How Long Will It Take to Localize My Game? – Localization Scheduling
Localization| 13 Aug 2022
How Long Will It Take to Localize My Game? – Localization Scheduling
By Jennifer O’Donnell, proofread and checked by Guido Di Carlo, Alyssa Wejebe, and Margaux Lanfranchi.
I want to start with a caveat that every project is different which means how long it takes will vary. As such, everything from here on out should be taken with a grain of salt and a heavy load of critical thinking for how things might impact your game’s schedule.
I highly suggest you do not plan your release date and then work backwards. Why? Because almost everyone underestimates how long localization takes and this often (always) leads to a very tight and expensive localizations.
Planning a game localization takes time and it’s important to schedule properly and, if possible, consult your localization team before you start. But here are some things to consider when planning your schedule!
Scheduling Loc – a Give-and-Take of Time / Money / Quality
A quick and cheap localization at high quality is impossible. You might think it’s doable, but it blows up in developers’ faces more often than not.
If you want high quality, then you have to spend either money or time on it—or both! If you want it to be quick, then you often have to sacrifice quality or spend more money on it—or both!
What you prioritize will impact your localization schedule, and therefore your release schedule. So, think about it carefully!
Before You Begin
Before you begin, you need to work out how big your game is, which means the total volume of words present in-game and how long it takes to play.
You will also need to consider an estimate for additional translations such as important development documents, time for familiarization and glossary creation, asset translation and creation, voice recording, store text, marketing materials, website text, and legal documents.
If you’re still in early development then you might have a rough estimate of word-count and game length, but add at least 50% to that. (I’ll explain why later under “Localizing in Early Development.”)
You also need to know what languages you want to translate your game into, what regions you plan to release it in, and what systems you want to release on.
This is important to know early on so you can start finding possible localization agencies or localization specialists and identify their costs, their workflow, and speed (which impacts scheduling). The languages you localize your game into will also impact your UI, such as accommodating for languages with longer words like German, or right-to-left text display in Arabic. Which will take time to program and implement correctly.
(UI for localization and localization agencies are two other articles that I hope will come later!)
The Localization Timeline
Localization isn’t just “translate game → put text in game → profit.” It’s actually a little more involved than that, with multiple steps in a long chain that spans development.
For example, if you have voice acting and want to include voices in other languages, then you need to have the source text and voice recording done in advance, plan the time to translate the scripts, polish the scripts, adapt the scripts, find actors, schedule actors, record the actors, polish the audio, implement audio, check the audio, then (ideally) plan and record pick-up sessions.
If you’re familiar with working on voice-overs in your native language then doing it for localization is very similar, only you’re doing it another 4+ times in different languages across different time zones, involving potentially hundreds more people! (Localizing for voice-overs could be a whole article of its own!) And that takes time!
Your translation schedule needs to factor in the following things:
- Time for finding, screening, and picking localization agencies.
- Time for finalizing estimates and contracts.
- Time for preparing the text and reference documents.
- Total text ÷ translatable words per day (this varies between agency/language, but 2000 words is about average.)
- Total text ÷ edited text per day (this varies between agency/language, but 4000 words is about average.)
- Total time to play game ÷ LQA hours per day × at least 3 (for at least three rounds of checks and regression.)
- Time set aside to prepare and submit for ratings in different regions.
- Time for planning and negotiating publishing in different regions.
- Time to translate, edit release-related text (store text, websites, marketing materials etc.).
Other development-related schedules to consider include:
- Time for localized asset/graphic creation and implementation (consult your graphics team).
- Time to select and test fonts.
- Time to plan and adjust the UI for other languages (consult your UI team).
If you have localized voices in your game, then you will also want to factor:
- Time for finding, screening, and picking multilingual voice recording agencies.
- Time for finalizing estimates and contracts.
- Time for finding, selecting, and scheduling actors.
- Time for script adaptation, checks, and reference file creation.
- Time for the recording sessions. (This can be based on how long it took to record the source language.)
- Time for audio delivery checks.
- Time for pick-up sessions.
- Time for voice implementation.
- Time for LQA for voices.
Depending on the size of the game and the number of languages, voice recording alone can take 6 months to a year, starting from looking for an agency and ending in final implementation.
Localizing in Early Development
This is when you’re considering localization before the source language is finalized. Your game levels and design are probably in flux too, but if your source text (the language you’re writing your story, items, enemy names etc. in) isn’t pinned down then it’s a little more difficult to plan and stick to your localization schedule.
Which is why if you’re developing a game and only have a vague idea of how much text will be in the final game will be, add another 50% to your estimate. Game developers often underestimate how much text will be in their game, so it’s always better to overestimate than under.
