Listen on :
Our Guest
Mark Lesun
Localization Quality Manager, Booking.com
Our Host
Alpi Mantry
Chief Human, Translate By Humans
Prepare yourselves, ladies and gentlemen, as we embark on our eighth electrifying tale, where we are talking to our QA experts in the localisation industry. We are thrilled to introduce our exceptional guest for today, Mark Lesun, Localisation Quality Manager at Booking.com. They’re determined to make the world of travel more sustainable, accessible, and inclusive, to create a positive impact on a global scale.
Join us as our CEO, Alpi, gets into conversation with Mark about his remarkable journey from a translator to a QA leader, the intricacies of quality assurance in localisation, the ever-evolving role of machine translations, and, to top it all off, some invaluable advice for fellow QA Managers in the field.
Welcome to the show, Mark!
Key Takeaways:
Key Insights:
Can you share your journey from being a translator to leading the QA department at Booking.com, and what has your experience in the language industry been like?
I began as a translator while studying Social Psychology. I transitioned to automotive industry translations, facing imposter syndrome initially. Progressing from translation to editing and coaching, I later became a reviewer for Google. Then, I became a language manager at Google, focusing on coordination and collaboration. After five and a half years, I joined a vendor briefly before moving to Messagebird, where I set up localisation processes. Eventually, I joined Booking.com as the head of the QA department, overseeing quality and managing a diverse team.
What are the key differences in approaching a localisation project when you are the vendor, the project manager/coordinator, or from a QA perspective?
The differences in perspective depend on whether you’re in QA or on the vendor/client side. In QA, you set expectations and ensure everyone understands them. On the other hand, you gather input and sometimes nudge for materials. It’s about managing stakeholders and teams or adopting a client vs. service mentality. Both are fascinating with unique challenges.
Can you provide insights into the QA role, localisation queue management processes at Booking.com, and the associated challenges you encounter?
In the journey from translator to QA department head at Booking.com, I’ve followed standard quality management processes like monitoring, issue resolution, and root cause analysis. Collaboration with colleagues and staying updated on industry best practices is crucial to avoid “reinventing the wheel.” Challenges can vary due to company-specific setups, with tools adapting to each situation. At Booking.com, we have 40+ linguistic teams, offering opportunities but also requiring efficient logistics and communication due to the size.
Okay, so do you engage with any Language Service Provider (LSPs)?
We collaborate with LSPs, and our external resource program is evolving to meet our growing needs. We aim to create a future-proof system that minimises frequent adjustments. Working with LSPs has challenges, but it’s a convenient way to outsource tasks and ensure clear communication with vendor teams.
So, considering all of this, do you think communication has been a challenge to keep everybody on the same page?
Localisation is always a challenge, whether you’re on the human translation services provider (LSP) or client side. It involves many people with different habits and communication styles. Information can easily get lost or forgotten. Vendor linguists work with multiple clients, so keeping guidelines accessible is crucial. Creating centralised guidelines helps but requires constant follow-up. Ongoing feedback is essential for improvement. These are the main challenges in this setup.
Are there any additional steps you'd recommend to address these challenges besides centralised guidelines and a feedback loop?
To ensure success in localisation, sharing critical resources like style guides, glossaries, and reference materials is essential. Complex content types, such as marketing materials or presentation decks, benefit significantly from shared references so the team understands the intended look and feel. Collaborating closely with stakeholders for clear instructions and maintaining open communication channels between translation and review teams fosters community and teamwork, leading to a successful localisation program.
Yeah, good one. Okay, thank you, Mark.
Since you are our first QA specialist on this series, I wanted to ask how you maintain brand consistency and tone across languages and regions. Markets?
Maintaining a consistent brand language can be tricky, but it’s essential. Style guides help define our brand’s voice for different content types. For instance, marketing needs a friendly tone, while finance content should be more serious.
Localisation adds complexity. Each market has unique needs. Local input is vital.
Collaborating with regional teams and calibrations can help strike the right balance between global consistency and local flavour.
It’s a continuous process of finding what works best.
Yeah. I mean, I've dealt with clients. As they grow, this is one thing they need help with: brand consistency across regions because, as you said, there is no definitive solution to it. But it's, you know, the most minor steps being taken, depending on the situation and the scenarios that you face. So, I agree with what you've said there.
Now, Mark, what's your take on machine translations? You must have encountered situations where machine translations have been referred to cut costs, but what is your take from a QA perspective there?
Discussing localisation in recent years inevitably involves addressing machine translation (MT), which is becoming increasingly integral. Whether we embrace it or not, it’s a force we must reckon with. While I’m not a tech expert in MT, our larger team boasts an MT specialist who’s diving into technology and shaping our future strategy. The impact of MT on the localisation landscape is poised to be substantial. Certain types of content will see a surge in MT usage.
From a quality assurance (QA) perspective, it’s a fascinating realm to explore. Many companies, including us, are already dabbling in it. We need to redefine our quality expectations based on content types. As former translators or localisation specialists, we naturally yearn for perfection everywhere, but it’s rarely attainable. Sometimes, we accept ‘good enough’ clear and understandable translations. For instance, highly technical documentation doesn’t need literary flair; engineers want precision, not poetic language.
Similarly, for instructional materials, brevity is critical. Machine translation will likely become the go-to for such content types. Human translators will shift towards QA and spot checks, ensuring no glaring errors slip through.
On the flip side, I recently discovered emerging technologies that automate QA assessments, some already in use. These tools go beyond detecting common bugs; they pinpoint errors, annotate them, and assign quality scores using methods like QF. You can receive a translation job with an automated scorecard detailing various error types like spelling, punctuation, and grammar. No human involvement is needed. While I haven’t seen it in action, it requires extensive training on bilingual files and assessed translations. However, it’s just a matter of time until we have tools capable of fully automating translation assessments. This will only replace human linguists partially, especially for high-visibility content or model training, where human expertise remains crucial.
MT is reshaping localisation, but human linguists will still play vital roles in refining machine translations or ensuring quality in the final output.
Automated QA tools are on the horizon, promising efficiency but not complete human replacement.
Absolutely, Mark, I concur entirely. Human translations are here to stay, but what's exciting is how we can harness the power of machine translations in tandem with human expertise.
Again, these translation engines, at their core, rely on human-generated content or human-guided training. Their functions are evolving, not vanishing.
Thank you for that, Mark.
As we wrap up, I would like one piece of advice you may want to give a QA Manager like yourself.
Don’t be a stranger in your professional journey! Reach out to your colleagues and connect with others in your industry, especially fellow quality managers. We all have much to gain by sharing ideas, best practices, and solutions. Imagine creating a Quality Managers’ Community where we can gather to tackle our daily challenges. Even if such a platform already exists, not everyone may know about it, so let’s expand our connections. In a rapidly changing world, staying efficient and relevant means breaking out of our company bubbles. Staying in touch with the real world and what’s happening out there is crucial.
With the ever-evolving role of machine translations, it’s clear that adaptation is key. Mark’s insights shed light on the path ahead, offering advice that resonates with both seasoned QA managers and those aspiring to join their ranks.
In a world where words and context matter more than ever, Mark Lesun’s wisdom is a guiding light for the future of quality assurance in localisation. We thank Mark for sharing his expertise and look forward to the continued journey towards excellence in translation and beyond.
Are you ready to catapult your content into the global spotlight? Your content is a goldmine of potential, poised to enrapture audiences worldwide. At Translate By Humans, we do more than translate – we orchestrate transformations.
Don’t hesitate any longer. Initiate the journey to unlock the enchantment of localisation today! Book your free consultation call with us, and let’s embark on this metamorphic expedition together!