In spite of the fact that word-count is standard practice in the localization industry, when it comes to languages with no spaces, character count becomes the normal practice:
Normally it is a price per character (for Chinese). In Chinese, a lot of words combine two-tree characters but it is not possible to account automatically the words in a Chinese text because of the same character is used in many combinations doing different words. On the other hand, in Chinese grammatical words are less in number than in another language. Japanese combines alphabetic characters (from two alphabets) and Chinese characters. There are more words in Japanese for the same sentence because of the grammar structure of Japanese. A Chinese text condenses a big volume of information in a small volume of words and therefore the translation of this text becomes somewhat bigger.
The most logical method of text volume evaluation in Chinese is character count. E.g. a 1000 word English text translated into Chinese might be around 1300-1800 characters long
Japanese, when translated into English (or any other language), is usually charged per source character because, outside of particles, it is not ossible to tell without reading the text where one "word" begins and another ends. The standard formula for estimating how many English words will come out of a given Japanese source text is to divide the total number of Japanese characters by 2.5.
This formula can be used to make a rough estimate.
In the JP>EN market, most fees are based on the source character count.
2 J=1 E: 1000 J chars = 500 E words.
Although it can be somewhat lower (as in less than 500 E words per 1000 J chars), depending on the field. A good rule is 2J=1E.
Modern Korean is written with spaces between words (unlike of Chinese or Japanese). Traditionally, Korean was written in columns from top to bottom, right to left, but it is now usually written in rows from left to right, top to bottom. This means that the traditional word count scheme, when a word is counted on a spacing basis can be applied.
Languages that don’t have spacing and require character count include: Chinese, Japanese and Thai.