In this chapter, we're diving into the exciting world of technical tools and software. Now, don't fret if you're not a tech wizard—these tools are generally user-friendly, and we're here to give you a quick rundown of what to expect.
In this chapter, we'll focus on four main categories of tools that you'll likely encounter in UX localization and other related fields:
We'll explore each of these categories in more detail to help you understand what these tools are and how they can support your localization efforts. Then, we'll also discuss some ways to future-proof ourselves by working with advanced and new tech.
By the end, you'll be able to confidently say "yes" when a product team asks if you're familiar with a particular tool.
1. Machine translation
Love it or hate it, there's no denying that it's become a significant player in the localization industry.
MT has been around for decades, but recent developments in AI and natural language processing have made significant improvements in its capabilities. While many translators are quick to dismiss MT as inferior to human translation, it's crucial to remember that the tech industry is, at its core, a business focused on making money.
Commercial interests drive innovation and progress, so we want tech companies to be successful and manage their budgets effectively. This, in turn, means that they'll have the resources to invest in high-quality localization services. We can't deny that MT is here to stay. Often, companies can't justify a 100% human localization process due to costs, but a hybrid process involving both humans and machines working together can be a feasible solution. The sooner we embrace this reality, the sooner we can help clients understand the importance of human review in localization and how a professional UX localizer can make a significant impact on their business results.
When we approach the conversation without judgment, we can guide product teams to use MT wisely. Some copy types work well with MT, while others suffer from its limitations.
How does machine translation work?
MT engines learn by analyzing vast amounts of content and identifying patterns between words and phrases, similar to how humans learn. However, just like human translation, MT engines need context to be accurate.
With plenty of context, MT can handle texts well. But UX copy often lacks context, which can lead to poor MT results. So, it's all about balance—MT isn't inherently evil; it's a tool to be used wisely and in moderation.
Case study: Airbnb
Take a look at this screenshot from the Italian version of Airbnb's website. Notice that the UX copy—titles, buttons, links—is professionally translated by humans. However, for larger chunks of copy, like reviews and listing descriptions, Airbnb opted for machine translation.
By clearly stating that some parts were automatically translated, Airbnb is being transparent with their users. They even provide an option to view the original copy. This well-thought-out hybrid process demonstrates Airbnb's commitment to offering a great experience for their Italian-speaking users and their global audience in general. But they took this a step further, too.
In a Slator podcast episode featuring Marco Trombetti, CEO of Translated, he explained that Airbnb wanted to machine translate all user-generated content. However, using existing models like Google, Amazon, or DeepL would have been too costly and wouldn't deliver the desired results.
Remember how machine translation models "learn" from the content they're fed? To perform well with Airbnb host descriptions, the model needs to be trained on high-quality, similar descriptions. This helps the model learn patterns and produce better results.
Airbnb took a unique approach. They continued to pay for manual, human translations during a "grace period" and collected data from these translations. They then used this data to train their own proprietary language model. Today, all user-generated content on Airbnb is machine translated by default using their custom model.
2. CAT tools
Let's go back in time and talk about CAT tools. CAT stands for Computer Assisted Translation. In plain terms, CAT tools are computer software designed to help people translate better. You can divide them into 4 rough categories:
Gen-1: Standalone software
These types of tools have been around for quite a while, and if you've done any translation work, especially through an agency, you've probably encountered some of the more traditional CAT tools. Early CAT tools were software you installed on your computer, with no online capabilities. You'd receive a translation file from your client, work on it, and then email it back to them once completed.
Gen-2: Going Online
With widespread internet access, translators began using CAT tools with online capabilities. These tools synced translations with client servers, eliminating the need for sending files back and forth. Some even offered web versions, allowing for more flexibility and convenience. However, these tools were still geared toward traditional translation.
With content stored on your computer, teams only gained access to it after you synced your files. This setup limited visibility, real-time collaboration, and communication.
Gen-3: Cloud Applications
To address these limitations, a new generation of CAT tools emerged as cloud applications. With tools like Memsource, Smartcat, and XTM, everything is stored online, and nobody needs to send or sync files. While the lines between generations are blurry—Memsource merged with Phrase (a Gen 4 tool) and Smartcat continues to develop new capabilities—this categorization helps illustrate the available tools.
Gen 3 CAT tools offered similar features to previous generations but still couldn't meet the tech industry's needs for effective UX copy localization in every language.
