Insights from the Localization Trenches: A Recap

by Drew Evans
3 Minute Read

Building a successful services model is not something that comes by chance. Often, it takes years of experience of trial and error to understand the nuances and specifics requirements that come with delivering a result that wows.

We recently hosted a webinar with Samantha Reiss, Lilt’s Head of Services, where she shared her learnings from her experience in the trenches of localization. She’s worked to build localization programs for countless customers over her over 15 years of experience, and she’s taken lessons from each stop in her career.

Building a Localization Program

When embarking on the journey to create a scalable, world-class localization program, Samantha points to a handful of crucial questions that every organization needs to ask itself. While they may seem obvious, they can cause issues down the road if not addressed up front. Since the goal of the program is to deliver content from start to finish, it’s important to know:

  1. What is the work?
  2. Who is going to do it?
  3. How will they do the work?
  4. How long will it take?
  5. How much will it cost? 
  6. How will we know it’s right?

Ultimately, the localization industry is based on transforming content so people around the world can read, understand, digest, and utilize it in their local target language. But before the content can make its way onto the screens and pages of its intended audience, proper goals need to be set to simply understand how to get the content translated.

However, Samantha shares a warning: don’t over complicate your process. Follow the 80/20 Rule - roughly 80% of your architecture should fall into a repeatable, predictable process that is known to work, while the remaining 20% can adapt and change to suit the specifics of your project. This not only saves time, it also reduces headaches when problem solving down the line. 

Tackling the Misconceptions 

In her years of experience, she’s also encountered a number of misconceptions about workflows, processes, results, and more. 

For example, one of the common misconceptions she’s run into is the idea that the delivered translated content is accurate, even though a company’s in-country project manager needs to rework the translation upon receipt. How can that be? In her specific instance, it came down to a lack of context. 

“Using numbers is great - it allows us to show trends and compare one day to another. But what we also need is feedback and collaboration,” Samantha says. “Having the ability to take the feedback and come out with constructive action planning to improve is really what we want to harness.”

Metrics on their own require context - without that context, a 99% accurate translation may mean something very different to the project manager than it does to the translator. Getting input and insight, even from those that may not be directly involved with the localization process, can help shine new light on potential problems and make the team rethink how they define success. 

Another common misconception that Samantha has run into in her career is that the translation quality is bad and that it sounds like a machine. This is often the source of angst for everyone involved - translators, language service providers, and customers alike. 

However, because Lilt uses a human-in-the-loop approach to adaptive machine translation, the translator involved is able to provide immediate feedback that the system can learn from to prevent this problem. It helps to avoid the more literal and robotic output from raw machine translation and, in many cases, workflows where post-editing is involved. 

To learn more about building a scalable localization program that can adapt to your company’s needs, watch Insights from the Localization Trenches with Samantha Reiss on-demand.