This article is originally published at translation-blog.multilizer.com.
Translation quality is one of the most important issues in translation crowdsourcing. Based on our analysis of dozens of crowdsourcing articles and cases we elaborate the quality aspect of our earlier article about top reasons for crowdsourcing translations.
The common view is that crowdsourced translations are usually better than professional translations. The main reason is said to be that “users know best”. That is, the end-user experience is the ultimate measure of quality and when end-users themselves do the translation they will use terminology and translations that they are comfortable with. Another reason is that translations often contain a number of small, perhaps secondary errors. Due to cost, these would not be worth fixing by professional translators but end-users are able to fix these in a couple of minutes.
There are also some hints that end-users might be more tolerant towards translation errors when they know that the “crowd” has made them. On the other hand it must be noted that e.g. with Facebook most users don’t know that the translations have been crowdsourced. With small languages professional translators with good knowledge of the subject matter may be very difficult to find. Therefore the choice could be between crowdsourcing and professional translator without knowledge of the subject matter. This obviously greatly affects quality.
To achieve good quality with crowdsourcing is not self-evident, however. It requires good processes that enhance quality. For example, without first creating glossary and agreeing on terminology, the translations can not be consistent. This problem is highlighted when new versions of products are developed. The time between publishing product versions is often long and the original translators might not we available any more. The translation discrepancy between products is a real problem but it can be enhanced with good crowdsourcing processes.
One aspect of translation quality is brand identity. Crowdsourced translation might not take this aspect into account sufficiently. Also regional variations in languages are a challenge. People will translate to their own language and sometimes will not even notice that their own language is actually a regional dialect.
In general the quality of crowdsourced translations is good, often even better that by professional translators. However, crowdsourcing does not automatically result in good quality. There are some challenges but it’s possible to overcome them. The most crucial thing is to have proper crowdsourcing processes in place.
This article can be re-published freely providing that it is not modified and the original source of the article is linked to.