FAQ FAQ | Contact Us Contact Us | Client Portal Client Portal
Professional Language Services

Blog

Blog
Nov 2

SAE J2450 Translation Quality Metric


How can we assess the quality of localized products?


This question is frequently discussed by localization practitioners, clients, and educators. As a company practicing localization and translation for over 20 years, Language Intelligence is also taking an active part in this discussion and believes that translation quality metrics can greatly facilitate the evaluation and measurement of localization and translation quality.

The translation evaluation process that Language Intelligence has developed is largely based on the SAE J2450 Translation Quality Metric which was created by a task force of the Society of Automotive Engineers. The metric was developed to establish a consistent standard against which the translation quality of automotive service information can be objectively measured, regardless of both source and target texts and of whether the translation is performed by a human or a machine. Even though the metric was originally designed for automotive industry, it was later modified and adapted for other industries, including Life Science (medical device, pharmaceutical, etc) and manufacturing.

The metric consists of four parts:

  • Seven primary error categories which cover such areas as terminology, meaning, structure, spelling, punctuation, completeness, etc.
  • Two subcategories (serious & minor )
  • Two meta rules to help evaluators make a decision in case of ambiguity
  • Numeric weights for each primary and subcategory

The aforementioned parts of the SAE J2450 are reflected in both the translation evaluation form and qualification evaluation form which Language Intelligence uses to evaluate its approved and prospective translators, respectively. Evaluations are performed by approved Language Intelligence translation resources who work in the same language pair and who have related experience and subject matter expertise.

Language Intelligence puts great emphasis on the periodic evaluation of its approved resources. Based on the company’s policy, any project of approximately 800 words in length is eligible for evaluation. A project with a word count in the range of 800-3000 is usually evaluated in its entirety.  For projects exceeding 3000 words, project managers identify a representative sample of 800-3000 words for evaluation. Over time, evaluation scores are carefully analyzed by the company’s project and quality managers in order to identify trends and appropriate actions.

Over the past few years, the metric has been extensively used by the Language Intelligence team. Even though it has its limitations and needs to be extended for certain projects, the SAE J2450 has proven to be a useful tool for objectively evaluating, measuring, and improving translation quality.

For more information on the SAE J2450 or Language Intelligence’s quality management policies, please contact Rick White, Director of Client Services and Quality Control Manager.

Rick White

Rick’s sixteen year tenure with Language Intelligence has allowed him to participate in many different aspects of the localization process. Rick’s time spent in the DTP team, as a Localization Project Manager, and as a Key Account Manager, have created an excellent base of experience for his current role directing the Client Services function. By combining this experience with his creative ability and strong communication skills, along with the insight he has gained into our clients industries, Rick is able to produce solutions that are a perfect fit for each of our client’s specific needs. Rick has a BA from SUNY Fredonia and has studied German.

TwitterLinkedIn


← Back to Blog