<<“In experimenting to see what worked and what didn’t,” says Mariliam Semidey, 37, who served as Viki’s senior manager of customer and community experience from July 2016 to June 2021, “we might’ve made some mistakes.” Time and time again while Semidey worked there, the question came up during corporate meetings: Should Viki get rid of its volunteers?>>
5 Likes
I may have used for you the wrong word as ‘‘investigation,’’ but ironically you did just what I wanted to be done, and I don’t know… what word you would have used? I could also call it research; which you did a fantastic job for what I read so far, but sadly you stopped at the good part. I wished someone cared enough to stop these unfair distribution of volunteer work by exposing the systematic way they ‘‘hog’’ projects.
By breaking these unfair pattern of behavior that has been so prevalent for so many years here at Rviki, and by as I suggested making a ‘‘blocking’’ system that can prevent ANYONE from ‘‘hogging’’ projects, we can finally have a fair distribution of volunteer/paid work for ALL/EVERYONE that is willing, and practically begging to be given a chance to offer their great services as volunteers.
The Quality of the subtitles will be maximize in high volumes without ever sacrificing ‘‘Quantity’’ (speed of services) since better qualified volunteers will be able to work at a faster rate of time; unlike SOME of these hoggers that mainly work with translation tools that requires working ‘‘extra time’’ slowing down the availability of the subtitle in the drama/movie etc., frustrating the viewers, and taking away the chance of acquiring more paying subscribers since they would be much more satisfied with the improved quality in translations in any OL.
nectarine_tart
This system has been developed and maintained by Viki, either through conscious intention or negligence or some combination of the two.
That’s when I realized that ‘‘viki’’ is run and ruled by the controlling group we have here, and NOTHING will ever be done to change these SO unfair practices. Now I know why they talk about ‘‘viki’’ like ''its a ‘‘friend’’ a ‘‘she’’ ‘‘their buddy’’ bc that ‘‘viki’’ ‘‘have their back covered.’’
1 Like
I think they need to get rid of the contribution count reward method that makes some volunteers ‘‘greedy’’ for numbers, and they want to ‘‘hog’’ as much as they can, to get those numbers up. They resort to creating control groups that ‘‘hog’’ most of the work done by their ‘‘buddy system.’’
They need to realize that SOME paid subbers although may be ‘‘faster’’ workers; not all of them are providing quality work in their subs. They need to concentrate on HOW to provide Quality and Quantity work at the same time by creating a system that don’t keep abusers/translation tools users For ALL of their translation working as volunteers. Those type of volunteers are not productive for this site.
WE are Human and we might need to use a translation tool for a word or two, but if they don’t know the Language proficient enough they have no business offering a volunteer work that will lack quality, and will affect quantity since is time consuming.
Getting rid of volunteers won’t solve the problem because some PAID subbers have given crappy work here too. I have seen some subtitles that paid subbers have done, and they leave a lot to be desired when it comes to Quality work. AI? GOD forbid they grow on that idea since they will make the biggest mistake of their life since AI subtitles in my opinion are the worst ones to have in dramas/movies (unless we have great honest editors that know the OL proficient enough to provide the viewer great quality editing).
When RVIKI allowed controlling groups of volunteers (paid/non paid), to close/stop the availability of having more qualified volunteers to be able to work here at RVIKi: THAT’S when they need to rethink; WHERE did Viki go wrong when implementing the volunteer program? Who is the main culprit here? IN my opinion the ‘‘hoggers’’ that have taken ‘‘control here’’ like they ‘‘own’’ the place.
2 Likes
I think they’re understaffed. We’ve seen a lot of Viki employees leave and not be replaced. If it’s not that, then it’s just laziness.
In any case, Viki doesn’t keep its promises. Once again, they don’t respect their own guidelines. We’re becoming more and more of a closed community and Viki is only interested in taking advantage of us. We used to have a reputation for having the best subtitles, but that’s no longer the case.
