I know. I am terribly late with my report. But here it comes. If you want to see to the future of language services, have a look at global GDP development and the digital content explosion in the Asia-Pacific area in non-native English. Automated normalization of English may be the basis of finally succeeding in machine translation. This is one of the angles we learned in this year’s KITES Symposium that focused on language technology and international business development opportunities for Finnish companies.
The impact of language on brands and growth is huge. Even more than language and technology, business depends on multicultural communication, for there is no business without communication. The outlook for the language services industry is good, because currently only a fraction of existing digital content is being translated. Approximately 20 languages are needed to reach 80% of internet users, of whom only 27% understand English. Managing multilingual business is thus an essential part of any company’s strategy, while linguistic and cultural skills are essential for communication.
We heard an excellent example of best practices from F-Secure. Their documentation and user interfaces require massive translation volumes, millions of words localized to various languages. With a tight budget, an even tighter process, and extremely fast production pace, they’ve learned that the cost of opinions in language matters is high to an organization, and they’ve found a solution in creating translator freedom. F-Secure says their translation is 20% more productive paid by the hour versus per word. Among millions of items handled, only 30 real bugs affecting end user experience were found. For practical reasons, they use dedicated translators and seem to encourage translators’ networking and empowerment. There is no absolute business value in a corrective process as such. The only thing that matters is the user experience. Translators can learn from their “mistakes” during the next round of updates. No complaints on user experience is a sign of well-planned terminology, and a lean process is based on analysis, automation, and waste minimization just like they do in process technology.
Novita is another successful Finnish brand seeking international growth by multichanneled marketing. They have established a worldwide knitting club called Novita Knits and profiled their target groups based on universal motivation factors instead of cultural differences. Yet they need cultural insight when conveying the messages to new markets. Their company values and CSR are communicated through content and actions, not statements. To ensure success, their story should not be translated but retold and taken to heart in another culture. I detect a definite demand for multilingual narrative and transcreation services.
After the business track, Kites Symposium 2014 got geekier towards the end. We learned that Google has launched a universal dependencies project that includes the Finnish language in order to create a universal parser and to advance the development of machine translation. The Finnish Internet ParseBank developed and maintained in the University of Turku currently holds a huge corpus of circa 3.5 billion tokens which translates into 270 million sentences. Parallel work is going on in the University of Helsinki while both language analysis projects are suffering from the strict copyright law of Finland. Usually lack of copyright creates a setback for translators but in this case it seems to be the other way around. Already looking forward to what updates we get to hear next year at #KITES2015.