The language services industry represents companies that provide translation, localization and globalization services. They are commonly referred to as translation service providers, or more commonly, as language service providers (LSPs). These companies range from $400 million companies, to “mom and pop” businesses. Language Service Providers are typically hired by other companies to manage the translation process and provide timely, high quality translations. Some of these companies focus on doing translations for one particular company or industry, while others work with many companies across several industries. Some LSPs specialize in particular languages, while others cover every language imaginable. In general, the services provided and the technology used varies, but the goal remains the same: effectively translate information and manage the translation process.
The case for translation automation
Language service providers often manage all of the translations for some of the world’s largest companies and as more companies “go global,” the demand for accurate, timely translation continues to increase. However, the margins that LSPs earn continue to shrink and many of them feel pressure from their customers to shorten the turnaround time and decrease prices without compromising on the quality. In an industry where service and translation quality are key differentiators, it is difficult to improve either one without incurring additional cost.
With this in mind, many LSPs have begun to embrace automation. Many have done this for years using translation memory tools. However, now many of these LSPs are beginning to add automation to their workflow using tools provided by companies like across Systems GmbH and Idiom Technologies, Inc. Using applications like across Language Server and Idiom WorldServer, LSPs can now minimize the number of manual steps required in order to improve overall translation productivity.
Taking automation one step further, language service providers who utilize translation memory and/or have added automation to their workflow, are working with Language Weaver to add statistical machine translation to the translation process.
Typical deployment for SMTS
The purpose of deploying statistical machine translation software in translation workflow is to increase translation productivity. Language service providers will be able to translate more data in a shorter amount of time, leading to quicker turnaround without compromising quality.
A typical deployment in translation workflow is illustrated below:
As shown on the left side of the graphic above, data can come from multiple sources. All text is first sent through an existing translation memory application. Exact matches pass through to a human translator for review; new content and “fuzzy matches” go through Language Weaver translation software that has been customized for a company or subject domain. The output is then sent to a human translator for editing and review. The translated text can be posted to several locations for use throughout an organization.
This entire process can be automated with workflow management software or using tools developed by a LSP in-house.
Translation Productivity: More Information
Language Weaver and across Language Server
Language Weaver and Idiom WorldServer
Additional Workflow Deployment Options
Customization requirements
Language service providers are hired to provide high quality, human quality translations. Therefore, to see the greatest productivity impact from SMTS integration in the translation workflow, customization is required. However, customization of Language Weaver software is dramatically different and significantly faster than customization of rule-based MT. For more information, please visit our Customization Process page.