Translating from English has become difficult, with subtitles that make no sense. I don’t understand why the episodes are being released when the English edition is clearly not finished. Nor do I understand the point of making viewers read incorrect subtitles.
People are never satisfied anyway. They’re the loudest, but I don’t think they represent the vast majority.
Viki chooses to listen to those who pay a subscription as if the work of volunteers didn’t make money or attract subscribers. 
8 Likes
Viki asks this from CMs. They tell them not to have a TE (without whom even the best English editor, if s/he doesn’t know the source language, risks leaving mistakes in meaning), the role of the GE is all but gone and very often not even a CE is allowed!
And when they do allow you a CE, they tell you to release to other languages immediately as soon as the subtitles are translated, before the CE starts working. (Sometimes it happens that not even segment Quality Control is allowed!) If the English editor is quick enough, the editing will be done while the translation in other languages gets done. The translators may not have the best version, but by the time the OL editor comes, the English edition will also be complete. Therefore there can still be a good outcome in a short time.
The viewers often don’t care for excellent subtitles, they just want to know what happened in their favourite series. Even before, when we had proper TE, GE and CE, they often chose not to wait, they watched right away, and even when it was only partly translated and missing pieces here and there!
5 Likes
I think the CM could be selected using a simple algorithm that would collect important/pertintent data from each applicant’s background in an easily digestible form. This could include factors such as experience, when they were CM for the last time, how many parallel projects they have (including editor, moderator, subber, segger) with preference for people who seem to have more free capacity, and prioritizing a healthy rotation of capable individuals. There should be maybe a quota for newbies who will be assisted by more experience co-CMs?
My point is that if they don’t have staff for this, this could be done by a program and Viki staff would just need to check if what the program comes up with is OK.
4 Likes
Those releases happen after QC. When the segmenters and CS are done, OL is allowed to translate and at the same time, the English team, whether it’s one person or more, starts editing. Then when the CE is finished, OL moderators/editors can start editing.
Of course, there are also cases where the CM (or Viki?) doesn’t even bother to hire a CS…
This may work for the big teams, but if you are translating on your own, you’d rather wait with translating until the English subtitles are edited.
At least the ones who complain about the speed don’t. I guess most of them wouldn’t even recognize an excellent subtitle. I on the other hand can get so annoyed by bad subtitles that I switch to subtitles in another language. If that isn’t an option, I’d rather wait longer to watch a show.
And now, even the percentage of subtitles is more important than the subtitles themselves. At least on one show (and I’m sure it won’t be the last one), we are now supposed to immediately fill up any segment we might create cause the percentage of subtitles are not supposed to go below 100%. We fill them up with an invisible character (for both presubbed languages) so it doesn’t make a difference for the viewer experience; it’s just that no viewer can complain about missing subtitles this way. Extra work for us, with no actual benefit.
So bad subtitles or no subtitles, it doesn’t matter to Viki as long as the viewers don’t (seem to) have reason to complain.
I think the problem is that there is only 1 person in charge of choosing CMs. This person may not only be busy but also biased. I don’t think the choosing is so much work that they can’t do it at all. After all, the most unfit options (those who are not QC yet) have already been filtered out by the system, and there are no application letters to read. It just might make the process a bit more fair if the decisions were taken by a (small) group of people instead of just one.
6 Likes
I know that. I was asking why Viki ask to do this. In the end, wrong subtitles spread to all languages. You say it’s because of viewers. Why does their opinion count more than ours?
This isn’t always true. On some projects, we’re told that the English edition has been done, not that they’ll do it later.
4 Likes
That’s because Viki doesn’t worry about all projects equally. The more popular the show, the more chance there is that viewers get impatient and they want to prevent that from happening. Also, some CMs just release before editing because they think it’s the norm nowadays while others do things the old way unless they get specific instructions not to.
6 Likes
While this is true, the person choosing the CM has to consider, say, 180-200 applications. How do they make a decision if not looking at the applicants’ backgrounds? That involves looking at all those applicant profiles. That is a lot, I think, for 1 person UNLESS… unless the information they need is presented to them in a spreadsheet, for example, or if there is an algorithm using selection criteria (I suggested a few above), which presents the results in an easily digestible way.
