Elearning Localization: How to Take a Course into New Languages Without Rebuilding It
How do you localize an e-learning course into several languages without rebuilding it? You work from the source project, not from screenshots: export the course text, translate it with your glossary and translation memory, import it back, adapt the layout for each language and republish. That is elearning localization in practice, and it is how DEMA Solutions handles training courses.
Which training owners suddenly need their courses in more languages?
This guide is written for the people who own training content at a manufacturer and have just been told it needs to exist in many more languages. You may run the dealer academy, the onboarding course for service technicians and the annual safety refresher. You may sit in L&D and have built the courses yourself. Or you may work in technical documentation and notice that the course uses the same component names, warnings and procedures as the manuals you already manage.
All three roles meet the same problem. A course is not a document. It is a small piece of software made of slides, layers, triggers, quiz logic, media and a player that wraps it all together. Translating the words is only part of the job. The other part is making sure the course still works, still fits on the screen and still teaches the right thing once every word has changed.
Translation vs localization for e-learning
Translation turns the text of a course into another language. Localization makes the whole course work for a learner in another country. The difference sounds academic until a German technician clicks a button labelled in English, or a Spanish dealer is asked a quiz question about a measurement in inches that nobody uses in that market.
For a typical product or service course at a manufacturer, the parts that need attention go well beyond the narration script and the slide text:
- on-screen text, including titles, body text, callouts, hotspot labels and text that only appears on hover or in a hidden layer;
- images and diagrams that contain words, such as annotated photos of a machine, flowcharts or screenshots of the control panel;
- buttons, navigation and player labels, the “Next”, “Submit”, “Menu” and “Resources” texts learners click on every slide;
- quizzes, answer options, feedback messages and the results slide, including the logic that decides which answer is correct;
- units, number formats, dates and examples, which should match what the learner meets in the field;
- safety content and references to regulations, which must say exactly what the source says and may need checking against local rules by your own experts.
Two of these deserve a closer look. Screenshots of a machine interface are easy to overlook because they are pictures, yet they are often the most important visual in a procedure. If the machine itself shows its screens in the learner’s language, the course should show the same screen in the same language, with the same labels. Otherwise learners have to translate in their heads while standing at the machine, which is exactly what training is meant to prevent.
Safety training is the other. A course on lockout procedures or on handling a pressurised circuit carries warnings that should match the wording in the instructions supplied with the product, word for word where possible. Localization does not change the safety content. It keeps the meaning intact and flags anything that may not apply in the target country, such as a reference to a national standard, so that your subject experts can decide what to do with it.
What to send: the source project, not screenshots or PDFs
The most common slowdown in an elearning localization project happens before any translation starts: the supplier receives a PDF of the slides, a set of screenshots or a link to the published course. None of these can be turned back into a working course. Someone would have to retype the text and rebuild every slide by hand, which costs more, takes longer and introduces errors.
What is needed instead is the course as your authoring tool stores it. In Articulate Storyline that means the .story project file, which holds the slides, text, triggers, animations, interactions and embedded media. Articulate Rise and Adobe Captivate have their own source formats and export routes, and the principle is the same: whoever localizes the course needs the source, not the output.
Alongside the project, most authoring tools can export the translatable text as XLIFF, the cleanest route because it keeps every string linked to its place in the course. A Word or Excel text export also works, as long as it is re-imported with the same tool and version that produced it. It helps to add the published HTML5 or ZIP package, so the localization team can see how the course behaves, and the media assets on their own: images with editable text layers, original video files and any existing subtitle files.
Two habits save a lot of time. Freeze the source before you send it, because changes made during translation have to be repeated in every language. And say which version of the authoring tool you use, since projects saved in a newer version may not open in an older one.
How the elearning localization process works, from export to republished course
The round trip has five stages, and knowing them makes it easier to plan who does what on your side.
It begins with the export. Either your team or the localization team exports the translatable text from the project, usually as XLIFF. The export is checked first: are hidden layers included, are variables and tags protected, is any text trapped in images? The content is also analysed against your translation memory to see how much is new.
Next comes translation. Linguists translate the exported text with your approved glossary and your translation memory, so that a component, a fault code or a safety instruction is named in the course exactly as it is named in the documentation. Quiz questions and their feedback are translated together, so the logic still holds: if the correct answer is “the pressure relief valve”, the feedback must refer to the same valve under the same name.
Then the translated text goes back into the project. The file is imported into a copy of the source course for each language. This is a technical step, and it is where a clean export pays off: if tags and variables were protected, the import places every translated string in its correct slide, layer and state.
The fourth stage is the one that is most often underestimated: the layout check. Every slide in every language is opened and reviewed for overflowing buttons, callouts covering diagrams and quiz options that wrap badly. The next section covers this in detail.
