![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
The promise of work...
May. 2nd, 2005 11:00 pmMore jobs came in today, which is good, though I devoted mucho time to considering carefully worded responses to people wanting to know about my rates, and who are particularly curious about discounts for volume and for using translation memory tools such as Trados. I am not the only one troubled by such issues; they crop up frequently on virtually every translator forum and discussion board.
The client's side is fairly straightforward: We're offering you a lot of work, so you chould give us a volume rate. Also, when using a translation memory tool, some sentences may crop up that are either very similar to or, indeed, identical to ones you have translated before, so you should accept less money for such text. (In some instances, when a client provides an existing database, the translator may even be faced with using similar/identical sentences translated by others.)
There are almost as many responses from the translator's side as there are translators.
Sure, it may make sense to offer what I think of as a token discount for volume, but the problem is that a lot of clients aren't looking for a 5% discount. Heck, I've had some prospective clients ask for a rate cut of between 30% to 50% of my normal rate for "volume" work. My gut response is to think: "I can't afford it."
The problem from the translator's side is this: the effort to create a "unit" of translation - be it a word, a sentence, a certain number of characters, or a standard line - does not really benefit from any economy of scale. It takes roughly 100 times as much effort to translate 100 units as it does to translate one. It's not the same as the situation for manufacturing a physical product, where the effort required to make the 100-th widget is a fraction of the effort required to make the first.
A different argument attaches to variable rate translation using translation memories. The idea of a variable rate is that the translator is paid "in full" for units that are below some cutoff threshold of "similarity" (as determined by a sophisticated pattern-matcher in the translation memory program), meaning that any such new sentence is sufficiently different from any existing sentence in the data base that it is, essentially, considered a completely "new" sentence for translation.
Under this scheme, you get paid some fraction of the full rate if a new sentence is determined to be between, say, 80% and 90% "similar" to a previously translated sentence, the idea being that you merely have to "edit" the previous translation to turn it into a perfect match for the new sentence. Similarly, you get paid a lesser fraction if the new sentence is between 90% and 99% similar, on the theory that you have less to massage. Finally, you are paid a minuscule fraction (or perhaps nothing at all) if a sentence is identical to a previously translated sentence, since - in theory - there's really nothing for you to do.
There are several wrinkles in the wallpaper here. First, unless the translation memory has been reviewed or is one that you trust because, say, it is composed of your own translations, it is likely to be, um, defective. Once, I was considering taking a job once where the client scale provided zero payment for 100% matches, whereupon I asked them to confirm by email that - pursuant to their policy - I was to pay no attention to 100% matches in the course of the work, whereupon I was informed no, that wasn't what was intended: in case of a 100% match, I was expected to verify that the matched sentences were accurate and make any necessary changes; it was just such changes wouldn't be tallied when it came to getting paid.
You can imagine the resulting impasse, but I digress... (and didn't take the job).
Second, although massaging less-than-100% matches is often likened to editing (which traditionally pays less than translation), it is editing of a very special, bad kind: making corrections to what, from a human, would be considered an incompetent translation. (Consider that a rate of one "major" translation error per page is viewed as mediocre work by many clients, and now consider a page of "95% match" translations. Assuming 10 such sentences per page, that's like editing a translation with at least 10 major translation errors per page.)
And now, the kicker. If a client gets to use the resulting translation file from the time you submit the assignment until the end of geological time, it stands to reason that the client will be able to reduce future translation expenses even further. Now, since you are - in effect - providing more than just a translation, shouldn't that logically command a somewhat higher price, just as there is (or should be) a difference between selling "first North American serial rights" to a short story and "all rights"?
Yet in such cases, clients often wish to combine the best of both worlds and get a translation memory at a discount!
Now, I am perfectly aware of how the market operates and of what some translators will do to get work. Some will cut their rates to the bone to undercut the competition. Most will "go along with the crowd" and agree to deep discounts and discounted variable-rate TM translation simply to get work.
