Localization in a Startup


About a year ago, in the middle of the pandemic, I made the bold decision to move to Barcelona with my family to join Preply, an ed-tech startup company which opened its doors in 2019. As the company’s head of localization, I would be building their Center of Excellence from scratch.

Working at startups is in my genes. After graduating from college, I started working for my dad at his recruitment agency. A few years later, wanting to stay closer to my educational background, I joined a translation agency as a project manager. The agencies I’ve worked with since then have always been small, fighting for their place in the industry. 

In 2012, I decided it was time to move to the buyer side. Up until then I had worked to deliver great translations to clients, and I felt it was time to understand what was going on on the other side of the fence. Why was it that we always had to work with such tight deadlines? Why didn’t the client ever want to hear about our challenges?

I joined the Dutch gaming company Spil Games as a localization project manager in a team of six people, all from different cultural backgrounds. What an eye-opener. Th is was an eight-year-old startup with a very creative vibe but simultaneously it felt like we were constantly in competition. And we were. We had quarterly, sometimes monthly, all-hands meetings in which the latest numbers were shared. I’d never been in a company that tweaked its yearly objectives as frequently as this.


One of the biggest things I learned during the seven years I worked at Spil Games was that in order to drive a successful localization strategy you need to understand the main challenges of the business. You also need to be flexible enough to adapt to change quickly. Decisions you make will often not be about internal team topics such as how to achieve better quality, building your in-house team, or setting service level agreements with your vendors. It may very well be about how we, as a localization team, can contribute to the overall success of our company. Th e smaller the company, the more individual  contributions count. 

The speed at which a company operates, the lack of bud-get — especially for localization — and an overall messiness of context can be challenging. Obviously, a company that is just getting off the ground or one that has only existed for a decade, will not have the same solid structure and framework as a large corporation. This can mean it will be harder to secure a budget for your projects. Get used to not taking any support for granted on your initiatives. Instead, business cases, proof of concept, and return on investment (ROI) are extremely important to get the go-ahead. 

And as you know, the combination of “proof of concept” and “ROI” when it comes to localization is challenging, to say the least.


Startups are fast-paced, often with a shorter-term vision for the company. The main focus is innovation, dealing with competition, possibly securing funds, or finding investors. As a startup, the pressure to look for opportunities on the market, as well as threats, is high.

My husband has been working in the pharmaceutical industry for the last 12 years or so, usually in huge corporations. The biggest difference between those companies and a startup, in terms of localization, is that while responsibilities are well–defined but limited in a multinational, in a startup you are often the only person responsible for localization. If you’re lucky, you’ll have a mini-team to work with. On the other hand, the great thing about this is that you have easy access to management. 

This means you will get to communicate your ideas for innovation, how to deal with new markets, etc. You have an enormous amount of freedom to do what you think is best. 

The startup will also give you a chance to learn about business impact much faster, which helps in making strategic decisions about localization in the future.


In any role — but especially key ones with a lot of visibility — your first 90 days are crucial. Of course you cannot change the world in just three months, but you can set a solid basis and get the buy-in for you as a professional and for your localization strategy. There is a lot of great literature out there that I highly recommend reading, to help set up your plan. I also have a few tips on how to go about it from my own experience.


It is appealing to either just go and make sure localization is done within a company, relying on the existing solutions, or to do a full sweep straight away to make your mark. Startups move fast, and people expect to see results quickly. However, it’s extremely important to establish your benchmark. You need to know what has been done already, how it is done, and therefore, what exactly you need to do to create a mature localization strategy.

Talk to people as much as you can. See what they do, how they have been managing the localization process before your arrival. Even if there was no localization manager before you, all companies with a multi-market strategy will have done some form of translation or localization. 

Create a list for yourself with the topics you need to cover. A few examples: 

  1. The business — The absolute first thing to do, is learn about your new company. What are the main objectives and challenges? What does success look like? Who is the customer, if anyone? What metrics are important?
  2. Internationalization — How prepared are they currently to scale?
  3. User experience — How do customers perceive the product in the source market and the localized user experience.
  4. Processes — What process is there for localization? How do people manage it? Is it centralized or not? Are agencies used, and if so, one or multiple? 
  5. Market opportunities — How to decide what market to tap into? Are you localizing into the right languages? Are there any missed opportunities? Or are you maybe focusing on the wrong markets? 
  6. Linguistic assessment — What is the quality of the output? How do you decide which variant of the language to select (e.g. European Spanish vs Latin American Spanish or Euro-pean Portuguese vs Brazilian Portuguese)?
  7. Gather all of this information and write it down somewhere. 

Eventually it will be part of your localization strategy.


Internationalization is an important, yet often overlooked, factor. Establishing the status quo also means looking at how internationalization best practices have been applied. Is the website coded in a way that allows for scaling into new languages? That is to say, is the text hardcoded, how are strings managed, and are best practices used to localize dates and currencies? These are all questions you need to find answers for. Doing so is important to creating a fuller picture of how things work.

 This can be tricky, especially in a startup, depending on your technical knowledge of localization. But again, in this phase, what you need to do most of all is simply gather as much information as you possibly can. Just try to understand how it works. Finding the right person in the company to help you understand is a big part of the job.


