Successful localisation is more than just “good translation”. Translation is only one step in a complex setup where people and intentional planning are as important as tools and technology.
Enterprise localisation can feel like a black box. It’s not always clear how things work. That’s why we’ve found an inside man to show us around the machine room so to speak.
Philip Philipsen has been building and managing localisation projects for LanguageWire customers for many years. In his position as Senior Operations Excellence Specialist, he has learned everything that's needed in the intricate machinery of a successful localisation setup.
Come along behind the scenes as he shares his insights.
Start with brilliant basics
Localisation setups need to be built with intention and architected correctly from the beginning. If workflow setups are based on the wrong assumptions, decisions, and priorities, fixing them later can be difficult.
Philip recommends operating from a philosophy he calls “brilliant basics”.
To Philip, it simply means taking the time to focus on the essentials up front. In short, getting the basics right, so you can hit the ground running.
Most people want to get off the ground fast, thinking they can optimise later. Most of us have heard the advice that starting with “good enough” is preferable to endlessly perfecting. And that may right in some cases, but not when building the foundation for excellent localisation outcomes.
The good news is that there’s a sweet spot to aim for. Do the foundational work that gets you to the famous 80% up front. Any more than that and the curve starts to flatten: you simply won’t get enough return on the effort you put it. Over time, you can optimise by another 5-10%, but we’re getting ahead of ourselves.
Map the context of the whole organisation
Part of the necessary groundwork is making sure you understand the entire context before getting started.
This means mapping out values, priorities, and tradeoffs, across teams, not just your own. You may understand your own department, but you may not know who works with e-learning or technical documentation.
From the perspective of the provider, Philip says, “we need to understand how it’s all connected so we can be an effective facilitator between the needs and the output”.
It’s like a playing card, with two sides. One side represents the company. The other side represents the translators, validators and other skilled professionals who do the localisation work. You need something in the middle – the card dealer – to connect the two.
Understanding the context, Philip says, is a crucial part of “playing the card” connecting everything.
Orchestrate automation, infrastructure, and the technical foundation
With the context in place, we can now begin orchestration: Setting up the technical foundation, infrastructure elements, and automation options.
This includes populating your Termbase with your own industry and company specific terminology. It means deciding on tools and processes. What do we want to use AI for, and when and how much do we need a human to step in? How do we want to exchange files and who should own the final approval?
An example: At LanguageWire, we’ve set up a connector linking our website’s CMS to the Translation Management System (TMS). We can then order translations directly from the website. A Project Manager assigns the task to our preferred human creative translator and a validator from the LanguageWire Team, all sharing the same Termbase with approved terminology.
We’ve decided on this workflow, because we prefer creative translation for our customer facing material and we like having an internal reviewer for final sign-off.
Once the project is finished, the result is sent back to the CMS. The localised content is automatically placed in the correct locations on the website.
An even more important aspect to get right is the people who do the work.
Pick the right people
Finding the right experts, whether translators, reviewers, or language specialists, and training them is a crucial and sometimes time-consuming task. But don’t leave home without it.
You have to select the right people for the different steps in the process.
It's not just about professional ability, but also about personality. Do we want someone fast and “always-on”? With highly specific skills? Do they take feedback well and are they easy to work with?
Philip’s advice for anyone who is about to build a “brilliant basics” localisation setup is to spend the necessary effort to properly enable the experts who work on the projects. Preparation matters.
“Most language experts work in solitude. Human interactions and genuine engagement matter more than you think,” Philip says.
Human connection with the experts is best established over a call, with video if possible. Communicating digitally has many benefits, but nothing replaces the rapport that happens through direct communication.
“Almost nobody does calls anymore, which is why you should."
Once you have a sense of each other, you can start exploring if the expert is the right fit. Not everyone has the same strengths, communication style, or skills. When you let the experts know they’ve been chosen because you’ve assessed that they’re right for the job, it fosters a sense of commitment that you can’t get if you just automate right off the bat.
“Doing things efficiently after you’ve created a shared understanding is fine, though – that’s when automation can work to everyone’s advantage,” Philip says.
Let the setup do the heavy lifting
If everything has been set up with thought and intention, localising content will be a (relative) breeze.
Depending on your setup, you should be able to seamlessly order projects, update your Termbase, and participate in feedback loops. The setup can basically run on its own, with only the touchpoints you’ve agreed on in advance.
It could look like this:
A translation need arises within your company, for example in the marketing team.
The team orders the type of translation they need in the TMS or connector, for example, creative translation.
The experts are already handpicked and onboarded for your content specifics, so all the project manager has to do is assign them to the task. The Translation Memory and Termbase are already up and running.
The Project Manager can be easily contacted by the team that ordered the project, in case there are changes to the project or questions from the experts.
Once the task is complete, the files are automatically made available to the team, and Translation Memories updated for future use.
Make metrics meaningful
Now that localisation has been set up, and projects are being ordered, how can you know whether you’re reaching your goals?
Philip encourages teams to focus on metrics based on the desired outcome.
The metrics should help verify whether the hoped-for result manifests, keeping in mind that nothing lasts, and that continuous adaptation and improvement will be necessary.
Metrics can act as a benchmark to see, retrospectively, if your strategy is working. But they can also act like stripes on a road, guiding the direction of the teams in real time.
Metrics shape behavior. Knowing that should impact the way you set them.
If you focus on speed and cost, those will become the team’s top priorities. If you measure end customer impact, teams are encouraged to think about how translation quality supports business outcomes.
Incentives can lead to people “gaming the numbers”, as Philip calls it. That is, working towards achieving the right digits in a dashboard without thinking critically about whether the output helps produce the desired outcome or not.
Measuring impact typically requires longer cycles to produce meaningful insights. Looking at results over a longer period—such as three months or longer—helps smooth out anomalies and gives a more accurate picture of whether the setup or strategy is working.
What metrics can’t do is truly capture the value expert knowledge and experience brings to the table.
According to Philip, expertise is equal parts experience and unquantifiable skill. "Not everything that counts can be measured, and not everything that can be measured counts,” to quote a well-known saying.
Don’t overoptimise
Now we’ve come full circle. We have set up our foundations, projects are being managed, and metrics are in place.
What is next?
Philip’s opinion is clear: Don’t let over-optimising get the better of you. While it’s true that “you can always improve things”, beyond a certain point, it doesn’t make practical sense. If you got your brilliant basics right, and your localisation setup works to consistently produce the outcome you want, then your additional efforts are likely to meet with the law of diminishing returns. At this point, your focus should generally be to preserve the excellent baseline you already have.
Good localisation lies in planning, skills, workflows and technology
Good localisation is more than translation. It’s the combined result of intentional planning, strong technology, solid workflows, and the right people for the job – both language experts and the localisation team that orchestrates the setup.
Building a solid foundation means that you’ve got the rails for a high-quality baseline. And that scaling later becomes a matter of refining the setup to reflect your volume and the individual services you require, not rebuilding everything from scratch.
Wondering what the right localisation setup looks like for your business? Let's talk.This article is the March 2026 edition of Lost & found in translation, a monthly newsletter sharing insights, opinions, and reflections by real people who work in localisation. Subscribe to get notified when the next edition comes out.
Subscribe on LinkedIn
