel natural, keeping a brand voice consistent, or recreating the emotional effect of a campaign.

That is why translation programs work best when they assign the right method to each content type.

Smartling's LanguageAI Platform helps teams manage AI translation, human translation, review workflows, and linguistic assets in one place, so organizations can apply different translation approaches across content, languages, and markets without managing each workflow separately.

Watch our five-minute demo to see how Smartling can provide the best software for your strategy to help you reach more customers and scale your business.