In this discussion, we ask the question that everyone wants to know in relation to their translation needs: how long does a translation take?
Ian:Let’s begin this topic by outlining some of the basic timeline parameters, how many translation words per day is realistic?
Sonja:When a human translator is working on a project, we have a time budget in mind of around 2,000 words per day, or about ten pages of A4 text. Some translators are faster, and some are slower so of course there’s variation, but this is a reasonable baseline.
One of the most significant factors that impacts time is the difficulty of the text. If a piece of text is very coherent and flows nicely a translator can get into it and establish a good pace, getting through as much as 2,500 – 3,000 words in a typical working day. If a piece of text is rather disjointed or very terminology heavy, which requires the translator to do quite a bit of research to determine appropriate terms, the productivity rate will be much less, perhaps only 1,000 words in a day.
Ian:How do other things like proof reading have to be factored into a project’s ‘time budget’?
Sonja:The vast majority of our translation projects include and allow for a proofreader’s time. A proofreader generally gets through 4 to 5 times as many words as a translator in the same time period. This ratio is of course affected by the quality of the translation; the proofreader may find herself having to make considerable edits and amendments to the text which will decrease her productivity.
The single largest time requirement from a project management perspective is the time required to put together the best suitable translation team for each project.
As you’ve discussed previously with Tim and Kristin, when the translator finishes her work on a job, the proofreader doesn’t immediately begin her work on the text. There’s almost always a gap between the translator finishing her work and the proofreader beginning hers as they are almost certainly already working on another project. So, while the proofreader can get a lot more done than a translator in a comparable time period, there is a lag between the two.
Ian:Are there any other factors that have an impact on the turnaround time for a translation?
Sonja:There’s a file preparation stage in which we have to get the original content as supplied into a format that enables the translator to access the translation memory and the term base at the start of the project. At the end of the project we have to get the content back in to the original format that it was supplied in, we check that answers to queries have been implemented and go through a final QA check to see if the customer specific terminology has been used and tags and formatting are right. And, of course, check the final document to see if the layout or code adjustments are correct. All this takes time.
The single largest time requirement from a project management perspective is the time required to put together the best suitable translation team for each project. We work regularly with something like 500 translators in a typical month and the assignment process itself can take some time. It can also vary depending on the time of year, as there are fewer translators available during summer and Christmas holidays and everybody else is extremely busy during these dates.
If other services, such as multilingual DTP, are needed then another workflow stage is added to the timeline, as the DTP engineer needs to check that the layout is correct and appropriate for each of the target languages.
Another additional stage can be Client Review, and this can be very time consuming, too, as typically you’re asking someone to review the text who already has the responsibilities of his day job. So, there are a number of things other than the translation and proof reading that have to be considered in the delivery schedule.
Ian:If the base line is 10,000 words translated in a week, barring complications, what options are available if a client requests that a job be completed more quickly than this?
Sonja:The feasibility of getting multiple translators working on the text at the same time is the first thing we’d look at in that circumstance. This is something we do on a fairly regular basis when there’s an urgent need for a particular translation. There are complexities involved in working this way because people obviously will translate the same text in different ways, so you need to employ a robust approach to terminology to ensure that different translators are using the same terminology.
When a customer needs, say, 30,000 words translated within days rather than weeks, that breaks the translation words per day “speed limit”. It’s a request which is almost impossible to fulfil using human translators, so machine translation with human post-editing could be the answer.
A key aspect when using multiple translators is to use one proofreader going through all the work and regimenting any discrepancies in style and terminology. So the proofing person is essentially functioning as an editor. What we want to achieve eventually is that the translation reads like one person wrote it.
It’s much more difficult working this way and there’s always a level of compromise when trying to do a translation very fast with multiple translators – it’s not how to get the very best result but sometimes it’s what we have to do to get the job done.
Ian:How can Translation Memory help with requests for expedited translations?
Sonja:What may start as a 10,000 words translation may actually end up being only a 5,000 word ‘new’ translation based on Translation Memory matches from previous translations and repeated text within the document. Translation Memory can play a critical role in transforming what may look at first sight like an impossible job into something which is very doable.
Terminology is key, too; when you’re working at speed having a robust terminology process really helps with getting the job done very quickly without compromising on quality.