by Luciano Oliveira, @TTCLuciano
Not long ago, any piece of content to be translated would mean new business for translation companies. Today, the share of content to be translated that actually goes to a translation company has been substantially reduced. A few ideas on why and how this happened, and what this means to the translation companies.
FIPO stands for “For Information Purpose Only”. This term was created to differentiate content that was “just for information” (like the name says) from “for publishing” translation. While it is obvious that anything being published (be it on paper or online) should have a perfect quality, this is not necessarily true for FIPO. When someone just needs to know the meaning of a certain content in a different language, they are more tolerant to inaccuracies.
FIPO translation represented a large chunk of the translation business in the past. Not many years ago, companies doing global business would translate lots of documents and content “just for information”. Examples:
The Translation Company Group LLC, also known as TTC, estimates that back in 2007 almost 50% of all translations were FIPO translations. 2007 is exactly the year when Google made available their machine translation service. Google Translate, a proprietary algorithm, based on statistical models, provides automated machine translation for free (more about Google translate history on Wikipedia).
As of January 2016, Google Translate supports 90 languages (some languages better than others) and serves over 200 million people daily. One may assume that the vast majority of these translations are “FIPO”. People are now getting cost-free translation for content they would be paying big bucks to translate 10 years ago. The quality of these translations is far from professional, but they work very well when you are willing to exchange quality for free cost. So, in our three examples above, instead of going to a translation company, translation clients are now going straight to Google Translate.
For TTC and other language companies alike, Google Translate meant a drastic reduction in FIPO translation business.
Two Things Going For Translation Companies: MT Challenges and Growing Global Demand
Machine translation has become great for FIPO translations, but it is decades away from offering a reliable translation for publishing purpose. For instance, there are hundreds of cases of mistranslations from Google Translate in the news. A recent example:
The reasons for mistranslations are multiple. There are still technical limitations in the field of machine translation, and human error or manipulation may also cause serious errors in the translations provided by free translation services. As Google Translate, for instance, relies on feedback (‘edits’) from translators using its platform, an orchestrated effort by a large number of translators may indeed lead Google Translate to assume a mistranslation is actually right.
In the “Russian Federation” case above, Google replied to The Guardian with a statement saying that “its translator tool works without the intervention of human translators”, which is obviously not true when you remember that “In the (Google Translate) web interface, users can suggest alternate translations, such as for technical terms, or correct mistakes” (source: Wikipedia). Google themselves ask users for help on their website:
And, we should remember that to get to this current state of quality (or “lack of quality”), decades of R&D were required. Automated translation of texts became a research subject for the first time at MIT in 1951 – meaning 65 years of investments in such technology. Even considering that the evolution of technology may be exponential instead of linear, most experts in the language industry don’t see a pure MT translation being used for publishing anytime soon.
Despite downturn in China and other problems worldwide, translation services business in general is actually posed to keep growing steadily. “As nonnative English speakers came to the United States and companies took products overseas, demand for translation services was driven upward. Though industry clients will continue to be price-, service- and quality- conscious, globalization and an increase in immigration will boost demand for industry services over the five years to 2020” (IbisWorld).
Curiously, even machine translation has created a new market for translation companies. Now, they offer the so-called “human edited machine translation”, which means a text pre-translated by a machine translation engine for a human editor to finalize it. This machine translation used by translation companies is not the same as Google Translate. They have more complex (and costlier) options, which may render a better result at the end.
These hybrid applications are ideal for large projects to be completed in short turnarounds. And, they are now representing an entirely new business line for translation companies.
The end result for translation companies is an actual increase of translation business along deep changes in the state of the art in the industry.
Year after year, there are more translation dollars in the market available to language companies. However, not all companies may be getting the same access to this growth. Larger translation companies are better equipped to leverage on technology and highly specialized professionals (e.g.: localization engineers) to handle translation projects that are more complex by the day (e.g.: the human-edited machine translation projects).
Our prediction is that small to mid-size translation companies will suffer and struggle to adapt, while larger ones will get the most from these changes. Small monolanguage agencies along freelance translators, where most of the hands-on translation actually occurs, will continue to enjoy an increasing demand even if translation eventually becomes an “editing job”.
Luciano Oliveira is the CEO of The Translation Company Group LLC, a translation company with offices in New York City and San Francisco. He lives in New York City along his wife and four kids.