Finally, each language version is republished in the format your LMS expects, whether that is a SCORM package, another e-learning standard or plain HTML5, and tested in the player before it reaches the learners. If your subject experts in each country want to review the course, this is the moment to involve them, with the course running as learners will see it rather than as a table of strings.
Layout and text expansion: where courses succeed or fail
A slide designed in English or Italian is built around the length of its words. Buttons are sized for “Next” and “Back”, callouts are drawn to fit a short instruction, and the designer has balanced text and image on every screen. Translation changes all of that at once.
German is the classic example: compound nouns for technical parts can be much longer than their English counterparts, and sentences grow with them. French and Spanish also tend to run longer than English. Other languages run shorter, so a slide that looked full now looks empty. Every language has its own shape, and the course has to be adapted to each one.
The fixes are design decisions. Can the button grow, or should the label be shortened in agreement with the client? Should the font size drop slightly, or should the text move to a second layer? Does the callout need to move so it no longer covers the part it describes? These questions come up on almost every slide, and they need someone who works in the authoring tool every day.
Scripts add another layer. Courses for Eastern Europe, Asia or the Middle East may need fonts that support Cyrillic, Greek, Chinese or Arabic characters, and a font that looks perfect in English may simply lack them, so the tool silently substitutes another. Right-to-left languages such as Arabic and Hebrew reverse the reading direction of the whole slide, which affects navigation, the placement of images and progress indicators.
Interactions are the last risk. Drag-and-drop targets sized for short words, hotspots over labels that have moved, animations timed to the source text and triggers that depend on a text variable can all break quietly, until a learner in another country reports that the exercise cannot be completed.
This is exactly the work done by Foliage Solutions, DEMA’s brand specialised in multilingual DTP. The Foliage team has handled e-learning projects for its clients in Articulate Storyline, Articulate Rise and Adobe Captivate, working with the .story project, XLIFF and text exports, published packages and media assets. Their job on a course is the one described above: adapting the layout for every language, handling expansion and shrinkage, setting up fonts for non-Latin and right-to-left scripts, syncing subtitles and animations, checking the course visually across devices, and doing a final pass on readability and course flow. You can find out more about how Foliage approaches DTP for e-learning projects, and the same in-house expertise sits behind our desktop publishing services for manuals and printed training material.
Training videos and subtitles
Most technical courses include video: a technician replacing a filter, an installer wiring a control board, a product manager presenting a new range. These clips often carry the most practical knowledge in the course, and they need a plan of their own.
For manufacturers, subtitles are usually the most practical route. The original video stays as it is, the technician’s real voice and the real sound of the machine remain, and each language gets a subtitle track that can be updated cheaply when the video changes. Subtitles also suit workshops and shop floors, where sound is often off anyway. Voice-over and dubbing are a separate kind of production that we do not offer, and for most technical training they are not needed.
Reading speed matters more than in entertainment video, because the learner has to read and watch a technical action at once. Each subtitle should stay long enough to be read comfortably and never cover the part of the image where the action happens. Terminology must match the rest of the course, so the subtitles are translated with the same glossary as the slides.
On-screen text inside the video is the detail that most often gets missed. A part number shown in a caption, a warning burned into the image or a label on a screen recording will not be covered by a subtitle track. Decide early whether those elements will be re-edited in each language or explained in the subtitles. Our subtitling services cover training and product video, and the article on what changes the price of video subtitling explains the factors that affect time and cost for the video part of a course.
Keeping training consistent with manuals and the machine interface
Training and documentation often sit in different departments with different suppliers, but the learner does not see that division. A technician finishes the course, opens the service manual on the next job and then reads the control panel. If the three sources call the same component by three different names, the course has done part of its job and undone the rest.
The answer is to run course localization on the same language resources as your technical documentation. The glossary, or termbase, records the approved name of every component, function and warning in each language, and it should be the same file for the course, the manual and the interface. Our page on terminology management and glossaries explains how such a termbase is built and maintained. The translation memory stores every approved sentence, so a safety warning that already exists in the manual is reused in the course in exactly the same words.
This is especially valuable for safety training and for courses that walk through operating procedures, because these repeat content from the instructions supplied with the machine. The requirements for those instructions are becoming stricter, as our guide on user manual translation and EU rules describes, and a course that uses different wording from the official instructions creates a risk nobody needs. The same logic applies to the control panel: if your screens are translated, as discussed in our article on HMI and machine software translation, the course should quote those screen texts exactly as the machine displays them.
Updating courses without paying twice
A course is never finished. A new model joins the range, a procedure is simplified after feedback from the field, or the safety section is revised, and each change has to reach every language.
If the first round was done from the source project with a translation memory, updates are straightforward. The new version is exported, compared with the memory, and unchanged sentences are reused automatically. Translators then work only on what is new or modified, and the layout team checks only the slides that changed. If the first round was done by copying text out of a PDF and rebuilding slides by hand, the update often means starting again.
