Jan Hinrichs: how to localize mobile apps

Jan Hinrichs Beluga

Many developers start off by creating an app for the English-speaking audience.  It makes sense as the US market is one of the largest and a great number of users understand English.  Yet translating your app to other languages and making it available on other countries opens up new, often less competitive markets and helps get new users.

In this episode of AppInTop mobile app marketing podcast we talk to the founder and director of Beluga Linguistics Jan Hinrichs about all aspects of mobile app localization. Beluga Linguistics offers localization services in 26 languages.  It lists Last.fm, Swatch and Lookout amongst its clients. Listen to this podcast episode or subscribe on iTunes.

Can you truly have a widely successful English-only app?

Yes and no. Normally people start off with their native language.  This could be English.  This could be German.  This could be Spanish.

Most of our clients start with just their own language.  The first version of the app must demonstrate whether it is going to be well-received by the audience.  Does it works and do people get excited about it?

Then you go international.

Is it premature to think about going international before the app actually goes to market?  

By no means.  It’s key to have the global leadership in mind from the very first moment.  If you start just pushing out something successful that will need to be completely redone because you forgot about internationalization, you will be lost in the redevelopment process.

So, plan internationalization at the moment you start programming your app.

When you’re localizing an app is it best to start with the translation agency, hire freelancers or in-house translators?   

Have an early look at some of the tools we in the industry are using to do the translation and localization work.  These are often third-party services or platforms where you upload the strings and SDK files.  Then pull together your team of translators, reviewers and your managing staff and check out those platforms.

Is your code is suitable, it does it work well with this platform?  There is a  couple of platforms such as webtranslateit.com or transifex.com, which offer a lot of tips on how to prepare your files well so you can run one, two or fifty languages on it.

Is there a type of app that is not worth translating, that is not going to succeed outside of the local market?

Consider for example a service for restaurants, the restaurant recommendation service.  This is hardly scalable because you need a large investment in marketing to get restaurants to use your service and getting recommendations working well.

What normally happens is that if you’re building up a slow-growing service because it’s difficult to establish, you will find a copycat in the other markets.  It is especially true for China.

When you’re building your internationalization team, where do you begin?

You have to decide if you are going to do it yourself.  Then you can, for example, bring your friends into the process.  This is an approach that can work but it’s hard to maintain.

You can also choose a professional service.  Typically such services will have experienced people, with teams that are working together for a long time.  But then those teams will need to get familiar with your product.

If you use friends, crowd translation, or your employees, take into account that it will not be easier or less expensive than a professional service. It’s different.  It has a lot of advantages, especially when people are really motivated because they like your product.

Ask yourself, “Is my service really like Facebook?  Is it really so engaging that I can build up a system around volunteer translators?  Is it better to concentrate on my key markets and hire an agency, translation company or freelance translator working when I need them?”

How can you decide what languages to translate the app to first?

You have your road map in mind as a developer.

If you go international you decide what your preferred market is, where to make money or where you have better chances of finding influencers, venture capitalists or any other asset to help bring the app to the market. This is how you decide what languages to translate the app to.

One of the most difficult countries is Korea.  It is huge in terms of users but I haven’t seen a foreign app so far having success in Korea because Koreans seem to prefer their own homegrown apps. An app interface really needs to be local there.  And this is a process we can’t offer our clients because it means redoing the whole interface.

Chinese are more open to adapt foreign games or applications but Koreans less so. So we say, “Okay, go to Japan. Japan is open.  Go to China, to Taiwan.  We can do Korean, but Korea is really hard”.  And it’s even harder to find suitable translators there.

If you just do a translation of your app to Korean, the problem is the language because it’s really flipped around completely and an English sentence translated into Korean will most probably have little meaning to them.

Also it is a mature audience which may simply not want or need to use something foreign.

Lookout alone has been translated into 10 languages.  How many downloads did the app see?

Total number of downloads is 60 million.  We are lucky to be involved in some really impressive localization projects. Last.fm has millions of users around the world. In terms of translation, it all had to be custom-made.

How do you work at Beluga, how do your teams work?

This is really a virtual business because all our translators are freelance translators.  We work with 150 translators around the world.  They are all on board for years.

All communication is done via email and Skype.  If all of your translators live in one city, the culture changes their view of their home countries and their mother tongue.  So they will start to forget about the slang.  This is why we prefer people based in their home countries to make sure that people really know their environment well.

We use Skype and screenshots, video calls and screen sharing. We communicate a lot via Skype chat.  For example we say, “Hey, we need a translation here.”  In a couple of seconds you got a reply from Japan, from Korea, from Brazil.  And this really works pretty well.

What we do is human resource management in order to find the right person for the right project.

Take me through translating and localizing an app.  What process do you guys follow?  What’s an average length of time it takes from the very beginning to the very end?

When big clients such as Lookout or Swatch contact us, they already know exactly what they want.  And they are almost ready to start.  This is one case.

The other cases is when we get approached by an app developer saying, “Hey, we want to translate.  But we have no idea.”  In that stage I  explain the companies what they have to look at before they start internationalization.

