DeepL on DeepL: How we’re driving global success across our Localization teams
Calling all localization professionals! Have you ever felt like your Localization team is not living up to its full potential? Is your company counting on you to introduce Language AI into your workflow, but you’re not sure how to leverage it efficiently?
Well, you’re in the right place! Welcome to the first of our DeepL on DeepL blog series, where we’ll be pulling back the curtain on our own DeepL Localization team to show you how we use our own Language AI to maximize our team’s impact and efficiency.
A close look at localization at DeepL
First things first: How does localization work at DeepL? Our team is made up of an army of localization specialists, who expertly localize Marketing and UX content into 12 languages, being careful to balance quality with cultural nuance and accuracy.
But how do we enable our specialists to stay efficient, aligned with stakeholders, and equipped with the best tools and technology? In this article, we’ll dive deep into how we leverage cutting-edge tools and technology — together with thoughtful strategies — to drive our success.
Our tooling and technology
Every team relies on the right tools and resources to stay organized, and we’re no different. From project management systems to CMS platforms, we’ve got the tech to streamline workflows and foster seamless collaboration with our stakeholders.
But the nexus of our work stems from our translation management system (TMS) — the engine that drives our content localization. And this is where the DeepL technology truly shines.
What you see highlighted in the image above is the translation provided by DeepL, integrated in our TMS — Phrase — via the DeepL API, which gives our team direct access to our Language AI, enabling them to work faster and more efficiently.
Now, imagine you follow our lead and invest in the top-tier translation technology for your team. Sounds like the perfect solution, right? But here’s the catch: Even the most advanced AI translation won't reach its full potential without a clear content strategy, as you'll likely encounter various challenges, such as:
- All content is treated the same. When everything is considered "important", nothing truly stands out. Without clear priorities, your content loses direction, and your team ends up investing time and energy in projects that offer little return on the company’s goals.
- You’re drowning in a sea of localization requests. As your company’s demand for localization grows, so does the workload. With an ever-growing mountain of content to manage, your team can quickly become overwhelmed, leading to bottlenecks and missed opportunities for more strategic engagement.
A solid localization strategy is key to rising above these challenges. Here’s a close look at how we approach ours.
A “DeepL on DeepL” use case
To truly elevate your team’s productivity, it’s critical to follow these three steps:
- Connect with and understand your stakeholders
- Categorize and prioritize your content
- Apply the technology
What does that look like in practice? Let’s dive into an internal use case:
1. Connect with and understand your stakeholders
Our Knowledge Management team was hitting some bumps in the road when it came to localizing our Help Center. Specifically, the process of requesting and receiving translated content was overly manual, requiring the team to bounce back and forth between too many tools. In addition, turnaround time was slow, hindering the team’s ability to constantly refresh the Help Center with the latest updates.
2. Categorize and prioritize your content
After aligning on key pain points, the Localization team took a closer look at the content they were localizing and reassessed their priorities. The question became: How can we ensure we’re investing the right amount of time and effort in each type of content? The answer? A content matrix — think of it as your strategic roadmap for decision-making:
By evaluating the level of impact and visibility of each piece of content, you can find the perfect balance between human expertise and AI efficiency.
After having identified that the Help Center content fell into the “low impact/low visibility” category, our Localization and Knowledge Management teams hatched a bold idea: What if we fully transitioned to machine-translated content for the Help Center? The benefits were clear: smoother processes, quicker turnaround times, and an overall boost in efficiency, freeing up both teams to focus on higher-priority tasks.
3. Apply the technology
After completing this careful analysis, it was time to bring the plan to life. Both teams launched a new, streamlined workflow that incorporated our DeepL glossary to make sure our product and feature terminology was accurately translated and perfectly consistent across all of our content. This new approach quickly became the cornerstone for maintaining the Help Center's top-tier quality.
As the journey unfolded, our teams meticulously tracked key milestones in the Help Center translations:
- Total time invested per month plummeted from approximately 112 hours to less than 30 seconds.
- Total cost per month dropped dramatically from €2,475 to just €0.18.
- Average delivery time for translations shrunk from 29 days to a swift 10 minutes.
These changes reflect a remarkable 99.9% reduction across all metrics! In so doing, we revolutionized our workflow from two teams dedicating countless hours each week to an almost fully automated process. Content that once took weeks to publish is now almost instantaneously published in every language, showcasing our commitment to efficiency and excellence.
Our teams also uncovered some valuable insights. For instance, one crucial observation was the need to write content in a way that is easy for machine translation to process. This means avoiding idioms and unclear language, and using simple sentence structures to ensure the AI can provide accurate translations. By embracing these best practices, the team found that they could minimize imperfections and elevate the final output.
Curious to learn more about our workflows and stakeholder engagement? Check out this webinar, in which our Head of Localization, Morana Perić, unveils the inner workings of our processes and strategy, and our exciting roadmap for 2025.
And stay tuned for future posts in this series, as we bring you behind the scenes to learn more about how DeepL uses DeepL to maximize productivity and performance.