How to Localize a SaaS Onboarding Video into 5 Languages: Process, Tools, Mistakes, and Real Costs

I almost made a very expensive mistake.

We had just finished our onboarding video. The one that cut time-to-value from 14 days to 4. The one that took three months and too many arguments to get right. And now our head of growth was in my office saying the words every SEO dreads: “We need this in five languages by Q3.”

My first instinct was to treat localization as a simple translation job. Take the script. Send it to a translator. Get back five versions. Record five voiceovers. Done. I had a budget in my head. It was small. I was proud of it.

Then I actually looked at what we were asking. The video showed our interface. The interface was in English. We were going to dub a calm, reassuring voiceover in German, over a screen full of English buttons. That’s not localization. That’s a lie in a different language.

This is the article about everything that happened next. The five languages. The tools. The cultural disasters. The line items nobody budgets for. And the moment we realized that localizing a SaaS onboarding video is a completely different job from producing one—and that most teams, including us, get it wrong because they think translation is the hard part. It isn’t. The hard part is everything else.

We did this work in partnership with LeoStudio, whose approach to clarity you can explore at leostudiohq.com. They had warned us about this. We didn’t listen. This is what we learned when we finally did.

The Strategic Decision Nobody Talks About: What Exactly Are You Localizing?

Before we spent a dollar, we had to answer a question that sounds simple but isn’t. When we say “localize the onboarding video,” what are we actually localizing?

There are three layers inside every screen-heavy onboarding video. The narration—what the voice says. The subtitles—what appears at the bottom of the screen. And the product screen itself—the interface, the buttons, the menus, the tooltips, the on-screen text.

Traditional video localization handles the first two. It dubs the voice and translates the captions. It ignores the third. And for a SaaS onboarding video, the third layer is the one that matters most. A user can hear a flawless Japanese voiceover explaining “click the add data source button” while staring at an English button that says “Add Data Source.” They’re being told what to do in their language and shown how to do it in another. That gap is where onboarding breaks.

Research on this is clear. Dubbed narration over an untranslated interface tells the viewer the product isn’t really available in their language, no matter what the audio says. The screen is the truth. The voiceover is just narration.

We had to decide which layers we were willing to localize. And that decision shaped everything—the budget, the timeline, the tools, the mistakes.

The Three Levels of SaaS Onboarding Video Localization

There’s a cost model that most teams discover too late. It has three tiers.

Level 1: Subtitles only. The cheapest and fastest. You translate the script, generate subtitle files, and overlay them. The video works for comprehension. It doesn’t work for immersion. It’s fine for early market testing, where you just want to know if the market exists.

Level 2: Dubbed narration. You translate the script, hire voice talent or use AI dubbing, and replace the audio track. This is where most B2B SaaS demos live. Comprehension improves. Cognitive load drops. Viewers can focus on the screen instead of reading. But if the screen is still in English, you’ve only solved half the problem.

Level 3: Localized screen capture. You re-record the screen capture in the target language. The interface matches the narration. The menus match the buttons. The user sees a product that actually exists in their world. This is the only level that delivers a complete experience. It’s also the most expensive, because it requires your product itself to be localized first.

We started at Level 2 for four languages and Level 3 for one. That one was German. That decision almost killed the project.

The Process: How We Actually Localized Five Languages

Here’s the sequence we followed. I’m sharing it in order because the order matters. Doing these steps out of sequence is how you waste money.

Step 1: Lock the Master Script (And Lock It Again)

Before we translated anything, we froze the English script. Completely. No changes. Not even a comma. This is the step everyone skips, and it’s the step that costs the most when you skip it. Starting dubbing before finalizing the time-coded script leads to costly re-renders and re-recordings, especially when lip sync is involved. If a single word changes after you’ve recorded five voiceovers, you’re re-recording five voiceovers.

We created what our localization partner called a “Golden Document”—a version of the script with timecodes, on-screen text references, and UI element callouts. Every translation was built from that document. Not from the English script alone. From the document that mapped every word to a moment.

Step 2: Translate for Meaning, Not for Words

This is where we made our first real mistake. We hired a translation agency that specialized in legal and medical documents. They were precise. They were accurate. They were completely wrong for us.