All is not gloom, however. I know of at least three translators who periodically raise their rates to a point where the income gained from the higher rate balances the work lost due to the higher rate (the exact opposite effect, BTW, to what happens when granting a discount).
Me, I'm just trying to find my way.
Cheers...
The client's side is fairly straightforward: We're offering you a lot of work, so you chould give us a volume rate. Also, when using a translation memory tool, some sentences may crop up that are either very similar to or, indeed, identical to ones you have translated before, so you should accept less money for such text. (In some instances, when a client provides an existing database, the translator may even be faced with using similar/identical sentences translated by others.)
There are almost as many responses from the translator's side as there are translators.
Sure, it may make sense to offer what I think of as a token discount for volume, but the problem is that a lot of clients aren't looking for a 5% discount. Heck, I've had some prospective clients ask for a rate cut of between 30% to 50% of my normal rate for "volume" work. My gut response is to think: "I can't afford it."
The problem from the translator's side is this: the effort to create a "unit" of translation - be it a word, a sentence, a certain number of characters, or a standard line - does not really benefit from any economy of scale. It takes roughly 100 times as much effort to translate 100 units as it does to translate one. It's not the same as the situation for manufacturing a physical product, where the effort required to make the 100-th widget is a fraction of the effort required to make the first.
A different argument attaches to variable rate translation using translation memories. The idea of a variable rate is that the translator is paid "in full" for units that are below some cutoff threshold of "similarity" (as determined by a sophisticated pattern-matcher in the translation memory program), meaning that any such new sentence is sufficiently different from any existing sentence in the data base that it is, essentially, considered a completely "new" sentence for translation.
Under this scheme, you get paid some fraction of the full rate if a new sentence is determined to be between, say, 80% and 90% "similar" to a previously translated sentence, the idea being that you merely have to "edit" the previous translation to turn it into a perfect match for the new sentence. Similarly, you get paid a lesser fraction if the new sentence is between 90% and 99% similar, on the theory that you have less to massage. Finally, you are paid a minuscule fraction (or perhaps nothing at all) if a sentence is identical to a previously translated sentence, since - in theory - there's really nothing for you to do.
There are several wrinkles in the wallpaper here. First, unless the translation memory has been reviewed or is one that you trust because, say, it is composed of your own translations, it is likely to be, um, defective. Once, I was considering taking a job once where the client scale provided zero payment for 100% matches, whereupon I asked them to confirm by email that - pursuant to their policy - I was to pay no attention to 100% matches in the course of the work, whereupon I was informed no, that wasn't what was intended: in case of a 100% match, I was expected to verify that the matched sentences were accurate and make any necessary changes; it was just such changes wouldn't be tallied when it came to getting paid.
You can imagine the resulting impasse, but I digress... (and didn't take the job).
Second, although massaging less-than-100% matches is often likened to editing (which traditionally pays less than translation), it is editing of a very special, bad kind: making corrections to what, from a human, would be considered an incompetent translation. (Consider that a rate of one "major" translation error per page is viewed as mediocre work by many clients, and now consider a page of "95% match" translations. Assuming 10 such sentences per page, that's like editing a translation with at least 10 major translation errors per page.)
And now, the kicker. If a client gets to use the resulting translation file from the time you submit the assignment until the end of geological time, it stands to reason that the client will be able to reduce future translation expenses even further. Now, since you are - in effect - providing more than just a translation, shouldn't that logically command a somewhat higher price, just as there is (or should be) a difference between selling "first North American serial rights" to a short story and "all rights"?
Yet in such cases, clients often wish to combine the best of both worlds and get a translation memory at a discount!
Now, I am perfectly aware of how the market operates and of what some translators will do to get work. Some will cut their rates to the bone to undercut the competition. Most will "go along with the crowd" and agree to deep discounts and discounted variable-rate TM translation simply to get work.
All is not gloom, however. I know of at least three translators who periodically raise their rates to a point where the income gained from the higher rate balances the work lost due to the higher rate (the exact opposite effect, BTW, to what happens when granting a discount).
Me, I'm just trying to find my way.
Cheers...
