What is Translation Memory, and why do you need it?

Translation memory takes the words and phrases you’ve already translated and automatically puts them to work throughout your projects. Learn more here.


Cloud translation management systems (TMS) have truly revolutionized the way brands localize their content. Powerful features, like Artificial Intelligence and automation continue to improve translation quality and speed as well. These advancements have made reaching audiences easier by producing fast, accurate, and native content.

Computer-assisted translation tools, like translation memory, cut down translators’ work and provide cost savings for companies. Instead of retranslating and paying for the same content, translation memory automatically handles these known words and segments of text.

This article explores the potential for translation memory to reduce redundancy and improve translation results. It also discusses what translation memory is, how it works, and why organizations should prioritize it as an integral part of their translation management software.

What Is Translation Memory?

Translation memory is a linguistic database that saves previously translated sentences, paragraphs, or phrase segments from a source and a target language. When similar text appears in a new translation, the system automatically suggests these pre-translated elements to the translator through a process called automated translation.

This database can also work with flexible parameters to suggest the correct language and context. Top-performing translation tools offer perfect match (also called exact match) or fuzzy match options that widen the possibilities.

Translation memory ensures consistency in language usage, accelerates translation speed, and reduces workload by eliminating redundant translations. It also enhances productivity and maintains high-quality translations across various projects and languages.

The benefits of translation memory

Translation memory removes the repetitive elements of human translation. It also leverages the power of string searching and indexing to ensure that translators can focus on novel content versus common phrases.

Using translation memory in a TMS provides various benefits, such as:

Consistent translations: Human translators make mistakes or reinterpret phrases based on context. In large-scale projects, this variation leads to issues with consistency. A translation memory ensures that every instance of a word or phrase in the content remains consistent throughout the translation process.

Cost efficiency: Translation from scratch can be time-consuming, so repeatedly translating exact match phrases drives up the project's cost. Translation memory reuses strings without added costs. Additionally, with previously translated strings removed from the scope, the overall translation cost will decrease.

Improved quality: Both project managers and linguists can refine and optimize translation memory entries to adhere to brand tone and voice. This ensures that the translated content will be technically accurate and will reflect brand guidelines. This benefit also extends to future translations.

Centralized database: The translation memory within a TMS is accessible to every user and department in the organization. This ensures a consistent application of brand and voice, even in projects with multiple stakeholders and translators.

Faster time to market: Reducing repetition in translations speeds up the process and helps translators save time. It also reduces editing and post-production time. Overall, translation memory allows teams to deliver finished translations that are ready for market entry on a shorter timeline.

Higher revenue: Recent reports reveal that 40% of consumers will not buy from a website in another language, even though 66% use some form of online machine translation. A tool that produces native results will attract more buyers.

Expanded global options: Machine translation now supports over 190 thousand language pairs to ensure smooth content translation across any set of languages. With each improvement in technology, organizations have the ability to reach out to more markets and build a global business.

Translation memory in Smartling

Smartling is a TMS with robust features to support both machine and human translators and optimize the process in both categories. It features translation memory to speed up the production process and ensure that translators can focus on fresh content with consistent results.

Smartling improves upon basic translation memory capabilities with several unique features:

  • Support for multiple translation memory databases
  • Translation memory support across any content type from any location
  • Customizable string entries with move, delete, and edit functionality
  • Searchable translation memory by keyword, author, language, translator, and more
  • The ability to import an established translation memory
  • Tools to export and download whole or partial translation memory data

These features enable faster translations and more efficient workflows than simple translation programs.

Traditional translation memory management versus Smartling

Smartling provides brands with control over translation memory. All data is visible and accessible within the Smartling platform.

Users can perform terminology management functions like modifying existing translation memory entries, refining entries for tone or context, and removing data as needed. Brands maintain control over their translation memory database with easy access from within the platform.

The traditional editing process for translation memories is manual and error-prone:

  1. A user exports the translation memory into an Excel spreadsheet.
  2. The editor makes necessary changes within the spreadsheet.
  3. After editing, the user uploads and re-imports the database into the TMS.

This manual process may introduce errors into the translation memory or may result in a corrupted file. With Smartling, users update, edit, and remove individual entries within the TMS without affecting other entries or risking database corruption.

Users can also move content between existing translation memories, export memories, and use platform search features to locate entries by language, author, dates, and other parameters.

What is the difference between translation memory and machine translation?

Translation memory and machine translation are both important components of a translation platform. However, the two terms aren’t interchangeable, and each serves a different purpose within a translation project:

  • Modern machine translation uses advanced neural networks to interpret a source text and generate a translated document. This process is quick and helpful for understanding the source document’s major points. However, machine translation does not always provide accurate or culturally appropriate translations.

  • Translation memory, on the other hand, stores previously translated phrases and sentences from human translations. It recognizes similar phrases and inserts the corresponding translation from the language pairs defined in the database. It is a valuable tool for maintaining consistency and accuracy in translations while still allowing for human oversight and adjustments.

Users can achieve efficient, high-quality translations by using machine translation and translation memory within a TMS. Machine translation provides a starting point for translators to work from, while translation memory ensures consistency across the most common phrases and words in different assets.

Translation memory, glossaries, and style guides

Like a glossary or style guide, a translation memory helps professional translators maintain consistency in brand voice, tone, and style. However, translation memories are more complicated than style guides and glossaries, which establish guidelines for multilingual terms and basic usage rules. Translation memory instead houses source content and associated translated content.

When should I use translation memory?

Translation memory is most beneficial in large-scale or recurring translation projects where it’s more difficult to ensure uniformity across common terms and phrases. Below are possible uses:

  • Translation memory helps to ensure consistency and speed up the process when your content contains frequently repeated phrases, jargon, or industry-specific terminology.
  • It enables multiple translators to work on a project simultaneously and helps to maintain consistency across individual assets.
  • Translation memory is also helpful for files that require updates. It allows teams to reuse previous translations, minimize effort, and leverage better pricing.

Boost efficiency with Smartling’s machine translation and translation memory

Smartling offers organizations access to AI-enhanced tools and translation services, which make it fast and cost-efficient to recreate every important asset in multiple languages.

Machine translation cuts down translation time and provides accuracy. The translation memory tools within Smartling’s platform allow companies to further refine automated translation to preserve brand style, tone, and context during the translation process. The result is a cost-effective way to create personalized, native content for every target language.

To learn about building scalable content translation with a localization process that passes the test for native audiences, download our ebook: A Translation Solution that Grows with Your Content.