Their translations were technically correct and emotionally dead. Our English script opened with “You signed up to get answers. But right now, you’re staring at an empty screen.” Their German version translated that literally. It was grammatically perfect. It also sounded like a manual. It didn’t sound like a human acknowledging frustration.

We had to redo it. The fix was working with translators who understood SaaS, not just language. People who knew what “time-to-value” meant without looking it up. People who could translate the feeling of the line, not just the words.

The best practice here is simple: translate for meaning, not for words. If the English line is idiomatic, rewrite it in the target language to carry the same emotional weight. Don’t try to preserve the idiom. Preserve the effect.

Step 3: Anticipate Text Expansion (It Will Break Your Layout)

Here’s something nobody tells you until it’s too late: translated text is almost always longer than English. German can expand by 30%. Spanish by 20%. Finnish by even more. When your on-screen text callouts are designed to fit English, they overflow in every other language.

We had to redesign every on-screen overlay for every language. The font sizes changed. The line breaks changed. The timing changed. It was a week of work we hadn’t budgeted for.

The lesson: design your source video for expansion from the beginning. Leave room. Use fewer words. If you can remove on-screen text entirely and rely on the screen recording to carry the meaning, do it. It makes localization cheaper and cleaner.

Step 4: Localize the Screen Itself (For the Language That Deserves It)

This is where the German version broke us.

Our German market was our largest non-English market. It deserved Level 3. That meant the screen capture had to be re-recorded in German. Which meant our product had to be localized in German first. Which meant we had to coordinate with the product team, wait for the German UI to ship, re-record the screen capture with German menus and German buttons, and then sync it to the German voiceover.

That single language took six weeks. The other four took two weeks combined.

The lesson is brutal but necessary: don’t promise Level 3 localization unless your product is already localized in that language. Re-recording the screen for a still-English product is effort spent hiding the gap, not closing it.

Step 5: Voiceover, Captions, and Quality Control

We used human voiceover for German, French, and Spanish. We used AI dubbing for Portuguese and Japanese. The difference was noticeable but not disqualifying. Human voiceover carries warmth and trust. AI voiceover is getting better every quarter and costs a fraction of the price.

We also generated captions for every language, even the dubbed ones. Why? Because most viewers watch on mute. A dubbed video without captions is still inaccessible to the silent majority.

Then came quality control. This is where we almost shipped a cultural disaster.

The Mistakes That Still Make Me Cringe

I want to be specific here, because vague warnings about “cultural sensitivity” don’t help anyone.

Mistake 1: The Gesture That Meant Something Else

Our English video had a moment where the cursor hovered over a button, then moved to the next step. Simple. Universal. Except it wasn’t. In the Japanese version, the pacing felt aggressive. The cursor moved too fast. The pause we built for emphasis felt like impatience. Japanese onboarding content tends to be slower, more deliberate, more respectful of the user’s attention. We had to re-edit the Japanese screen recording to slow the cursor and add pauses.

This is not something a translation file catches. It’s a cultural pacing issue. It required an in-country reviewer to catch it, and we didn’t have one for the first version.

Mistake 2: The Color That Meant the Wrong Thing

Our video used a green highlight to indicate the next clickable button. Green means “go” in most Western markets. In some contexts, it also means something else. Our Portuguese reviewer flagged it. It wasn’t offensive. It was just slightly off. We changed the highlight color for the Portuguese version. It took ten minutes. It would have taken months to recover from a market that felt the product wasn’t for them.

Mistake 3: The Humor That Didn’t Travel

Our English script had a light, almost playful tone in one line: “Don’t worry, this part is easier than it looks.” It tested well in English. It flopped in German. The German translation was accurate but the tone was wrong. What read as reassurance in English read as condescension in German. We rewrote it. The new line was flatter. Less charming. More direct. It worked better.

Mistake 4: The Metadata We Forgot

We localized the video. We forgot to localize the file names, the descriptions, the tags, and the language metadata. For two weeks, our Portuguese users were being served a video tagged as English. The content was Portuguese. The discovery system didn’t know that. This is the unglamorous side of localization that nobody talks about, and it’s the one that quietly kills your reach.

