Man vs. machine: Is machine translation ready for retail?
Much has been made recently about the rise of machines, with a recent report from PwC claiming that more than 10 million UK workers are at ‘high risk’ of being replaced by robots within 15 years.
The translation industry is by no means exempt from these ominous predictions – with Google claiming that its innovations in Neural Machine Translation (NMT) have drastically improved the quality of Google Translate’s output.
This is likely to make machine translation an increasingly attractive prospect for online retailers shipping worldwide and needing product content in multiple languages – but can machine translation be trusted to never produce embarrassing linguistic errors? And should ecommerce brands put their reputation in the hands of robots?
We surveyed 400 consumers across France, Germany, China and Japan, showing participants side-by-side machine and human translated descriptions for fashion, homeware, technology and beauty products. Without telling them which was which, we asked for their judgements based on these three key metrics:
1. Level of understanding: Which product descriptions are hardest to understand?
2. Purchase intent: Which product descriptions make you feel more likely to buy the product?
3. Brand reputation: Which product descriptions leave you with the best impression of the retailer?
To find out whether machines or humans came out on top, download our report: ‘Man vs. Machine: Is machine translation ready for retail?’.