It's been a week since I here and I wasn't able to catch any job yet. In fact I only saw very few of them last week and they were instantly claimed by other translators. Is it a common situation here or it's only temporary due to covid situation? And if it's a lack of tasks in that language pair why are you hiring then?
14 comments
Hi jackal14131, my language pair is English → Portuguese. I'm with the same problem, last month I translated more than 3000 words in my language pair, but I don't know why, in this month the situation is very different, the jobs appear in my RSS but 1 minute late, so it is almost impossible for me to catch a job. And Gengo it is still hiring people for my language pair, very disappointed.
P.S: I write an email to the Gengo support team two days ago about this situation, and they didn't answer me yet :(
Hi everyone! The same problem here. My language pair is English → Russian. In February and March the situation was completely different, I was able to catch some jobs during the day but now it is impossible to get any. Jackal14131, please, tell me, when did you pass your test? Are you from Lionbridge? In my account it is said that the tests in English → Russian pair are closed. I suppose that this "instantly disappearing jobs" issue is connected to the fact that in April too many job seekers had been accepted as new translators.
Same here! Did you get any answer from Gengo Support Leonardo?
I'm facing the same problem. I haven't seen any job in En>Ru language pair for the last year at any time of day or night, neither for Standard, neither for PRO. However I'm getting regular e-mail statistic reports of allegedly existing orders. Do they exist at all or are they some kind of Cheshire cat?
Hello Dennise (and everyone else that has joined this thread).
Thanks for your question about the email statistics reports, as this is indeed something that a lot of people raise. In short, the work does exist. However, while the weekly statistics reports show the total number of units ordered in a particular language pair, they don’t account for how these units are distributed. If, for example, a lot of units were ordered but they were all part of a small number of large collections (rather than a large number of small collections) then this would mean less collections showing on the system. Additionally, the statistics reports don’t differentiate between units that are available to everyone, and units that are available to ‘preferred translators’ only, meaning that if the majority of the units correspond to an end customer that uses a preferred pool then non-preferred translators will not be aware of this work becoming available.
We’d love to be able to improve the statistics reports but as with everything else, this is a question of time, resources, and having the relevant teams available to help refine the way the statistics are displayed. In any case, we hope this answer has at least provided more of a background and as always, we’re always here to answer any more questions that you or anyone else might have.
Hi Katrina,
I thought I should observe the issue more closely before joining the list of complainers, that's why I deleted my previous comment. But there's definitely something wrong with getting English-Russian jobs. They are accepted by some software because they disappear in less than in 2 seconds in 100% of cases, the RSS doesn't help in any way. It can't be done manually by a human being. I suggest you to look fore some solution concerning this situation.
Hello dennise, thanks for commenting again and sharing your impressions of the situation. We do see that a lot of translators across various language pairs mention that collections are claimed very quickly, but part of the reason that collections don't tend to stay on the system for long is that they become available to everyone at the same time, and if there are enough people online then the chances are that another person will claim the collection sooner, either because their RSS feed reader refreshed more recently, or because they clicked on the collection faster. (As described in a help article available here, Gengo's caching feature allows RSS feed readers to place a maximum of two requests per sixty seconds.)
That said, we are aware that there is a possibility that translators might try to abuse the system by using bots or other automated solutions to pick up collections, and this is against the terms of the Gengo Translator Agreement. We'd always urge anyone who becomes aware of suspicious activity by other translators (such as repeatedly picking up collections and allowing them to expire) to report this either to me at katrina.paterson@lionbridge.com or to support@gengo.com. Provided that we have enough evidence to work from, we will look into this and take appropriate action if required.
Hello Katrina,
Thanks for your reply. Unfortunately our freelance activity gives us no chance to become aware and prove other members fraudulent activity, we can only detect this empirically. I think you should agree that no human can check what kind of order he accepts in 1 second. The problem is not the RSS feed and its refresh time, this is 1 second from the time the order appears.
Anyway, I thought what I can suggest as an improvement, considering I'm not that proficient in programming. Maybe there should be some penalty for accepting and then denying the order?
Hi Dennise, thanks for your response too.
These are all very good suggestions, and yes, I understand that as freelancers it is not really in the scope of your activity to be policing the Gengo workbench for bot users. However, as things stand at the moment, unfortunately the main way that we have of investigating suspicious activity is by other translators flagging it, since we don't have an automatic system that detects bot use (although this is something that translators have suggested in the past, and we have taken note of this feedback in case time and resources ever allow us to look into it in more detail).
As for orders getting picked up quickly - I would say that a lot of this stems from the first-come, first-served nature of the platform, which to a certain extent encourages translators to pick up a collection immediately rather than reviewing the source text beforehand. In this sense, creating a penalty for accepting and then declining a collection might potentially leave translators in the situation of working on a text that's not really suited for them, or it might discourage translators from picking up collections for fear of being penalised for then declining them. As this support article states, there is currently no penalty for declining a collection 'once in a while', though obviously doing this frequently, or leaving a long period of time to elapse between accepting a collection and then declining it, is not really best practice. As I say, we do look into the activity of translators if this is flagged and we act accordingly when needed, but in some cases it can be difficult to distinguish between ordinary human behaviour and bot use, which is why it always helps us to have as much evidence as possible when investigating such situations.
I do hope that some of this helps, but please don't hesitate to ask if you have more questions.
99% of jobs are taken from gengo pages before RSS are sent. The solution is Simple and I proposed it many times. The jobs shall be published on the page with 120 second delay to allow rss about available not long taken jobs to be sent. I asked a software developer about caching and he said that in theory caching can be circumvented by various means and obviously somebody is already using special tools to do it. I hope gengo finally implements this delay to allow the translators to see some jobs together with bot users
Another solution to fight bot users would be even a simple captcha while taking job. All websites are using captcha to prevent bots from abusing them but not gengo which has its own way and insists that holy caching works (obviously, it does not for some). And then no investigations, reports from translators, all those efforts will be needed. Just add captcha like everybody else use, the site already has it in login procedure
Hi dropinner, thank you for these suggestions. I've passed them on to the relevant team so that they are able to consider looking into them if and when time and resources allow. For now, all we can really say is to contact us if you find evidence of specific cases, so that we can investigate. We do very much appreciate your ideas (and please do keep them coming)
Hey, guys, the problem is still there! I haven't been able to land a job so far/ They are instantly taken! I have the RSS installed and try to take them right away but...alas. Would love the feedback on that one!
I started having the same problems. I can't exactly say when it started, but I noticed a huge decline in jobs overall and what was more shocking how fast they were being claimed. As some of you have stated, within a second or two. I also find this quite bizarre considering the fact that at time I was actually at the site as the job was being posted and still I wasn't fast enough to claim it. I'm wondering if there is an overall decline in jobs with Lionbridge overall.