Skip to main content

We Can’t Translate The Web Using Only Professionals

 

As national and cultural borders dissolve due to technology and our increasingly globalized world, the need for translation jobs is growing exponentially. Businesses want to be international and multilingual, as they should. There are around half a million professional translators out there right now doing this work. And they’re doing great work.

But of course, manual work isn’t scalable in any way. So long as translation remains human-based, it simply cannot keep up with demand. We’re obviously not at the place where we can eliminate human translation. I mean, have you tried using Google Translate lately? It can produce pretty hilarious results. It’s unusable for any company that wants to sound professional.

After all, translation is just another form of communication, and communication is built on human-ness (this may or may not be a real word). If we strip out that human touch, we can’t be sure if others actually know what are saying. Misunderstandings are at the base of the majority of human conflict and can have disastrous, expensive repercussions for a business.

We just can’t expect professional translators to keep up with this demand. There’s just too much. We did some basic calculations to relate the max capacity of translators to the sheer volume of content out there, content that ideally, will continue to be translated into new languages so that more and more of the world can enjoy it and learn from it.

The capacity of professional translators

A professional translator is able to translate roughly 250 words per hour. Not only do they examine nuance and subtlety to make sure the intended message comes through, and take into account the context of the document, but their capacity is also restricted by simple restraints like typing speed, and of course, the distractions we are all bombarded with constantly.

If someone translates 250 words an hour, 8 hours a day, 52 weeks a year, they cap out at around 520,000 words per year (to put this in context, the median length of a book is 67,000 words, so this would make about 8 books). There are around half a million professional translators out there, meaning their total capacity is approximately 330 billion words a year.

Now, that sure sounds like a lot. But to put it in context again, there are 288 billion indexed web pages (as of December 2014). And of course, there aren’t many single word web pages out there. To approximate on average how many words are on a web page, we can compare two standards. First, Google recommends at least 250-300 words on a page for it to index properly. And Medium, after studying their metrics, determined that the ideal length (for a blog post) is 1600 words. So conservatively let’s say each web page has, on average, 500 words.

Now, that means there are approximately one hundred and forty-four trillion words online, give or take a few hundred billion. For the entire community of pro translators to translate that into just one other language, that would take four hundred and thirty two YEARS. Wow. That’s a lot of coffee.

And this isn’t even counting the trillions of words of offline content (for example, Google estimates there are about 130 million books out there, meaning another 8 trillion words just in books that ideally, would also be translated. And we can’t forget that the amount of content online and offline is growing rapidly, faster and faster than ever before. The amount of indexed web pages is growing at a rate of 21% every two years (side note: this is based on data from 2010, so it’s safe to say it’s growing even faster now).

Of course, this is all very rough math, and it can be off by large amounts when estimating so many of the variables. But it’s definitely within the ballpark of reality.

What does it mean?

It means that professional translators, no matter how great they are at their jobs, can barely make a dent in translating all of the content we produce. And to have a truly connected world, we need to be able to include every citizen of the planet, no matter what language they speak. This is actually part of our mission statement here at Unbabel — the entire reason we do what we do:

We believe language should not be a barrier to communication. We believe in an inclusive internet, where all content is available to anyone no matter where they are, no matter what languages they speak.

For this to happen, we can’t just rely on human translators. We have to allow technology to do the heavy lifting, so that the human translators can focus on the art of language and translation.

Machine translation has a long way to go, but here at Unbabel we believe that when machine learning and real live humans work together, it can produce results far above and beyond what either tactic can do on its own. We use machine translation to do the bulk of the work, and our translators check every word, every phrase, every double meaning to make sure that the translated version flows as naturally as it does in the original language, and the personality and tone of your content can shine through.

We want to live on a more open, more connected planet. We believe in the power of connecting humans and technology to change the world. Join us.


 To learn more about Unbabel, visit us at Unbabel.com. To become an editor with Unbabel, sign up at our Editors page.

The post We Can’t Translate The Web Using Only Professionals appeared first on Unbabel.