Two habits make updates cheaper. Keep a clear record of which course version was localized and when, so that the next export can be compared with the right baseline. And keep the translation memory in a place where it belongs to you and can be exported, so that the work you have paid for is never locked in with one supplier. How a memory is built from earlier translations, including ones done elsewhere, is covered on our page on translation memory and alignment.
What drives cost and time
No two courses cost the same to localize, but the factors behind cost and time are predictable.
The volume of text is the starting point: how many slides, how many words on them and in the hidden layers, how many quiz items and feedback messages. Repetition reduces it, both within the course and against your existing translation memory. Interactions add work, because each drag-and-drop exercise, branching scenario or variable-based trigger has to be checked and sometimes rebuilt in each language. Media adds work too: images with embedded text need editing, and every minute of video needs transcription, translation and subtitle timing.
The number of target languages multiplies all of this, and the choice of languages matters as much as the count. Languages that expand strongly, or that need new fonts or a right-to-left layout, take more layout time per slide. A cleanly built source, with consistent styles and text kept out of images, also localizes far faster.
Finally, plan for review on your side. Your subject experts in each country, often a regional trainer or a senior technician, will want to check the course before it goes live. Their availability often sets the real deadline, so agree early who reviews each language and when.
Questions to ask an e-learning localization supplier
When you compare e learning localization services, start by asking which authoring tools the supplier works in directly and which file formats they can take. A team that can open your Storyline, Rise or Captivate project and work inside it will catch problems that a team working only on exported text cannot see. Ask whether layout adaptation is done in house or passed on to someone else, because that is where most of the quality and most of the delays are decided.
Next, ask how they will keep the course consistent with your manuals and interface texts. You should hear about a shared glossary and translation memory, and about who owns them. Then ask how they test the localized course before delivery: in the authoring tool only, or also in the published package, in the player and on the devices your learners use.
Ask how video is handled, including on-screen text inside the clips, and how updates will be managed: will the supplier show you which text is new and which is reused? Finally, ask how your own subject experts can review the course in each language, and how their comments flow back into the glossary so they do not have to make the same correction twice. Our broader guide on choosing a technical translation company adds questions on quality and process that apply here too.
How DEMA Solutions handles elearning localization
For DEMA Solutions, a course is one more channel for the product information we already handle for manufacturers as a provider of language technology solutions. We work from the files your authoring tool exports, return them ready to import and republish, and add subtitles to the training and product videos inside the course. Voice-over and dubbing are not part of what we do. What is included is set out on our page for e-learning and training translation.
Course layout stays in house. Foliage Solutions, our brand for multilingual DTP, has worked on Storyline, Rise and Captivate projects, so text expansion, fonts and the timing of subtitles and animations are sorted out before you see the result, rather than left to your L&D team. Translation follows our ISO 17100 certified process, in which a second qualified linguist revises every text, and the course draws on the same client-owned memory and termbase as your manuals and interface, delivered in TMX and TBX whenever you want them. With more than 20 years in technical documentation behind us, we give a safety refresher the same care as the manual it summarises.
Where TUBE fits
Courses rarely stay finished: a new machine version changes three slides, a revised procedure changes a quiz. TUBE handles exported course text in any of 91 file formats and reuses the sentences you approved for the last release. A leverage report shows how much of each course update came from reuse, and nothing has to be installed on your side.
Frequently asked questions
Work from the source project in the authoring tool. Export the translatable text, usually as XLIFF, translate it with your glossary and translation memory, import it back into a copy of the course for each language, adapt the layout slide by slide and republish in the format your LMS expects. The course structure, interactions and media stay in place.
Yes. Courses built in Articulate Storyline, Articulate Rise and Adobe Captivate can be localized from their source projects and text exports, with layout adapted for each language. The most useful starting point is the project file itself, together with the published course so the localization team can see how it behaves.
No. DEMA provides subtitles for training and product videos, translated with the same terminology as the course and timed for comfortable reading. Voice-over and dubbing are not part of our services.
A PDF shows the course but cannot be turned back into it. Text has to be retyped into every slide, hidden layers and quiz feedback are often missing, and interactions are lost. The source project avoids this and makes updates cheaper.
Languages that run much longer than the source, such as German, need the most adjustment. Non-Latin scripts need suitable fonts, and right-to-left languages need the whole slide mirrored.
Use one glossary and one translation memory for training and technical documentation. When the course is translated from the same approved resources as the manuals and the machine interface, learners meet the same name for each component and the same wording for each warning everywhere.
The new version is exported and compared with the translation memory. Unchanged text is reused, translators work only on new or modified content, and the layout check concentrates on the slides that changed.
Send us one course or module
The quickest way to see how your training would localize is to try it on a real course. Send us one course or module with its source file and your target languages, together with any glossary or earlier translations you already have, and we will show you what can be reused, where the layout will need work and how the videos can be subtitled.
How can we help?
Tell us what you need and we’ll get back to you with the solutions.