Back when we were still selling Triggertrap Mobile, translating the app into German gave us a 750% jump in sales. It also helps that we received a lot of PR from launching with a major web shop in Germany, of course, but there’s no doubt in my mind that having the app translated into German was a crucial factor to that happening.
And yet, when we launch version 2.0 of the app, it will only be available in English, at least in the short term.
Internationalisation — or I18N among friends, so named because it starts with an I, ends with an N, and has 18 letters that are way too slow to type out in full — is a key part of any business delivering software apps. The first time we did the translation was a pretty complicated process: It’s impossible to correct the copy of an app without doing a full software update, which is less than ideal for many reasons. Obviously, that also means that if you have a typo in your app, it means going through the whole iTunes Connect app submission process and — if you’re a particularly careful developer — the full suite of pre-launch tests. Obviously, this carries a cost, both in terms of time and money.
Languages and quality control
I speak in squares. You speak in circles. And some stock photography really makes absolutely no sense.
At Triggertrap, we’re proud of the way we use language. The blog posts have character. The tweets have personality. How we communicate is an integral part of the Triggertrap (dare I say it…) brand identity.
Of course, when the words and phrases we use are being pored over by half a dozen people, tweaked, loved, and adjusted all the way throughout the development process, the translation of said lovingly crafted prose is going to be important, too.
For context, I speak three languages fluently (English is my Third language), I get by in a couple of other ones, and I can get the gist of a written piece in about a dozen languages. This means I have a lot of opinions about what ‘fluency’ means in a language. Yes, I did work as a journalist for the biggest newspaper in Norway, and at the time, my Norwegian wordsmithery was fluent. But that was the best part of a decade ago, and I haven’t really used my Norwegian since. Nonetheless, I did translate Triggertrap into Norwegian. Is it factually correct? Probably. Does it have spelling or grammatical mistakes? Potentially. Does it have the pizazz of the English version of the app? Absolutely not.
It turns out that some of the jokes and phraseology we use in the Triggertrap Mobile app are untranslatable, because they rely on cultural concepts, idiomatic twists, or daft puns. We love ‘em, but puns, especially, are notoriously hard to translate. The other aspect of that is that even people who speak English to a non-native-level may find themselves struggling with our silly little jokes. And so, we found ourselves facing a choice: Do we make our language simpler? Should we eschew some of our brand identity in favour of clarity and ease of use?
When I spent a bit of time talking to our customers, we discovered that the quality of the translations varied wildly. Some of them were pretty close in quality and feel to our English version, whilst others were shambolic and scarcely useable. To be frank, it was fiercely embarrassing… But at least as soon as we knew, we were able to dig deep and figure out how we get here.
In a small company, I18n is unreconcilable with agile.
Finding someone who will translate your app for you is easy. It’s cheap, too — for most languages, we were able to find translators (via Elance, oDesk, and similar sites) who were able to do the full translation for around $400 or less. But the biggest cost was the challenge of Quality Assurance. It’s bloody difficult, because no matter how you look at this, translating the better part of 400 phrases into 17 languages, is 6800 opportunities for Getting Things Wrong. The additional challenge is that there are different degrees of wrong.
“Press the button”, “Activate the switch”, “Tap the control”, “Fondle that round thing”… They’re all perfectly valid phrases. Some of them will work in some of the 17 languages. Some of them will not. Some languages have absolutely bonkers grammatical effects, too. In Russian, for example, the word “hour” would change based on the number of hours. English may have 30 words for ‘switch’, but Norwegian isn’t going to be that lucky.
All of these things are well-trodden ground in terms of translation challenges. There’s a simple solution: Pouring a lot of money into it to make quality soar.
Our biggest discovery was that loading up our app with languages is great for marketing and international customers, but it’s lethal to rapid and agile development.
The big problem is that the project management overhead of managing 17 translators, 17 copycheckers, and then finally checking every screen of the app in every language to ensure that international glyphs (Cyrillic, Japanese, Chinese), long words (German, Finnish) and right-to-left languages (Arabic, Hebrew) don’t show up in unexpected ways takes up an immense amount of time and project management bandwidth, which is very hard to reconcile with the overall priorities of a small, agile company.
An alternative approach
From our Triggertrap Mobile version 2.0 we’ve decided to take an alternative approach to I18N. Yes, it’s frustrating to a lot of our users that the app will only be available in English in the short term, but starting from scratch means that we’re able to assess which languages our customers are actually using, and from that point, we can ensure that the translated apps are much higher quality than before.
Sometimes, you have to take a couple of steps back to be able to move forward.