Client context and challenge
What is Post-Edited Machine Translation?
Post-Edited Machine Translation (PEMT) is the use of high-quality machine translation in a workflow to allow new content to be quickly edited, as opposed to translated from scratch.
PEMT can be a cost-effective and time saving alternative to the regular human translation workflow, as post-editors only need to review and edit previously-machine translated content.
To be most successful, PEMT needs engines that are tuned to match the content you translate. At Lingo24, we test using live data, and only introduce PEMT if productivity gains are certain.
PEMT helps clients to make the most of their existing translation budget and unlock new growth opportunities.
Lingo24 conducted an assessment of the end-to-end workflow involved in the translation and review of content for international launches to identify acceleration opportunities.
Three key actions were identified to reduce the time to market:
- Build high-quality customised Machine Translation engines based on the large set of previously translated data available
- Introduce PEMT to move from translating new content from scratch to post-editing high-quality engine output – making the translation phase faster
- Once happy with the output from the engine, gradually introduce exceptional in-market review of the content, making the review time shorter.
This approach was trialled on a single language using the client’s existing language assets to build an engine tuned to their exact content which would significantly outperform generic engines available on the market.
A trial period was conducted to test the process and assess the feedback received. Along with Coach’s productivity metrics, this information confirmed the initial expectations that this approach resulted in the same quality at a faster pace.
Client was able to reduce their time to market in international markets from 2 weeks to 1 week, allowing them to grow international revenues from 15% to 55%