When you start makes a big difference
The more “solid” or “fixed” the source game and the text is, the better it will be for the localization schedule and costs. But it also means the longer you have to wait between finishing the game and releasing it.
I’ve seen games that had to wait a year or more after completion for localization to be finished before they could release. Which cost them a lot of time.
But also, games that tried to start localization too soon ended up translating tens of thousands of words, only to scrap practically everything, re-write the whole story, then re-translate everything from scratch. Which cost them a lot of money.
Neither of these strategies are “wrong”, however. How your schedule plays out will depend on you and your project. The important thing is that everyone (on the dev team and localization team) is aware of the schedule and expectations. You have to consider your overall development schedule, how it will impact localization, as well as how localization will impact it—and communicate that with everyone involved.
Constantly editing text and changing the level design will impact localization
It’s also important to remember—especially if you start localization early in development—that changes you make to the source text or game will impact localization.
Obviously, any changes you make to the source text will need to be reflected in all languages. But other changes such as the number of enemies, or the gender of an NPC, or the object a character picks up, may also impact other languages even if they don’t impact the source language.
If you leave localization until after your source text is finished and find yourself with 8+ months while localization is happening, try not to spend that time polishing or tweaking the game. Or if you do, be prepared to pay for re-translations and re-checks!
Communicating these changes to localization (or at least adding/updating comments for them!) can also take time. If you don’t communicate these changes then they may (or may not) get picked up during LQA (localization quality assurance), which can make LQA run on for longer.
Expect the Unexpected
No matter how well we plan, something unexpected will always happen.
It might be an issue with ROMs for LQA failing (which can add a day or two to LQA), or PMs or translators leaving part-way through a project (so you and/or the localization agency will need to find replacements), or even a global pandemic that forces everyone to stay home for at least two years!
The best we can do is add buffers to every step of the schedule. 20% might seem like a long time, but you’d be surprised by how much a 20% buffer can save time in the long run!
Start Thinking About Localization Now
Localization often gets filed under QA, but you should start thinking about localization long before QA starts.
The act of translating the text does happen towards the end of development (as I mentioned, you want the game and source text to be as “stable” as possible), but other factors such as UI, graphics, and culturalization, will come into play a lot sooner than you might think.
Not to mention finding good localization partners, setting up contracts, answering questions, etc. can add unforeseen time to your schedule.
Hypothetical Localization Schedule
Let’s say I have a game that’s halfway through development. I think it will have 20,000 words, but just to be on the safe side I will calculate my schedule based on 30,000 words.
I want to translate my 30,000-word game into European French, Italian, European Spanish, and German (often called FIGS) from English. I don’t need to record any voices, but I do want all graphics translated.
Step 1: Finding a localization agency
It takes me about a week to look up three localization agencies I want to contact, then another week to send a query and set up the test. They have a week to complete the localization test. Then I send the test to other agencies linguists to get feedback on the quality, which could take another week.
Tip! Comparing quality with price will help you gauge their value for money – remember, cheapest doesn’t mean best value! And most expensive doesn’t mean highest quality!
I decide with my team members which company we want, then we set up a rough schedule and contract with the company, which could take another 2-3 weeks depending on how much back and forth we need for the contract.
The final part of this step is to arrange a kickoff meeting! This is my chance to work out workflow with the localization agency. What format do I/they want the files in? How will I answer translator questions from the translators? What will the overall workflow be? Etc.
Tip! Kickoff meetings are great and both experienced and not experienced companies should have them with all outsourced vendors.
Total time: About 8 weeks
Step 2: Reference preparation and familiarization
Once I’ve had my kickoff meeting, I should have idea of what format the text delivery will be in, and what reference files I want to share.
How long reference materials will take to prepare will depend on how much I’ve already prepared during development, the complexity of the game and text, the number of characters, etc.
Tip! See High Quality Localization? Help Loc Help You! for more details on what kind of documents you might want to prepare for localization.
For my hypothetical game it’s going to take me about 2-3 weeks (between my other tasks) to prepare reference files and the text for translation.
Also, because I want the translators to familiarize themselves with the reference documents, I’ll send those the week before I send the text for translation.
Total time: About 3 weeks
Step 3: Text translation and editing
My game is still in development, so I’ve planned for three batches of about 10,000 words.
The translators can translate 2000 words a day, and their editors can edit at a rate of 3000 words. Which means one batch of 10,000 words will take the localization agency 8 working days (not including the weekend!) I add an extra day in case we’re slow to reply to questions or there are any technical difficulties, so 9 days.
