3

Whenever I work on a collection, I always have to correct segments already translated by other translator(s) which are full of all kind of mistakes, as once you grab it you are responsible for the collection’s full content and you must correct (for free) all the errors found.

At this point I ask Gengo if, as it used to happen in Orwell’s “Animal Farm” that some animals were more important than others, also in Gengo there are translators more important than others. Otherwise it wouldn't be understandable that since I began to work with Gengo (about 5 years ago) I have always been penalised for every single comma, for every single space, for correct words that did no like the LS, etc., while the different translators who have worked on the mentioned different segments full of mistakes (errors going around from collection to collection for months) were not penalised or even their mistakes detected by the LSs reviewing them?

The fact is that the theory on which Gengo bases its translation relationship with translators, their duties and rights and the support offered to them by Gengo sounds great. All that theory, though, is not applied in practice because the real fact is that we translators only have duties but no rights and no real support on Gengo's side (I exclude and deeply thank those in Gengo who always are ready to help in moments of difficulty, unfortunately only one or two).

Regarding the mentioned issues, a sincere answer by Gengo (not explanations with empty words, please) would be appreciated.

5 comments

  • 3
    Avatar
    kvstegemann

    This sounds as if you were working on a TM job where parts of the translation were prepopulated from the TM. I am not sure how Gengo creates these TMs, but maybe the TM was not from Gengo but from the end client, so that the end client is also responsible for the bad quality of the TM. Normally I don't take TM jobs because the quality of the TMs is unpredictable, and a badly maintained TM can be a disaster for a translation project. A TM is only useful if it is managed carefully, but since Gengo mostly goes "self-service", things can deteriorate quickly.

  • 2
    Avatar
    Ballmar

    Hello, Kvstegemann, the problem is not TM jobs which are an option that I freely accept. The problem is that when grabbing a collection, it is composed of different segments, some untranslated and other already translated by different translators (not by a TM). The majority of them have different types of errors and these segments go from collection to collection for weeks or even months.

    I’m talking from my experience along the time I’ve been working for Gengo and what I feel is that the fact of having been penalised from day one for every single comma and every single space makes me feel more and more disappointed with the way that Gengo treats translators by understimating their work. I have also come to the conclusion that not all the works are reviewed under the same quality standards: if the LSs were so strict with all the translators’ jobs and their work were reviewed so meticulously as mine, those segments that I repeteadly have to correct for free, would have no errors at all.

    I’m conscious that when people are lowly paid, they are lowly considered, no matter how good professional they could be and how hard they could work: human nature is like that and so is the society we are living in. Consequently, as I don't have the power to change the situation, I accept it. Gengo, though, instead of stimulating translators with more fair evaluations as compensation for the low reward they get for their work, it adds an extra stress on us by the way our work is evaluated, which means that the sword of Damocles is permanently pending over our heads. If 10 reviews with a 10/10 score in a row is needed in order to have our score increased 1 tenth, and the slightest thing would deduce the same (1 tenth) from our score, does it make sense? This is totally unfair and discouraging, no matter these unbelievable theories about “metrics” that someone in Gengo, "with great consideration towards translators", invented to make their life more comfortable (that system may be appropriate for other sectors/situations but not for this one).

    That is why I mentioned that, maybe, as happens in Orwell's book, here some translators are more equal than others...

  • 3
    Avatar
    kvstegemann

    @Ballmar, I am still not sure how this kind of collection occurs. I do not see this very often - if the already translated segments do not come from the TM, I can only assume that another translator has abandoned that collection after working on it for a while. That happened frequently in my first years at Gengo, but at some point Gengo changed the rules and translators will now not be paid if they abandon a collection. Consequently, you should receive the full payment for the collection if you pick it up after someone else abandoned it half-completed.

    I am not aware of any other reason how this kind of situation can arise, but maybe @Lara can shed some light on this. Until then, I think we should not make too many assumptions.

    Apart from that, I don't worry too much about my score. It's just a number. As long as it is sufficient to get the pro jobs, it has no other consequences - you won't get more or better jobs if your score is higher, the only thing you need is a quick finger on the mouse. So why bother?

  • 2
    Avatar
    Lara Fernandez

    Thanks for trying to explain, @kvstegemann, I appreciate it :)

    @Ballmar - like @kvstegemann stated, what you're speaking about is TM (Translation Memory) applied by the system to said jobs, hence why you see the same translations "going around from collection to collection for months" (in your own words.)

    I already explained this to you in detail if email, if not last year, then the previous one, but for the sake of clarifying everything for anybody else who may come across this post, let me repeat it below :)

    At Gengo, there are only three instances in which you may see and be able to edit the work of another translator:

    • Edit jobs, in which you will be expected to proofread and edit a translation submitted by another translator. You will be paid in full according to Edit rates.
    • Declined jobs, which a translator worked partially on, but then chose to let go. In this case, the first translator has not submitted the job to the end customer and has not received payment for it. You are not required to use their previous work for reference, and can delete and start from scratch. You receive payment in full for these jobs.
    • Rejected jobs, which a customer has rejected and which have been reopened to the pool. Again, the first translator has not received payment for this job (and has likely received a low score for it, which guaranteed the rejection), and you are not required to use their previous work for reference, and can delete and start from scratch. You receive payment in full for these jobs.

    If what you were referring to were Edit jobs, and you find that the quality submitted by the first translator is not up to par, you can always flag the job and contact support to report it, so we can investigate. 

    What you are most likely referring to, however, and especially taking into account that you mention the same segments "going around and around" is simply translation memory. Of course, translation memory was saved at some point based on the translation that someone submitted in the past. However, it is applied automatically when the system detects a match, and please note that matches may vary in percentage as explained in this article. As such, the TM applied by the system may fit perfectly within the text, or may require some correction and adaptation to the new context. Please also note that, depending on the match percentage, you are still being paid for these words (though the rate may vary) so you're not really correcting it "for free." 

    With that out of the way, please note the following:

    • TM is saved at the time of submission of a translation. That doesn't mean that the job containing the initial TM has never undergone a GoCheck or that, if there were errors indeed, the translator hasn't received feedback about it. However, because all this happens after submission, it wouldn't be reflected on the TM.
    • Because matches can be fuzzy, sometimes the TM identified by the system is not necessarily a 99% match and can be as low as 75%, which means that naturally the translation provided may not match perfectly and may need some editing, which you are required to do, as you're receiving payment for it.
    • As per our TM FAQ article:

      Who is responsible for bad translations within a TM?

      Translators are responsible for ensuring that they only use accurate matches. You'll need to make sure the final translation you submit to the customer meets all the quality requirements.

    Please also note that if you're repeatedly seeing bad TM translations, you have the option to flag them (also explained in the TM FAQ article).

    If you need further explanation on how TM works and how it affects you, I'm happy to explain anytime. Also, if you suspect that that's not the case at hand, please feel free to send me a screenshot (along with the job and collection number) of any of the jobs where you see these pre-filled segments so that we can look into them and provide the most accurate explanation for that one specific case.

    Thanks, and have a great week! 

    Lara

    Edited by Lara Fernandez
  • 0
    Avatar
    Ballmar

    Thanks for your comments, Kvstegemann.

    Lara, as always, thanks for your time, and a nice week too.

Please sign in to leave a comment.