Another important element to familiarize yourself with is the current localization model. Any company will consciously or unconsciously follow a model. They will either use one agency, multiple agencies, freelancers, or in-house native speakers. Look at what the current model is, what the pain points are (quality, efficiency, scalability) and what model would work best for the company. 

Your chosen localization model will be part of your overall strategy. There’s not really a right or a wrong way, but there will be a model more suitable for the required speed, quality or seamless integration with the business you want to create. 

Following are the most important localization models, and their pros and cons.

01 – Outsourced, single language service provider (LSP) 

A very common model is that of the single LSP. It makes the process very easy, as you just have one point of contact. It’s also scalable, as most LSPs have a broad language pool. The down-side to this model is quality, mostly. As a single vendor, they take care of both the localization and quality control for you. You could potentially mitigate that by having in-house subject matter experts who can double check the agency’s output. However, the smaller the company, the less people will have time for this. That leaves the customer as the ultimate feedback channel, which may be risky. 

02 – Outsourced, multiple LSPs  

In this model one agency is responsible for localization and another for review, but it’s also possible to work with individual agencies for specific languages.

A two-vendor setup where one trans-lates and the other checks can definitely be entertaining; as the two agencies are in fact competitors, it’s very tricky to achieve a completely objective process. We’ve had instances where we passionately followed a string of arbitration emails between two agencies, with nice suspenseful music in the background — or fighting kitties, whatever suited the context best. 

The upside to this model is that you can keep a critical eye on your localizations, which should lead to better quality. 

03 – Outsourced freelancers 

What is great about this is that the linguists really feel like they are part of the team. They are usually very knowledgeable about the product and the context of what they are localizing. 

However, dealing with freelance linguists requires a lot of overhead, and it’s difficult to scale with this model. 

04 – In-house linguists 

If your company has the budget for this, that’s great. It allows for flexibility, speed, and quality. However, the budget is important. In order to scale, you need to hire more and more people and managers to oversee production. Most startups will not go for this option, simply because it’s too expensive.

05 – Hybrid model

This is the model I currently use. We have a team of five language managers, a project manager, and an external team of 14 linguists. This model allows for speed, quality, and flexibility. The downside is scalability. If the in-house team and the pool of freelancers grows, you need more resources to manage everything. A hybrid model can be more scalable if you work with in-house specialists and an LSP.


Once you establish the status quo within your agency, it is time to build your strategy. Take all the knowledge you’ve gathered, look at where the pain points are, and provide a solution through your strategy.

Make sure the strategy can be understood by management. Refrain from language industry specific jargon, such as LSP, LQA, TAT, leverage, TM, etc. Instead talk about language service providers and agencies, quality control, delivery speed, reuse of content, translation memory, etc. 

Also make sure to connect your localization strategy to the overall business objectives.


For example, it’s helpful to ask yourself the following questions:

Is it in your company’s goals to grow three times their cur-rent size within the next two years? Make sure you focus your strategy on how to achieve that. Does it mean you need to find a cost-effective solution? Drastically increase your quality? Is it time to finally fix that incredibly poorly done Arabic website your team put together God-knows-how-long-ago? Whatever seems most important to the company, focus on it. 

In my case, a goal for the company was to seamlessly integrate localization. So my strategy was focused on automation solutions and scalability.


The first months in your job at a startup are crucial to make your mark and get localization on the map. It’s therefore important to not try and tackle the biggest challenges immediately, but rather to focus on things that are easy to achieve and still impactful.

What we did as one of the first projects was a content audit of our main languages on the website. The audit itself took a few weeks. We then proceeded to do a mini-research project with our users. This gave us a great opportunity to work together with the research team, and learn about setting up simple research projects. The outcome of that research project would confirm whether our assumptions about the quality of the localization were correct or not. 

Finally, we implemented the changes on the most important pages of our website — the landing page, search, and check-out page — and then began measuring the key metrics; traffic, bounce rate, conversion. We took those insights and communicated them to the company. 

It was a fairly easy project to execute, and the results were so interesting they inspired an internal dialogue about localization. This is the kind of impact you need, to get your show on the road.


Throughout this article, I’ve mentioned metrics and key objectives a number of times. Data and localization are a challenging necessity to formulate and successfully execute a localization strategy. In the end you need something to show what success looks like, otherwise it’s all just a bunch of fluff. You can have localizations of the highest quality and the fastest turnaround times possible, but if this high quality and speed cannot be reflected in any of the metrics that matter to the business, your efforts will go unnoticed.

In the case of Preply, we are an online marketplace. We offer space for tutors to present themselves, so we help learners to make the best decision. Metrics that support this business are in the area of traffic, conversion, and retention. We did a very cool project where we localized emails and created videos in the tutor’s native language, to give them an extra push to complete registration. Quality is important here, as well as a natural sound. But the metric we will report success on is conversion to registered tutors.


I can say I have never been happier than I am working in a startup. It is full of creative, entrepreneurial people, who aren’t afraid to fail and learn. My colleagues’ drive makes me want to continue to learn as well. It’s very gratifying to be in an environment where there is so much energy and passion to make the business a success. Use this energy and passion to learn from other departments, solve the puzzle you find when you first join, and then come up with a solution you think works best.

Listen, learn, connect, design, and execute. And of course, rinse and repeat. Because, as you well know, the localization industry does not sit still. The solution you thought of today may not be the best one tomorrow. Keep challenging yourself, and see if it can be done better.