Gen-4: Localization platforms
These systems are designed specifically for app and digital product localization, focusing on the tech industry's needs. They integrate with a company's tech stack, giving localization teams the tools they need to produce high-quality results.
Localization platforms have many traditional CAT tool features but also offer localization-specific options, like pseudo translation and machine translation. Some well-known Gen 4 tools include Lokalise, Localazy, and Phrase (which recently joined forces with Memsource).
As localization professionals, you'll most likely work with CAT tools, ideally Gen 4 tools, as they're better suited for app and software localization.
Now that we've covered the different generations of CAT tools, let's dive deeper into localization platforms and explore their features. If you're an experienced translator or localizer, some of this may already be familiar.
In the previous chapter, we learned that there are three key things you need to know to localize an interface effectively:
1. User and brand information
2. Product and business goals
3. Scenario and flow
These are the basics you'll want to keep in mind throughout the project. However, there's more to it than just that. When you get down to the nitty-gritty, you'll also need:
4. specific information on words and sentences
5. Terminology
Also, a few weeks back, we touched on the importance of:
6. Communicating with product teams.
These are all critical things that are needed to ensure we get a smooth process as we craft the perfect UX copy.
Now, the perfect localization tool may not exist yet, but some Gen 4 platforms are coming close. Let's take a look at an example.
CAT tool example #1: Lokalise.
Here, translators have access to all languages, not just their own. This can be incredibly helpful when facing challenges in translation or when you need to cross-reference with other languages (1).
Lokalise also offers a glossary (2), screenshots for context (3), and unique key names for each line of copy (4). These key names can reveal crucial information that can assist you in creating accurate translations. Additionally, the platform allows for different versions of strings to accommodate language variations (5).
Most Gen 4 CAT tools have similar features, though they may look different. Let's see another example.
CAT tool example #2: Crowdin
Crowdin provides many of the same capabilities as Lokalise, such as source strings (1), translations (2), character limits (3), and string-level discussions (4). You have access to a list of the strings in the file for context (5), get TM suggestions (6) and search previous versions, too (7).
These localization platforms offer a lot of useful information, like string-specific details, terminology, and some context about the scenario and flow. However, they don't provide everything you need. You'll still need to learn about user information and goals, as well as more context for the scenario and flow.
Can you choose your own CAT tool?
Depending on your role in localization, you may or may not be the one choosing the platform to use. Regardless, it's essential to be familiar with the different tools available. This knowledge can help you assist the product team, stand out, and even land new projects if you're freelancing.
If you're not into experimenting with apps, you can join the Localization Advocate's loc tools study group, which runs free demo events twice a month with reps from various localization tools. It's an excellent opportunity to learn about the available options and pick up some handy tips for using them. Just follow the link in the slide to find out more and join if you're interested.
3. UX design tools
These tools might not be as familiar to the localization community, since localizers didn't need access to them until recently. Originally, these tools were designed for product design, helping designers create wireframes and prototypes for apps and software.
Figma is probably the most well-known and widely used design tool today, but there are others like Adobe XD and Sketch. These tools are used to design the user experience and create mockups before moving on to development. Often, localization occurs at this stage, even before developers start working on the feature.
Sometimes, you'll come across tools like Invision and Zeplin, which are specifically created for collaboration in design teams. They serve as a central storage place for up-to-date designs, acting as a source of truth for teams to refer back to.
Now, you might be wondering why I'm talking about UX design tools when you're not designers. Well, these tools can be quite useful for localization purposes. Teams often use design tool links as a visual reference, which can be helpful when localizing specific pieces of copy.
Additionally, some teams even do the localization directly in the design tool. While this provides full access to the user flow, it doesn't offer other essential localization features like translation memories or glossaries. So, while it might work for smaller products, it's not ideal for scaling up.
By accessing and using design tools wisely, you can complement the information you get from CAT tools, providing a fuller picture of the product or feature you're localizing. However, this means you'll need to familiarize yourself with design tools and their features.
We won't dive deep into Figma or Zeplin training right now, but we will cover some of their main features so you know what's available. If you need more guidance, don't worry; there are plenty of help centers and YouTube videos that show you exactly how to use these apps. Just remember that exploring UX design and writing tools can open up new possibilities for your localization process, so it's definitely worth the effort.
Design tools meant for prototyping
Imagine a file in one of these tools as a giant, infinite canvas filled with various items. You can scroll in any direction and add as many items or screens as you want.