The Real Costs: What We Actually Paid

I’m going to give you the numbers, because I wish someone had given them to me.

We localized a 90-second onboarding video into five languages: German, French, Spanish, Portuguese, and Japanese.

Subtitles only (Portuguese and Japanese): $150 per language. The translation and subtitle files. Turnaround: three days per language. This was the cheapest line item and it delivered exactly what we needed for markets we were still testing.

AI dubbing (Portuguese and Japanese): $80 per language. We layered AI voiceover on top of the subtitles for these two languages. Total per language: $230. The AI voice was good enough for onboarding, where the goal is comprehension, not cinematic performance.

Human voiceover (German, French, Spanish): $400 per language. Professional voice talent, recorded in-studio or remote, synced to the existing video. The quality difference was real. The German voice sounded like a person. The Portuguese AI voice sounded like a very good AI.

Screen re-recording (German only): $1,200. This included coordinating with the product team, waiting for the German UI to ship, re-recording the screen capture, re-editing the timing, and re-syncing the audio. It was the single most expensive line item in the entire project. It was also the only way to make the German version feel native.

Cultural review and in-country QA: $600 across all five languages. We hired native speakers who worked in SaaS to watch the localized videos and flag anything that felt off. They caught the pacing issue in Japanese, the color issue in Portuguese, and the tone issue in German. This was the best money we spent. Without it, we would have shipped versions that worked technically and failed emotionally.

Total for five languages: Roughly $3,800. That’s the honest number. It’s less than a full re-production of the video. It’s more than I had originally budgeted. And it’s the reason our activation rate in German markets matched our English activation rate within six weeks.

For context, traditional studio dubbing runs $100 to $500 per minute per language, and takes one to two weeks per language. AI dubbing can run as low as $30 to $150 for the entire video, depending on the tool. We landed in the middle because we mixed approaches—human for the markets that mattered most, AI for the markets we were still testing.

The Framework We Now Use Before Every Localization Project

If you’re planning saas onboarding video localization, here’s the checklist we run before we spend a dollar.

First: Which markets deserve Level 3? Only localize the screen if your product is already localized in that language, and only if that market generates enough revenue to justify the cost. German deserved it. Portuguese didn’t. That decision saved us thousands.

Second: Who is translating? Hire translators who understand SaaS, not just language. A legal translator will produce a legally accurate script that sounds like a robot wrote it. A SaaS translator will produce a script that sounds like a human wrote it for humans.

Third: Who is reviewing? You need in-country reviewers. Not translators. Reviewers. People who live in the market, use the product category, and can tell you when something feels wrong. This is not optional. This is the difference between localization and translation.

Fourth: Have you designed for expansion? If your source video is packed with on-screen text, you’ve made localization three times harder. Strip it down. Leave pauses. Say what a button does, not where it is. “Click the button in the top right” will break when the layout changes in German.

Fifth: What’s your maintenance plan? Your product ships updates. Your localization files will go stale. Build a refresh budget into the original project. Treat localization as ongoing maintenance, not a one-time task.

Why This Matters Beyond the Metrics

I’m an SEO. I care about activation rates and time-to-value. But there’s something else here that doesn’t show up in a dashboard.

When a user signs up for your product in their language and sees a video that speaks to them in their language, over an interface that’s also in their language, something shifts. They stop feeling like a guest. They start feeling like a customer. That feeling is hard to measure. It’s also the thing that makes them stay.

We localized our onboarding video into five languages. It cost $3,800. It took eight weeks. It almost broke our timeline. And it was the single best decision we made for international growth that quarter.

If you’re thinking about saas onboarding video localization, start with the strategy, not the tools. Decide which layers you’re localizing. Decide which markets deserve the full treatment. Hire reviewers, not just translators. And before you spend anything, go to leostudiohq.com and read how they think about clarity across languages and contexts. Their approach to onboarding video production is the reason we even understood what we were localizing in the first place.

Tell them the SEO guy sent you. Tell them you know the difference between translation and localization. They’ll know exactly what you mean.