Unique words and vocabulary variety
In the detailed analysis WordCounter.Cloud shows a value called "Unique Words". It tells you how many different words your text contains. Together with the total word count it gives a rough idea of how varied the vocabulary is. This article explains how the value is calculated and what it can and cannot tell you.
How unique words are counted
WordCounter.Cloud splits the text into words at white space, exactly as for the normal word count. Then every word is converted to lower case, and each different spelling is counted once.
Two details are important:
- Upper and lower case do not matter. "Report" and "report" are the same word.
- Punctuation stays attached. "report", "report," and "report." are three different words for the tool.
Example: "Text, text and Text." has 4 words, and WordCounter.Cloud also counts 4 unique words: "text,", "text", "and" and "text.". A human would say there are only two different words.
A second example
"The report is good. The report is short. The report is clear."
This text has 12 words. WordCounter.Cloud counts 6 unique words: "the", "report", "is", "good.", "short." and "clear.". Half of the words are repetitions. Here the value shows the repetition clearly, because the repeated words have no punctuation attached.
What the value does not do
- Word forms are not merged. "write", "writes" and "wrote" are three different words. In languages with many endings, such as German, this increases the number of unique words.
- Synonyms are not recognised. "big" and "large" are two words, even if you use them for the same thing.
- Frequent small words count once. "the" may appear twenty times but only counts once as a unique word.
Punctuation as a source of error
Because punctuation stays attached, the same word can count as unique several times: once in the middle of a sentence, once with a comma and once at the end of a sentence with a period. In texts with many short sentences, the number of unique words is therefore somewhat higher than a person would count.
Calculating the share yourself
The tool shows the number of unique words but not a percentage. You can calculate the share yourself: unique words ÷ total words. In the second example this is 6 ÷ 12 = 0.5, or 50 %.
This share depends strongly on text length. In a short text almost every word can be different. In a long text, small words like "the", "and" or "is" keep coming back, so the share goes down. Only compare texts of similar length, ideally two versions of the same text.
How to use the value
- Spot repetition. If the share drops noticeably after an edit, you may have repeated certain words more often.
- Combine it with the keyword list. The keyword density shows which content words are repeated most often. Unique words only tell you that there is repetition, the keyword list tells you where.
- Don't chase variety for its own sake. Technical texts often have to use the same term consistently. Replacing it with synonyms can make a text less clear.
Try it
Paste your text into the WordCounter.Cloud editor and open the detailed analysis to see "Unique Words" next to the total word count. If you want to know how the total word count itself is determined, read the article on why word counters show different numbers or the guide on typical word counts.