4 Likes
The choices they make don’t give me the impression that they thoroughly investigate everyone’s background. It’s more like, “oh, that person was fast on their last project (which was not long ago so it’s easy to remember); let’s give the new channel to them.”
Whether they actually did a good job in the true sense of the word, whether in case they didn’t, it was actually their own fault or not, or whether there might be other great options is irrelevant to Viki. They just need the guarantee that the viewers will stay satisfied.
7 Likes
I know. Actually, I said the same a bit above. To clarify my point: If the sole reason Viki lacks a genuine application process for CM positions is due to insufficient staff to manage it, and if given the opportunity, they would embrace a proper process, then the solution I proposed seems feasible: a streamlined collection of applicant data in a format that’s easily manageable and reviewable.
4 Likes
For that Viki would first have to have some quantifiable database of its QCs. Which they don’t.
Additionally, some things you mentioned are neither quantifiable nor predictable (such as availability and work capacity).
5 Likes
What makes you think so? They operate a website where all this data is available.
Why not? Availability is perfectly quantifiable and Viki already quantifies it with moderator slots. These slots could be extended to all roles. For example, in this imaginary spreadsheet or other kind of database, they could see the current roles of a given applicant: CM - 1, Editor - 2, Moderator - 5, Subtitler - 18. Whatever. This will give a clear picture of who has accumulated a lot of roles, so Viki could find a candidate with more capacity, say, CM - 0, Editor - 0, Moderator - 2, Subtitler - 4. Or fewer. I don’t know. Capacity is CURRENT ROLES, Experience is PAST ROLES.
2 Likes
Do you remember how Moderator slots were chosen?
They asked us. They asked what would be a fair number of projects per contributor. Some said 20-25. Some have the capacity for that. They didn’t do any fancy calculations (which they could have).
Yes, Viki could gather QC statistics to make an algorithm of any sort. But this is a website which struggles with server space to the extent that inboxes of more active volunteers are permanently under “Too many requests” error. And too often than it should, the whole site is affected my the same error.
Again, Viki could indeed gather QC statistics, but they are refusing to for YEARS. That’s how long we have been asking for some sort of contributor profile which would have basic calculated info/coefficients included so that the CMs and Mods could much more easily choose members of their team.
Availability is not actually quantifiable because with many average QCs it fluctuates as their own personal and professional life allows them.
Viki couldn’t even automatise when a CM has “completed” their project because completion is relative. Is it when the English team is composed and ready to start? Is it when Moderators of major languages have been recruited? Is it when the English team releases their final episode to the OLs? What about projects which come presubbed to English - are they inherently completed from the very beginning?
Sorry I’m being so negative. It’s not you, nor your suggestion. I think I just spent too much time on this site and I have little faith in Viki’s computational agility.
The first and foremost thing we need are rules, rules and rules on how a CM is chosen. That would have to be the start of your algorithm anyway.
9 Likes
I think Viki managers do not have a clue about the quality of subs in other languages. Let us say I am a manager: what tool do I have to check that the subtitles in Arabic or Indonesian are correct? No one. So they just do not care. Implementing a verification system would require a Copernican revolution in the office that deals with volunteers. The system initially made sense: a series of cascading edits that at least ensured the subtitle passed through many eyes. In Italy we have the triple edit standard, but if a moderator wants to make just one edit, the system allows it. On the other hand, the existence of the edit is only on word, it is not a work that can be verified: I could take a film, distribute it to the subbers, wait for them to finish and move on to something else, without ever even going to look at what they have written. I think you know that a Serbian moderator subtitled an entire drama by putting three dots, and was accidentally discovered. The only system Viki has to verify the quality of translations is user feedback. But if users simply write a comment in their language at any point in the drama, who will read those comments? Nobody. Also in this case we would need automatic data collection with a format on which to evaluate the quality of the translation from 1 to 100: but then we would need data analysis work, and even in this case no one does it. There is already something similar, with five stars, I would like to see who checks those feedbacks. So we write in the wind.
