Machine Translation - Not A Good Idea.
We just received a machine translation offer from a company in Russia. Unfortunately, it looks like they’ve been over using their product. The brand name is “Promt”. Any English native will see the problem with this. Whether deliberate or accidental, it looks like a typo. It won’t help them be taken seriously. It isn’t even phonetic, as the second “p” in “prompt” should sound. So is machine translation a good idea. The answer of course is that it depends how you want to use it. If you want to understand the gist of a website, then it has some merit. However, in a professional context, machine translation will damage your brand.
Machine translation – not a good idea
The machine translation industry quotes 70% accuracy. However, you should be very skeptical about this type of claim. At best machine translation gives you some understanding of the gist of a text. For example it will probably give you enough to understand whether you have received a genuine quote enquiry. You will know whether it is just a computer generated junk e-mail which you can safely delete. Hopefully you will know that, but equally you might not…..
Even at the level of positive and negative you may have problems. The wording will always be literal, and any word with multiple definitions will produce rubbish. More information about this type of issue can be found in this article on free translation. We always human author our translations. We carefully select our translators and proofreaders. They are highly skilled.
Marketing is a major challenge for machine translation. The problem is that marketing translations are never literal, and word by word. The required style and tone in the target language may be different. The text may actually need very different wording. For example, marketing often uses plays on words. An English play on words is very unlikely to work when translated, the target language would need to have the same double meaning for it to work.
OK, so not marketing, but what about a technical text? Unfortunately for the machine translation industry, technical wording doesn’t work well either. The issue is that a good technical translator, is in effect a technical author. They need to understand the technical terminology of a particular text in both languages. The target text still needs to flow and make sense once translated.