Today, we introduced a change that affects the word count of all projects with Japanese, Chinese, or Korean as the source language.
Up until now, the word count for phrases in these languages was calculated by splitting the phrase on whitespace characters (tabs, newlines and spaces) and counting the pieces produced. However, the industry standard states that because whitespace characters in these languages are not used the way they are used in other languages, the word count should take into account the number of symbols instead.
With today’s change, the word count for Japanese, Chinese or Korean will be calculated by multiplying half the number of symbols in your source language by the number of languages Transifex hosts translations for. If your source text consists of 50K symbols, Transifex now considers that number to be 25K.
You can find more info in our documentation.