The volume of a translation job is not a single figure; it is a distribution that shows how much of the text is new, how much is repetition and how much comes from the memory. The analysis report presents this distribution in the form of a table and forms the basis of the quote.
Definition
An analysis report is the table that shows the distribution of the text across the match tiers. The tool produces this report automatically. The result is numeric.
Rows of the report
Each tier is shown on a separate row. The rows carry a word or character count. The total sits at the bottom.
The context match row
Matches at the highest level of confidence are shown separately. This row represents the lowest effort. Its coefficient is low.
The exact match row
Identical sentences are gathered on this row. The context has not been verified. It requires review.
Fuzzy match bands
Fuzzy matches are divided into percentage bands. Each band requires a different effort. The division is defined in the tool.
The internal repetition row
Repetitions within the document itself are shown separately. They do not come from the memory. The distinction is preserved.
The new text row
The part for which no match is found is gathered here. The real effort is on this row. The proportion is decisive.
Which memory it was produced with
The report depends on the memory used, and a different memory gives a different result. Which memory was used is recorded. Transparency is ensured.
The effect of the settings
The threshold and segmentation settings change the result. The settings are recorded. Comparison becomes possible in this way.
The effect of preparation quality
A poorly prepared file lowers the match and the report looks poor. Once the preparation is corrected, the rate rises. The order matters.
Sharing it with the client
The report is presented together with the quote. A calculation whose basis is not shown becomes a matter for bargaining. Sharing is standard.
Its readability
The report is a technical table and requires explanation. A table without explanation is not understood. A short note is added.
Explaining the tier definitions
What each row means is explained once. This explanation is not repeated in later jobs. It becomes permanent knowledge.
An input to the time estimate
The distribution determines not only the volume but the time as well. The proportion of new text directly affects the time. Planning rests on this.
In jobs with many files
The analysis is run on all the files together. Repetitions across files become visible in this way. Collective analysis is preferred.
In update jobs
The analysis of the new version brings out the parts that have changed. The scope is narrowed accordingly. The gain is large.
Archiving the report
The analysis report is kept in the job file. When the calculation is questioned later, it serves as the basis. A record is kept.
Timing of the report
The analysis is produced before the quote is given, because it is the basis of the calculation. A report produced afterwards tries to explain a quote already given and is not convincing. The order is not broken.
What it is confused with
An analysis report is frequently thought to be a definitive document. Yet it depends on which memory and which settings it was produced with; the same text gives a different distribution in a different set-up. For this reason the conditions of production are recorded.