The importance of quality in localization and how to measure it

September 21, 2022

Click play to listen to the article

There is a lot to say when talking about quality. During the pandemic, audiovisual content has increased exponentially. And in times when clients are very focused on quality and all Language Service Providers (LSPs) claim to have the best quality in the industry, it is a challenge to understand what quality means and why it is important.

In this article, we will focus on what quality is, how it is used and measured, as well as on various factors that affect it.

What is quality?

First and foremost, quality is presumably the absence of errors. In order for a translation to be acceptable, it must be error-free. But there is more to it than accuracy. If that wasn’t the case, then it would be next to impossible to comply, as no-one is able to produce an immaculate text, because all humans make mistakes or, at least, subjective choices.

The definition of quality will also depend on who is judging it. Everybody has their own view of what it means for a text to be perfect. Linguists could be language experts, but the client has the best vision of who their target audience is and what the text will be used for, as well as a full picture of the product and the subject. Translation is, after all, a form of art.

In the end, quality means that a text is error-free, the language flows, the users understand it, and it serves the purpose it was created for. So, why and how do you measure it?

What is the goal of measuring quality?

The goal of a quality program is to identify the level of quality produced and to take the necessary action to improve it. Both LSPs and clients should have a quality process in place to measure the acceptability of the delivered content.

LSPs undoubtedly need an internal quality process. The principle “if you can’t measure it, you can’t improve it” will require quality-related data, the main asset of any quality department. The analysis of said data and any conclusions you come to will affect the linguists’ onboarding, training, and improvement measures. LSPs need to correctly set expectations and manage projects accurately, and this will be measured by both reliability and quality metrics. Quality assurance not only aids in the detection and elimination of errors prior to the delivery of work, but also helps rank linguists based on their performance and develop measures and strategies to reduce errors in the future.

Clients rely on LSPs to manage the whole localization process without realizing the importance of their own involvement, of learning what localization entails and which points are critical for success. Not only would they gain more control by having a basic quality program, but they would also be able to support LSPs by clarifying expectations, providing feedback, and adding clarity.

How to measure quality

We already established that each individual focuses on different criteria depending on their point of view when judging a translation, which results in disparity during the evaluation. Just doing a review is a simple task, but setting up a whole process is a more complex endeavor.

Quality assurance is a cumbersome process that involves many steps, some of which are listed below:

  1. What to measure: The company should define which information needs to be collected. All data gathered should have a purpose and be aligned with the business objectives, and only focus on what is important. Metrics may include the number of errors as per the total number of words, or pass/fail vs total tasks, number of poor feedback received, etc.
  2. How to measure and by whom: To ensure consistency, the same team should be in charge of the evaluation, using the same process and principles. Defining the sample size and using a specific set of criteria or questions that are meaningful and reflective of our goal is a cornerstone of consistency and fairness. Finally, a metric to assess samples is needed, such as 1 to 10, fails vs passes, a percentage of 0 to 100%, etc.
  3. What the baseline is: In other words, define what quality is and what quality-related performance indicators are. Having a clear picture of what is acceptable vs. what is unacceptable will serve as the foundation for determining how to improve.
  4. What to do with the data: All data gathering should have a purpose and a use, such as analyzing quality metrics and improving processes or other aspects of the business.

How to achieve quality

Achieving quality is more than simply measuring it and taking the necessary action. Below are some other processes that should be in place to help achieve quality.

Qualification and training

All LSPs working in the localization industry should include an onboarding process for linguists to ensure their primary skills. Not only their skills need to be evaluated –– all linguists should go through a trial period and receive training to straighten the learning curve. Linguists should get help to become experts in both the task and the client’s content by being monitored and provided with regular feedback, as well as with guidelines and reference materials.

Reference material and guidelines

As previously mentioned, it is in the best interest of clients to provide reference material and feedback. Clients are the experts in their own industry and products. Compiling existing style guides, glossaries, previously translated materials, or reference materials before the start of the language process can greatly shorten the learning curve and guarantee the quality and consistency of final translations.

Quality data analysis

Following the collection of quality-related data, the next step in the quality assurance process is to extrapolate trends and patterns. This analysis will help understand the level of quality desired in comparison with the actual quality and determine the course of action, such as make changes in the process, offer training, provide additional guidelines or references, or initiate a special process should this be needed for a given linguist.

Feedback and communication

Client’s attention to guidelines, expectations, communication, and feedback is an essential part of the linguistic process that should be replicated with linguists. Sharing examples of excellent quality and the issues found is extremely important, so linguists could learn, use or avoid certain patterns in the future. Client’s feedback should be tracked and incorporated into all future translations, and root cause analysis should be performed on all feedback to identify and avoid flaws, as well as to implement good practices.

Other considerations

Finally, quality will also be determined by the cost, visibility, and speed of the service provided.


Having a quality program is beneficial, no matter how basic it is, as both clients and LSPs should be able to determine the level of quality they provide. But before getting lost in the nitty-gritty, we must remind ourselves of the importance of the process. Instead of focusing on every minor error, the purpose to be served is to make sure that audiences from different markets can access the content, that the original meaning of the message is preserved, and that the brand is expertly positioned in the market. Focus on what needs to be measured, and take action based on the data gathered to make the most of it.

At Dotsub, we have a quality program that centers on the analytical data as a means to evaluate the quality that follows industry standards and to ensure that clients receive the highest quality possible, but adding a holistic level can prove to be useful to identify if the translation has served its purpose, especially in industries focused on stylistic nuances, such as the production of audiovisual content. 

If you want to know more about our processes or about how we handle quality, feel free to reach out to us at

Want to learn more? Contact Dotsub today at

We’re committed to your privacy. Dotsub uses the information you provide to contact you about our relevant content and services. You may unsubscribe at any time. For more information, check out our privacy policy.

Share This