Plus, you can easily zoom in and out to see specific screens or the entire flow. This is super helpful for getting visual context and understanding the user journey.
Figma files, for example, can be built in different ways depending on each company's preferences.
There's a pane on the left with two sections: Pages on top (1) and design pieces at the bottom (2). You can click through Pages to reach different parts of the design, but most of the time, the content you need to localize will be on a specific page.
The bottom section (2) displays all the design pieces. Frames contain smaller pieces inside, and within each frame, you'll find screens with various elements. For instance, a screen might include text boxes, images, shapes, and icons. Developers can then turn these elements into working code.
If you have editing access to the tool, you can use the type tool (1) to edit the copy and enter your localized version. Working directly in the design can be useful for testing, even if you eventually input the actual copy elsewhere.
Now, let's talk about communication. clicking the text bubble icon (2) opens the comment sidebar, allowing you to add comments and view all comments on the document.
You can sort comments by date, show or hide resolved comments, and filter to see only the ones relevant to you. When you click on a comment, you can read it, reply, and resolve it.
Design tools meant for collaboration
These types of tools were made for collaboration with non-design professionals, like localizers and developers. InVision and Zeplin are two great examples. The industry is always changing, and the lines between these tools and others might blur, but understanding their basic features is still important.
Instead of an infinite canvas like in Figma, these tools present screens in a gallery view, organized like files on your computer, with folders based on projects and categories. This enables easy commenting, discussion, and storage of different screen versions, keeping everything neat and organized.
InVision and Zeplin have similar interfaces with different colors and aesthetics, but the main principles are the same.
Screens can also be organized in a flow, giving you full context of what users see before and after.
This is super helpful for localizers to understand the user journey.
Adding comments in these tools is similar to Figma. Just put your comments directly on the screen, and you're good to go.
What can we do with design tools in localization?
There are three main ways you can use design tools in localization:
Use them as visual context, to understand the flow, layout, and how everything looks.
Localize directly inside the tool by replacing the source copy with your translation.
Use comments to communicate with the product team, ask questions and provide feedback.
Unfortunately, options 2 and 3 are not ideal for any localization flow. Let's see why.
Localizing directly inside a design tool
When you localize in the design tool, your translated copy isn't linked to the source copy or other localized versions. This means once you overwrite the source, it's gone, and accessing other translations becomes a hassle. Editing each text box separately can also be time-consuming, and you won't have access to translation memories, glossaries, or other helpful resources.
Using design tool comments to communicate with the product team
When you work this way, questions end up being linked to screens, not the copy. This makes it hard to track questions, especially if screens are deleted or changed. Plus, other localizers might not be able to find your questions easily.
However, there are some advantages to using design tools for communication with the product team. Namely, you can expect faster replies and better visual context for your discussions.
So, should you use design tools for comments and discussions? It depends on the alternatives available. If you're stuck with a spreadsheet or an old-fashioned CAT tool, you might opt for the design tool. But if there's a new-generation CAT tool with built-in context, that's obviously the better choice.
In the end, it's all about finding the right balance and using the right tools for the job. As a localizer, it's important to be adaptable and make the most of the resources at your disposal.
4. UX writing tools
While design tools treat copy as an afterthought and localization tools see it as a source to work from, UX writing tools put copy center stage. This makes them perfect for crafting microcopy.
As you know, creating UX copy is no simple task. It requires focus, collaboration, and a deep understanding of the user experience.
Traditional CAT tools display copy amidst a sea of other data: translation memory results, machine translation results, glossaries, versions, and more. This clutter can distract from the crucial aspects of the user experience, such as the UI, layout, and flow.
UX writing tools, on the other hand, prioritize the UX copy. They display it in two ways:
1. As visuals, similar to prototyping or design management tools, with screens shown side-by-side for easy editing.
2. As texts, grouped by screens or displayed as a list of all text in the app.
Take these breadcrumbs, for example. These links indicate the user's location within a hierarchy. With UX writing tools, you can navigate and edit copy while viewing its context within the app.
Frontitude's game-changing feature is the live editing of copy, which you can see immediately reflected in the visual. You can also edit directly in Figma, with changes syncing to the UX writing tool. This means full context without compromising design files, and your team can track all changes.
Localization is supported too. Once the product team sets a language, you can add copy for each string in every available language. You'll also have access to activity logs, comments, and discussions, providing the maximum possible context.
