Despite being more connected than ever before, we only get a taste of the massive volume of multilingual content coming from every corner of the globe. For businesses, the internet allows for an immediate international presence and low-cost distribution, but without scalable, quality translation, many foreign audiences and markets remain untapped.
Translation at scale was a common thread running through last month’s globalization events (Localization World, the TAUS Annual Conference, CrowdConf and a Women in Localization meetup). Traditionally, industry dialogue has focused on comparing and contrasting machine translation (MT), community translation and paid crowdsourcing models. A recent, interesting development is companies using these models in a complementary way.
Machine + Crowd
MT has come a long way from its controversial beginnings and now holds a firm place in most large-scale localization strategies. Quality-wise, MT is not going to be sounding fully human anytime soon, but that’s okay: with translation, “quality” now means “adapted to context.” This was a key takeaway from the excellent LocWorld keynote by Robert Lane Greene of The Economist and TAUS keynote by Intel anthropologist Genevieve Bell.
Another learning was that MT innovation has shifted from outbound to inbound. Major MT players are exploring how to use it to understand vast quantities of data coming in for keyword mining, sentiment analysis, automatic triage, sorting and filtering, and other activities.
These recent developments are behind crowdsourced translation’s growing popularity as an MT supplement. For outbound translation, it contributes human grasp of context, with appropriate translation quality, at scale. For inbound efforts, the crowd adds more sophisticated emotional and language understanding to big data tasks.
Volunteer + Paid Crowd
We noticed a similar trend in the community translation space where the paid crowd is increasingly being used to supplement volunteer efforts.
Building on the community translation experiences pioneered by Wikimedia Foundation, Facebook and Twitter (among others), companies are realizing that paid crowd overcomes common community challenges such as meeting deadlines or translating less visible and engaging content. Some companies are bypassing volunteer translation altogether and using a paid crowd from the start for more control without sacrificing good pricing, speed and scalability.
Across the board, companies are looking for innovative solutions to new problems by combining multiple approaches to translation instead of sticking with just one. At Gengo, we’re excited to see crowdsourced translation as a key item on the menu—a regular part of the toolkits companies use to understand and be understood in today’s world of rapidly expanding content and data.
Go global with Gengo’s people-powered translation platform.
or Contact us