The 10th day (a Friday) I spend checking the deliveries and preparing and sending the next batch so the translators can start working on it Monday morning.
Tip! Don’t expect agencies or freelancers to work over the weekend. Translators and editors need time to relax too!
However, my team have been editing text! So, I need to plan a batch 4 that includes changes we made to previously translated text. Ideally including notes that say what was changed and why so the translators can edit the text quickly. This will take another week.
Total time: About 7 weeks
Step 4: Finding an LQA agency
While the localization agency has been working on the text, I’ve been looking into possible LQA agencies.
This is similar to the localization agencies, sending queries for quotations and selecting the best value for money. You don’t normally test LQA companies though, so that cuts time down a bit.
Another kickoff meeting is essential for coming up with a plan for how they’re going to test (the systems they’ll test on), when I’m going to burn ROMs, the workflow for delivery, testing, bug reports, regression, etc. I ask the agency how much time they’ll need and schedule my testing schedule based on that.
Total time: About 6 weeks (during the translation period)
Step 5: Text implementation and additional translations
Before we send the text off for testing, we need to add everything into the game. Not just the main dialogue, but menu text, any graphics, etc.
Implementing the text into the game will vary depending on the team and how much text I need to add.
There’s also a chance that text I missed (like graphics) needs to be translated too, and this is the perfect time to catch that! So, I plan for another round of translation.
The game needs to be in a playable condition for LQA! So, this is the perfect time to make sure it’s almost finished. Any level design changes may impact the translation, so we need to make sure the testing company is aware of those too.
Total time: About 4 weeks
Step 6: Localization QA (LQA)
Once all the text is implemented it’s time to send the game off to LQA for testing!
Remember to schedule a few days to burn your ROMs, send them, and let the LQA company test the ROM is working correctly and has all the languages before the testers start.
The team will also need familiarization time at the start of LQA before they start testing, so they can get a feel for the game and understand how your reference documents relate to the game.
Tip! Familiarization time with a ROM and reference documents will save a lot of time in the long run! A team testing without familiarization can result in very slow test times, or the need to add extra people to the test, which can impact consistency and quality when it comes to testing localized text.
My game takes about 7 hours to play if you’re familiar with the game, but 10 if you want to 100% it (find every line of text in the game). So, I estimate 12 hours completion for LQA (adding a 20% buffer). Plus, the testers need time to write bug reports, consult the reference documents, etc., which means that 12 hours can easily go up to 20. Which is about 4 days total per round of testing.
We burn the ROM on Thursday, send it Friday, the localization agency tests the ROM works correctly on Monday. Testers start Tuesday, sending bug reports throughout the week. Text corrections are sent at their end of day Friday. We implement changes into the game before burning the next ROM for the next weeks round of regression and testing.
It seems like one round would only take a week but fixing bugs, implementing text fixes, and preparing for a new round can take a few days, so scheduling two weeks per round can’t hurt.
Not to mention we’ll need to burn multiple ROMs from the different types of gaming system we’re testing.
LQA finishes with a final round of regression checks to make sure all reported bugs and text changes have been implemented correctly. Any outstanding issues that are fixed after LQA ends will need to be checked by the LQA company using screenshots (but that can incur further costs, so best to avoid it!)
Total time: About 8 weeks.
Step 7: Ratings, promotions, release
Don’t forget your game’s localization doesn’t end when the game’s been finished!
Any marketing and promotional materials should be done by the translation team that worked on the game. They know the game better than anyone and can make sure terminology and phrasing is consistent with the wording in the game.
A poorly translated website that doesn’t reflect the quality of localization in the game can impact the final sales.
Total time: Varies drastically depending on publishers and release schedule.
If you look at the image below my very rough schedule for a 20,000 to 30,000-word game that takes 10 hours to complete would take about 7 months to translate and test.
Of course, every localization is different! This schedule should give you an idea of how everything fits together, but it does not mean your game localization will take this long. You cannot, for example, translate a 750,000 word game in two months.
A schedule could also change drastically depending on a number of factors—the condition of the game, national holidays in different countries, people getting sick, etc. Unexpected changes to the game could also mean re-translation and re-testing is needed.
The localization process and schedule varies drastically from game to game.
But I hope the above information has helped give you some context for the general flow of localization and will help you prepare better for your game’s localization!
Hi everyone – and happy timezone! We...
21 Sep 2022
Your monthly guide to NYC game dev news,...
New York - New York City9 Sep 2022
As the second week of the semester starts and...
Texas - UT Rio Grande Valley5 Sep 2022