Both Ditto and Frontitude offer activity logs, tracking the change history of a string, and facilitate communication with the product team through comments and discussions.
Product teams can add guidelines for individual strings, so you don't have to trawl through massive style guides – just focus on the relevant guidance. You can search the entire copy library, making it easy to find and maintain consistency.
And UX writing tools also let you control text components. These are set patterns, like "next" or "continue" buttons, that are always used consistently.
Sometimes, teams manage both mobile and desktop versions and want specific copy to be identical across platforms. Components ensure this consistency, syncing across all instances.
Do UX writing tools check all the boxes?
Let's circle back to the ideal localization tool we described before, and see how UX writing tools measure up. Here's what they offer:
String-specific info: You can add instructions as needed in the comment pane.
Communication: History and comments facilitate easy collaboration.
Scenario and flow: You can view screens in order to understand the user journey.
Terminology: Though not a term base, searching through live copy in use is even better. It avoids outdated translation memories and terminology lists that haven't been updated.
However, UX writing tools don't yet provide user and brand information, or additional information about the product and its goals. For now, this information comes separately in the form of a localization brief. It would be great if these tools could incorporate such crucial details in the future.
Personally, I'm a big fan of UX writing tools. They offer a fresh perspective on UX copywriting and are well-suited for UX writers and localizers. While they may take a few more years to fully mature, I believe they'll eventually become a solid alternative to fourth-generation localization tools. These tools are evolving to better accommodate the unique needs of UX writers and localizers, making the localization process more enjoyable and efficient. So keep an eye on them, as they could revolutionize the industry.
Getting future-ready
As we wrap up this discussion, let's take a moment to explore the AI revolution and its impact on our industry. You've probably heard of chatGPT, Midjourney, and other similar tools that have recently emerged. Google even launched MusicLM, which creates music from a prompt! It's easy to feel overwhelmed by the rapid evolution of technology, but let's consider how we can leverage AI to power up our work. Don't be afraid to experiment. Most tools offer free tiers or trial plans, so give them a try.
For now, humans use AI tools through prompts, which serve as the user interface for AI. Prompts can be instructions, descriptions, or questions, and the type you choose depends on your goal and the tool you're using. Start with a simple prompt and refine it as needed.
Consider these questions to guide your AI use:
What parts of my work don't I enjoy?
What tasks take too long?
Identifying these areas can help you determine how AI might streamline or improve your work. For example:
1. Email writing
Localizers send countless repetitive emails. AI can help draft those messages, whether through a tool like chatGPT or an email client with built-in AI, like Canary mail.
2. Self-branding and marketing
AI can also assist with self-branding and marketing, generating copy and visuals for LinkedIn posts, blog articles, and more. Use AI to edit and proof your writing, with tools like Wortune that offer suggestions for alternative phrases and other helpful features.
3. Learning and self-development
AI can help you catch up on industry content by reading articles out loud or summarizing long articles into key points. Some tools, like Wordtune Read or Upword, make this process seamless.
4. Quality assurance
In the localization sphere, AI can be used for QA. AI-powered QA tools help ensure the quality and accuracy of translations. By using advanced algorithms and natural language processing, these tools can detect grammatical errors, inconsistencies, and even cultural nuances. This makes the entire QA process more efficient, as AI learns from previous translations and minimizes false positives, allowing localizers to focus on fixing real issues.
5. Term extraction
AI can also streamline the process of extracting terms for terminology lists. Traditional methods work based on statistical analysis, which means that they flag words as "terms" based on their frequency. With AI, term extraction is more context-driven, so it has a better understanding of the content - and you get more accurate terminology lists.
6. Ideas for localized UX copy
AI can be a valuable resource for generating creative and culturally appropriate ideas for localized UX copy. Machine translation serves as a starting point or a draft, giving localizers insight into potential translations in their target language.
Depending on the language, AI can provide ideas for UX copy without needing the source content. Some Figma plugins like Ghost UX writer even allow you to automatically generate UX copy. For example, Wordtune can generate alternative phrases for a given piece of text, while chatGPT can provide ideas for specific UX elements, like "add to bag" buttons for eCommerce sites.
These ideas are just the beginning. Embrace the AI revolution and explore the tools that can enhance your workflow. And if you discover more, feel free to share them with the community!
Tools & software
Chapter 8
