The size of a translation job is expressed most of the time by the word count. This measurement looks simple, but when what is counted has not been defined, it creates a difference between the two sides. The same document can give different results in two different tools, and this difference has commercial consequences.
Definition
A word count is the counting of the word units in a text. The count is taken on the source text. The result is the indicator of the volume of work.
Counting on the source text
The count is taken from the source text because it needs to be known before the job starts. The target text does not exist yet. This provides predictability.
The target text difference
The translated text comes out longer or shorter than the source, and its counts are different. Which of the two is taken as the basis is written in the contract. No ambiguity is left.
Differences between tools
Different tools calculate the count slightly differently. Constructions with hyphens and apostrophes may be split differently. The same tool is preferred.
The status of numbers
Whether or not digits are counted as words varies with the tool. In documents that are dense with numbers the difference grows. The rule is made clear.
What is not counted
Text inside images does not show up in an automatic count. These parts are calculated separately. If they are skipped, the volume comes out short.
Headers and footers
The content at the top and bottom of the page is not counted in some tools. These areas are checked. The shortfall is corrected.
Hidden content
Hidden rows and slides may not enter the count but may need to be translated. Their existence is checked. The scope is determined.
Footnotes and endnotes
Because notes sit in a separate flow, they may be skipped. These areas are counted separately. A check is made.
The status of repetitions
The raw word count includes the repetitions as well. Repetition analysis is a separate measurement. The two figures are not confused.
Raw and weighted count
The raw count takes all the words into account, the weighted count the matches. The two give different results. Which one is used is put in writing.
Variation by language
Some languages express the same meaning in fewer words. This produces a difference depending on the language pair. Comparison is made with care.
Languages written without spaces
In some languages the words are written joined together, so a word count is not meaningful. In these languages a character count is used. The unit changes.
Recording the count
The tool and the settings with which the count was taken are recorded. The record makes later verification possible. Transparency is preserved.
Sharing with the client
The result of the count is shared together with the quote. A figure whose basis is not shown becomes a matter for bargaining. The report is attached.
Resolving a difference in counts
When the client finds a different figure, the settings are compared. The difference generally comes from the method. The solution is technical.
In scanned documents
An automatic count cannot be taken on scanned files. An estimate or text extraction is needed. The method is communicated at the outset.
What it is confused with
A word count is frequently thought to be an exact and indisputable measure. Yet a figure given without stating which tool, which setting and which areas the count covered is incomplete. The definition is always put in writing.