What about TM discounts?
During the recent discussion on volume discounts, a number of readers brought up the (very valid) issue of translation memory discounts. I feel a can of worms about to open here, but I’ll inject some of my thoughts on TM discounts.
First caveat: I am neither a huge fan of nor a heavy user of TM. I own and use both OmegaT and Wordfast, and I really like both of them for different reasons. However because my favorite kind of work is translation that requires excellent writing, I only use TM when necessary. I find that because of the “chunked” nature of working in a TM tool, the flow of my writing really suffers when I use one. I do gain a pretty significant speed advantage when using TM, but I find that if the document needs to sound really natural in English, the extra editing time eats up the time I saved during the translation.
That being said, there are some projects where it would be suicidal not to use a TM. I work on several recurring projects for international development entities. These might involve multiple countries submitting funding applications using the same template, so it would be a huge waste of time to start each application from scratch. In those cases, I’m all about TM leverage.
There are two basic sides to the TM discount debate:
- Agency says: if some of the content in this document is recycled from a TM that we provide, why should we pay the translator’s full rate for the recycled portion? The TM is the result of work that we paid for originally, so we want to realize the benefit of it.
- Translator says: I purchased the TM tool and used my own time to learn to use it. I pay for the upgrades and for the computer system that runs the TM. The time that I save by using the TM is a result of this investment, and I want to realize the benefit of it.
Over the years, I’ve struggled with the issue of TM discounts. In some specializations (i.e. software and documentation), the situation may be different from what I experience. But I find that editing fuzzy matches (i.e. anything below 100%) takes at least as long and sometimes longer than just translating the segment from scratch. Therefore, the only TM discount that I offer is that if the client is extremely confident in the reliability of the TM that they provide, I will not charge for 100% matches if I accept them without proofreading them. Meaning that in Wordfast Classic, if a segment shows up green (Wordfast’s color code for a 100% match), I just click Alt-down (Wordfast’s command for accepting a segment) without reading it. I feel that for clients who put in the work to maintain a reliable TM, this is a fair compromise. But because I’ve run the numbers on the time it takes to edit fuzzy matches, I don’t offer TM discounts other than this.
Other thoughts?
Corinne,
I can understand being hesitant about providing TM discounts for someone else’s TM’s, but most of my work involves discounts for my own prior translation work where I have to say, I can’t find much of an argument for charging full rate.
I have also worked on group projects where the TM was a group effort (naturally of varying quality). I lost money on one side, I gained time by using the Trados concordance to look up individual terms. I don’t envy the proofreader, but in a project of over 800K words, consistency takes on increased importance.
The difference of opinion over CAT tools and TM discounts may have more to do with the kind of work than it does a general policy. I have another client that I do not give any discounts, but if I did they would not amount to anything because the texts require a more individual approach.
Thank you for taking up my question as a separate blog post Corinne. Due to my specialization (localization) I am often faced with requests for TM discounts and the nature of the translation often grants this possibility. But I am curious what others think and can agree with your arguments on editing fuzzy matches.
“But I find that editing fuzzy matches (i.e. anything below 100%) takes at least as long and sometimes longer than just translating the segment from scratch.”
I use a similar justification for never giving any discounts for repetitive passages, and there are so many repetitions of the same thing used over an over again in patents that they often sound like the mumbo-jumbo of a Buddhist prayer.
Sometime it is true that making sure that there are really no changes in repetitive passages takes a long time, usually it is not true.
But it’s a great line to use when somebody asks for a discount, which I never give.
Incidentally, I don’t use any TM tools and probably never will. I just finished a long German patent today (28,199 words) and the only memory tool I had were 3 post-it notes stuck on the bottom of my screen.
How I deal with this depends very much on which client is involved, but where some sort of scale is involved, it is usually with my content or content I have approved. Your observations about the time involved in editing fuzzy matches is the same as mine in many instances. So often I find that a 90% match takes longer to adapt than to translate from scratch that I consider a discount out of the question. I look at this technology more as a quality assurance tool for ensuring consistency. Real time savings are achieved consistently only with large blocks of 100% context matches.
Presently, we are living with the consequences of marketing lies introduced and perpetrated by an industry technology “leader” whom I am sure we can all name. Am I saying that an 85% fuzzy match will not ever be a time-saver? No. But in very many cases matches of this quality or above save very little time or cost more time, and we need to document this and make it clear to clients to counteract a decade and a half or so of misinformation.
