Agency docked 50% of agreed price on discovering use of SDL Language Cloud after delivery

I just delivered a translation for which I used SDL Language Cloud to provide me with a first-draft rough translation which I then carefully checked and edited segment by segment to create “my” translation.

Following delivery, the agency cancelled the original purchase order and issued a second one for 50% of the previously agreed price, arguing that I had delivered a machine translation (MT) which they cannot justify vis-à-vis their client.

The agency refuses to accept my argument that I used MT merely as a (legitimate) productivity tool to provide me with a rough skeleton as a starting point for my work.

I would be interested in finding out whether other colleagues out there share my opinion that it is for me as a translator to decide which tools I use, and whether they have encountered similar problems.

I would also be interested in the standpoint of RWS on this issue. I certainly haven’t seen any warnings about being penalised for using the software.

emoji
  • Hi Gerhard,

    I agree with you absolutely - it's entirely your business which tools you use to produce your translation (unless you have instructions to use a particular one, of course, but that's a different issue). What matters is that you deliver a quality translation on time. I don't use SDL Language Cloud but I do use DeepL for the same reasons, namely to give me a draft as a starting point to work with. There are some jobs where I can get the translation done much faster and others where the time saving is minimal. However, even then, DeepL provides me with inspiration and gives me ideas I might otherwise not have had. Like you, I'm fully aware that every segment still needs careful checking and very often editing. These are very useful tools but they're not (and can't be expected to be) perfect. You have to know how to use them properly, and that is your personal expertise that you've built up over the years. And in my opinion that expertise should be recognized and properly paid for. If the agency isn't your bread-and-butter customer, I would be seriously tempted to reconsider your relationship with them.

    That's my 2 cents worth, anyway!

    Martin

    emoji
  •  

    First of all I'd agree with the response from .  However, that's something you need to resolve in working with your client and maybe even review your working practice for this particular client.

    On your specific questions...

    I would also be interested in the standpoint of RWS on this issue.

    I can't answer with an official RWS stance because I don't think we can have one that would cover all use cases.  I think that ultimately these are things you need to have covered in your contract with your client before you start working on a project and then you will know the guidelines and how best to deal with them.

    I certainly haven’t seen any warnings about being penalised for using the software.

    Why would there be?  You could take this a step further and find yourself working for a customer who knows you have a particularly large Translation Memory that might be relevant for the task... and then penalise you for that.  It's not quite the same thing, but it's similar.  Machine Translation needs checking in the same way a fuzzy match needs checking from your TM, and might need correcting, or it might not.  I don't think it's our job to include warnings that the tools you are using could be the cause of you being paid less!

    I think it all comes back to the reply from Martin.

    Paul Filkin | RWS Group

    ________________________
    Design your own training!

    You've done the courses and still need to go a little further, or still not clear? 
    Tell us what you need in our Community Solutions Hub

    emoji
  • Hello    

    I am an old foxGrinning and I long ago anticipated the problem you Wieder faced/are facing, [BTW, Paul, my favourite MT is MyMemory and I am still using an unsigned plugin for my Studio 2022: couldn't you contact Translated and do something together to straighten this out?]

    So, when an outsourcer sends me an *.sdlxliff or an *.sdlppx as a source file, my very first step is opening the language project settings and adding my personal TMs and my MT (under the outsourcer's project memory in case of an *.sdlppx file).  When my translattion is completed, spell-checked and QA-checked, since I am supposed to deliver an *.sdlxliff or an *.sdlrpx file:

    • I go back to the language project settings and delete MyMemory and all my personal TMs (knowing which and how many I used is not the outsourcer's business!);
    • then i go to Batch Mode Tasks/Trados Batch Anonymizer (You must have downloaded & installed this plugin) and I apply it to the project, so all the project files are being processed.

    A piece of advice: when setting Trados Batch Anonymizer, don't check "Anonymize all values", it would look too obvious, just check "Remove/change MT provider" and no further box, it will be enough to make all "AT" vanish from the central comumn of the Editor pane. And so you will spare yourself unpleasant payment quarrels with your outsourcers.

    Best regards.

    Gérard

    emoji