Translation quality is not in the hands of the service provider alone. The clarity of the source text, the terminology list provided, the questions answered and the time allowed directly determine the output. This is said not to pass on responsibility but to show where improvement is possible.
Shared responsibility
Part of quality depends on preparation on the client side. This fact is explained openly. The aim is improvement.
The source text is decisive
A translation cannot go beyond its source. An ambiguous source produces ambiguous output. The ceiling is there.
Measuring source quality
At the start of the job the source is evaluated and problems are reported. The report is recorded. The expectation is set.
Providing the terminology list
In jobs where no approved list is given, terminology findings rise structurally. The list is requested. The effect is explained.
The consequence of having no list
The translator makes their own decision and the client may later prefer otherwise. A finding arises. The root cause lies in preparation.
Providing the brief
Without knowing the purpose and the reader, style decisions are made by guesswork. The brief removes this risk. The input is requested.
Answering questions
Unanswered questions are closed by assumption. The assumption often turns out wrong. The channel is kept open.
The effect of response time
A late answer holds the job up or moves it forward on guesswork. Both outcomes lower quality. Time matters.
Providing context
Screenshots and reference documents prevent context errors. Attachments are requested. The contribution is concrete.
Allowing realistic time
A schedule that leaves no time for review steps produces findings. The time is planned together. A balance is struck.
Format of the source file
When no editable file is given, preparation effort grows and the risk of error rises. The file is requested. The request is made early.
Finality of the version
Jobs started on a text that is not final produce rework. The source is frozen. The order is kept.
Quality of feedback
Concrete feedback with examples enables improvement. General dissatisfaction cannot be acted on. Detail is requested.
Timing of feedback
A finding that arrives in time corrects the next jobs. A finding that arrives late cannot prevent recurrence. Time matters.
Recording preferences
Accepted preferences are entered in the style guide. The same correction is not requested again. The loop closes.
Single contact person
Conflicting requests from different units produce inconsistency. The channel is reduced to one. A routine is set up.
Separating them in measurement
Findings that originate with the client are marked separately. The picture becomes fair. The root cause is sought in the right place.
Showing them in the report
This data is shared in the periodic report. The suggestion is made in constructive language. Improvement becomes shared.
The condition for mutual improvement
There is a quality ceiling that the service provider can reach alone, and this ceiling is set by the source text, the terminology list and the time allowed; raising the ceiling is possible only when the two sides work together. That is why the data is shared. Improvement is built jointly.
Limits of the approach
This fact does not lighten responsibility for errors that originate with the service. The two fields are kept separate. The limit is openly preserved.