7 Likes
Even if they do, Viki has no way to verify the language skills of the viewers. The ones who complain the most are the ones who prefer bad subs over no subs so how well-educated could they possibly be?
Would you want to use such system as a guideline or as the sole decision maker? I do agree with @bozoli that not all factors that play a role in why a CM should or should not be chosen could be predicted this way, which would make such a system as unfair as the current monopolized system. What could be done, in addition to the already-existing automatic exclusion of those who have not yet reached QC status, is to implement a blocking system to filter out those who have been CMs in the last 3 months or so. That way, at least we would get more variety of CMs and as a result, probably also of moderators, subbers, English editors, CSs, and segmenters.
8 Likes
Yes, you mentioned this before, and it baffled me then and I’m not sure I understand it now. Let’s suppose I am a Viki staff member in charge of selecting a CM for the next drama. I look at an application and check what’s going on for that applicant at the moment. Not over a period of time. That exact moment. It’s like a snapshot. If I see that person has enough on their plate, I rank their application lower than the others.
Yes, I agree that in my proposed system there would still not be any quality checks. I don’t dare to dream that big.
I also think that Viki will never introduce any form of user feedback on the quality of translations. They don’t want to open a can of worms. They don’t want to anger volunteers. Not even that legendary 3-dot person.
I don’t think I ever aimed that high. I just want a system that would ensure rotation. For me, that would be a hundred times better than what we have now.
4 Likes
@mirjam_465
"We fill them up with an invisible character (for both presubbed languages) so it doesn’t make a difference for the viewer experience; it’s just that no viewer can complain about missing subtitles this way. "
Sorry, but filling a segment with an invisible character is a violation of Viki contributing guidelines. https://contributorhelp.viki.com/hc/en-us/articles/18459867909523-Viki-Contributor-Guidelines " Responsibilities of a Subtitler. No inappropriate subtitling, which may include but is not limited to: Adding only HTML or punctuation marks into an empty segment." Although practiced frequently in order to give the false appearance of 100% complete subtitling, this behavior is a ground for reducing the number of contributions and even disqualifying the user completely.
Under “Manage video”, the All Languages moderator and the channel manager often see numerous languages at 100% when English is less than 100%. Far too often “other language” moderators or subbers put in a musical note or a period or other character and nothing more in the blanks cut for OST lyrics which haven’t been subbed in English yet.
The sad thing about the focus on 100% is that a misguided channel manager would prefer a wrong or plagiarized subtitle or html rather than a blank to reach 100% or that the channel manager won’t allow other languages to start subtitling if English is not 100%. (Unfortunately, lack of judgment is not a viki criterion for non-selection as a channel manager or moderator).
While good segmenting is the foundation of good subtitling, too strong a focus on the 100% puts determination of what video footage needs a subtitle in the hands of segmenters and/or a chief editor who may not know the source language well enough to realize the blanks are superfluous.
2 Likes
I remember doing it once, but it was very different from the cases you mentioned. I was Italian moderator on an old drama, and the CM couldn’t be easily contacted. There were numerous subtitles with “descriptions” of what was going on, such as “he put it in his pocket!” or “he’s shouting now”. Things like that. It’s not that this subber was doing it as a favour to deaf people, because it was only once in a while, not all subtitles. And some of those descriptions were of a visual action, not a sound. One of them was only a dot!
I refused to translate this useless and ridiculous stuff. But they were still too many per episode and the percentage of completion in Italian had dropped to 98%, although there was no dialogue missing from my translation!
I think I messaged the CM, but I don’t remember what happened. Also because after a little while the show sadly lost its licence, before I could even finish it…
Another one was Beethoven Virus.
Once upon a time subbers had a wild ride, it was the Far West times of Viki. Now, no one would dare do such things - at least I hope not.
So yeah, there are exceptions to every rule, I suppose.
Sorry for the slight deviation from topic.
5 Likes