The following is a recap of our webinar Translation Trends for 2015 featuring David Canek, Chris Wendt, Jan Hofmeister and Torben Dahl Jensen.
Additionally, the webinar featured case study presentations by two large language service providers that use cloud technology to automate and grow their business.
- Moravia’s move to Cloud translation environment.
- TextMinded becomes a translation solution provider.
Slides are available via Slideshare
In the beginning of the webinar, CEO of Memsource David Canek stated that from his perspective, the primary trend in the translation industry is the advent of the cloud technology.
Throughout 2014, Memsource’s user base more than doubled, reaching 30,000 users. Over 1,500 translation providers now sign up for the service every month, and approximately 100 million words are translated. According to a survey carried out in September 2014, more than 93% of Memsource users are satisfied with the service. These figures are indicative of a strong growth in the use of cloud tools.
The second trend emphasized by David was the rise of Machine Translation (MT). Since introducing analytics in Memsource, it is now possible to calculate exactly how widespread the use of MT has become. Over the past year, 47% of all translations in Memsource have been carried out with the support of MT. This is an impressive figure, since Memsource users are mostly professional translators and translation agencies and use MT with post-editing to improve overall productivity. In comparison, 72% of projects leveraged translation memory technology, while only 42% projects utilized term bases.
The third trend mentioned was the move from isolated “silos” in the supply chain to closer collaboration and sharing of language assets between buyers, LSPs and translators. In 2014, Memsource launched Shared Projects, Shared Jobs, Instant Quote and other collaboration features to facilitate this transition.
David Canek announced that in 2015 under the LSP Partner program, Memsource will focus on empowering translation companies to become solution providers. This business model implies that LSPs manage the linguistic assets of their clients, and provide a technology platform in addition to translation services.
Chris Wendt, Group Program Manager at Microsoft, introduced machine translation Post-Publish Post-Editing, or P3. The concept is to publish raw machine translation first, and then do post-editing on a live web page.
Post-Publish Post-Editing allows inexpensive translations to be produced instantly, and human translators can improve the quality later on in a “lazy fashion”. P3 can even leverage community translation and engage domain experts, enthusiasts, or the buyer’s company employees since time constraints are more relaxed in P3 than in professional human translation.
Chris explained that while MT post-editing is used to achieve human translation quality with less effort, in reality the productivity gains are usually not very impressive. This is due to quality control and management creating too large of an overhead.
In contrast, with P3 productivity gains are much greater, because only relevant pieces of content are localized. The rest is left to raw machine translation, which in many cases is good enough for practical purposes. Chris suggested that in 2015, Post-Publish Post-Editing will occupy the middle ground between raw MT and human-quality post-editing.
Jan Hofmeister, Senior Project Manager at Moravia, focused on his company’s transition to the cloud translation environment as well as a massive systems integration project to automate content flow.
There were a number of requirements that led to selecting Memsource as the cloud environment of choice. These included:
1) The ability to exchange files between PMs and vendors without resorting to email.
2) Agile content flow: Moravia’s clients often change source texts on the fly, and these changes can be shared with translators in real time.
3) Previewing of completed translations for specific file types.
4) API to integrate the platform with Moravia’s internal tools for QA, internal MT engines, accounting.
Once the integration was complete, Moravia had fully automated several steps in project management, including content pre-processing, uploading, task assignment, and more. The resulting solution is now equipped with analytics that calculate costs and deadlines for clients. This is a very welcome and appreciated feature. Jan concluded his presentation by stating that Moravia will continue with the integration of their own workflow and content systems with those of their customers and with Memsource.
Moravia localizes most of the content into 20 languages, and in some cases into 100 languages, with the average throughput measured in millions of words per month. Its technology approach sets an example for other providers. It is likely that more translation companies will follow in these steps and focus on automation and connecting their systems with those of the clients.
Torben Dahl Jensen, Partner and Language Technology Manager at TextMinded explained what it means to be a translation solution provider.
The model treats all people working on a translation project as a community – possibly including the buyer’s in-house specialists, LSP staff, freelance translators and subject matter experts – and aims to give this community a single platform to collaborate and share linguistic assets.
Torben provided an example: one of their important clients had an internal translation department, and they sourced additional volumes that they couldn’t process internally to TextMinded. Initially the translation company built their own terminology and term base for this project. The client then asked whether it was possible to share translation memory and terms. In response, TextMinded built a solution with simultaneous data access using Memsource.
At the end of his presentation, Torben emphasized the importance of sharing language assets and connecting different groups of specialists involved in producing translations for the same client organization. The move in this direction – sustainable communication – is going to be one of the major trends in 2015.