So I advise them not to do hard coding; to use international standards such as Unicode to creates strings that have variables in the translatable string, etc.

This is important, because when you have Turkish, for example the sentences are built the other way around.  So if your English sentence is, “Hi, my name is Michael.”  In Turkish it is, “Michael my name is.”  So if the user name is a variable, this variable must be changeable for the translator.

A lot of developers still think, “the less code I offer in the string, the safer the process is.”  And this is not the case because it would break your neck when you want to speed things up, go from one language to ten languages and so on.

Today, editing programs offer one line of code that you can integrate in your files and they take care of getting information that needs to be internationalized and translated.  And they pull out the PO files to have text online for the translators to do their work.

And when we start this process, before the translation begins, we usually have a look and add comments.  Normally the developers just push, “Here’s the app.  Here’s the file.”  And that’s it.

But we always have a problem with context.  You have a mobile app and you have strings.  All these are text strings.  If you get them out of the context and put them in the editor, you lose context.

All you have are clues like the key name or some sort of information around where the file comes from.  Sometimes it’s not enough because an application may be pretty complex.  You have a lot of text, which is hidden in the source code and is only viewable in a given situation.  For example, what if there is an error message?  What about male or female users?  What about geographical differences? You have a lot of ‘if’ clauses.  This is a hard part of our work: to offer the translators the context they need to get the translation right.

If there’s not enough information, we go to the clients, and get out the screenshots to give reference.  Text length is also very important, because translations can grow.  When you translate an English phrase into French, your sentences are 20 percent longer.

If you are just looking on the mobile screen, some parts are really restrictive.  So you need to put these limitations in the string information so that the translators can’t do more than 20 characters, for example.

What was the biggest challenge in Last.fm’s experience?  

We had an interesting process because once we have translated everything, we started forums, local forums on Last.fm to discuss our translations, to deal with user feedback.  And this is always tricky.

I’m telling the story because if you think about crowd translation, this is exactly what you will find — a lot of discussion around a given string of translation. You have French from France, you have Canadian French.  You have Latin American Spanish and European Spanish.  It’s mostly okay for Latin Americans to read Spanish from Spain, but Spanish users are very aware of translations done for Latin America.  This all leads to intense discussions about the translations, and there are many opinions.

So crowd translation is a tricky process because additionally you have to manage the emotions of the users, which are of course right because they’re the heavy users of the application.

But on the other hand you have linguists, which really know their grammar rules and the standards and have investigated every translation.  This is not coming out from nowhere but experience.

And you have to manage not to lose your translators because of some discussions, while offering the voice to the users.

To what extent does an app developer need to localize their content beyond language?   

Of course you have to be very conscious about your product.  You need to know if your product needs to be localized or if you can sell it around the world in the same format.

The concept of some games, especially the games that take place in Middle Age, doesn’t work in China.  In China they just don’t have any idea what the European Middle Age is.  But if you offer them a game themed around Chinese Middle Age, you have a pretty good chance.

So you need to consider your product.  Music for example can be global of course.  You need local content, but the interface can be the same.

Many of our clients create a standard dashboard for their services, but build local landing pages to give a special localized touch.  And that works well.

Are there any business processes within an app that need to be localized?  Payments, support moderation?

This depends.  You have more than 50 different types of payment out there.  So we usually see credit card payments.  There are 50 sorts of credit card payments.  And then you go to online payment wallets and the choice is getting huge.

And this is something you really need to check out with your product development teams for the different countries.  What is the preferred payment system in France or in Germany?  It may be not the credit card but a bank transfer.  And this takes time.  You can go pretty easy with PayPal right now.  But it’s not necessary the preferred choice of everybody.

And MasterCard or Visa are not key leaders in some markets.  And if you go to China, you have to investigate.

When it comes to overall support for an app in that particular language, how does that process work?

Normally it’s marketing agencies, local marketing agencies that helps our clients to get to bloggers, to do the local landing pages and that stuff.  And then there are the venture capitalists, which are doing a big job with bringing in contact the CEO’s of that companies with their respective partners in those places.

What we do in our job is we help our clients when it comes to keywords, keyword localization.  We do testing, etc.

Of course we would give them the feedback if we are asked “Do you think this makes sense to go up to that market?” But don’t forget we are translators and not a marketing agency.

How do you see the process of app localization developing in the next few years?

This problem of in-context translation will be addressed in the future, and it is especially relevant for mobile apps. We will see more editors like Colátris (late Babel SDK), which offers translating your app with your mobile phone.  So you have the layout on top of your application and then you can translate every sentence you see on the screen one by one.

It is something that is missing today.  It is really difficult to review and testing an app in the translation workflow.

And I think we’re going to see a big move in video and creating subtitles for video content as going international becomes more important.

This is an abridged and edited interview from the AppInTop mobile app marketing podcast produced by AppInTop, an automated mobile app marketing platform. Listen to this podcast episode or subscribe to the podcast via iTunes.