I am rarely asked to take fuzzy discounts based upon a TM provided by someone else. That is tiresome.
I am commonly asked to provide fuzzy discounts for repetition in my own translations though. I’m currently in the middle of an excellent example: an 80,000 word set of documents with about 50% exact matches (think texts on products sold in different sizes where only numbers change). I’m quite happy to accept 30% of my translation rate for the 40,000 words of repetition in my own work. What is usually means is that as I translate the second or third document I am in fact proofreading the first. I spot a few mistakes from the TM, so I go back and fix them in the earlier doc. It therefore needs much less proofreading than normal.
In the biggger TEnT picture, I love mine for its concordance and term recognition tools, which prevent wasted research and increase typing rate.
I agree with your post even though I’m one of those who (sometimes rather grudgingly) offer/accept discounts for fuzzies/repetitions. The important factor is the text’s subject matter, as you already said.
Asking for a discount for legal texts or literary texts or other creative texts would be stupid. I’d never give those discounts myself.
But product lists, simple user manuals, newsletters with the same introduction for each issue, etc. etc. – that’s a different field. I once did indeed achieve the improvements that certain vendors promise. I translated a lot of similar manuals for very similar products for the same client. And while the first manual might have taken 4 hours to complete, the third or fourth or fifth manual only took 1-2 hours to complete. If I got paid for 3 hours, then I was happy, and my client was happy too because they got a discount.
But before accepting any discounts, the quality of the TM and the requirements of the text must be checked thoroughly. I also have a client whose TMs contain inconsistencies and spelling errors and such – if they want me to use/check those TMs, they’d better pay me.
So, I don’t think there’s a clear “yes” or “no” answer to the question of TM discounts. It always depends on the subject matter, the quality of the TM, and the translator’s skills/preferences.
This is a good reminder that whether or not you charge for fuzzy and other matches, it is in your best interest to always (always) have a fairly clear idea of what you are charging (and earning) per hour on a given job.
I agree with your, Corinne. And there is even one step further of agencies that they do not pay or pay a minimum for those matches in the text without even giving you a TM, which means you have to first translate those matches and then reuse your own translation, which hasn’t been proofread yet!
On the other hand, I don’t believe on not paying for those matching in any case, precisely because, even if the client will give you an excellent TM you can rely on, sometimes the same sentence/word may be translated in different ways depending on the context. As a software and videogames translator, I find myself in this situation many, many times, so even if I have the same word 10 times, it might be that I need to translate it in different ways depending on the context around it. Why should I not receive money for this, if every single time I need to check where that word will appear, if it has variables around, if it is the title of a menu or just a message appearing on screen? That is time that I spend thinking about all those things surrounding the word and that time should be paid.
On the fair side, I also agree on what you said about, well, if you, as a client, provide me with a TM that you know will reliable and you also tell me not to bother proofreading or even reading those 100% matches, well then there I will just “Translate to fuzzy” and allow the program to work by itself, as I won’t receive any penny for those segments. But then the client must accept the consequences of that. Only in that case, I think it’s fair not to receive any payment for those matches.
@Karen: I agree about the “proofreading” aspect of the repetitions being a nice thing, and my policy is very similar to yours on charging that type of content (manuals, data sheets, etc.)
I’ve noticed though that when content is repeated between press releases and product documentation or training materials, the difference in context often calls for a full rewrite, and any discounted rate would be unreasonable. Where there is a good trust relationship with the client I often simply make an estimate of costs if asked to, then adjust up or down afterward to reflect real effort. Often this pays some decent dividends for the client.
Hi Corinne! Great article! I must say I agree with you completely and hardly ever offer any TM discounts, the main reason being that I also spend more time proofreading and checking the text than it would take me to translate it from scratch.
I don’t use TM programmes that often anyway, because most of my work is in the videogames localisation industry, so it’s a very creative type of translation that does not benefit from the use of these particular programmes. If I use them to translate things like the menus, which can be repetitive, I am prepared to offer discounts to reuse my own translations, but as an exception and only if I have a good relationship with my client/it’s a long project that’s going to keep me busy with work for a while. After all, as you said, I bought the tools and spend hours of my own “unpaid” time to learn how to use them. In addition, there are other benefits for the client, such as increase in speed and consistency of my translations.
I work mostly for direct clients and they have no clue to what are CATs, TMs, fuzzy matches, etc. Some clients I charger per word, others I give a closed price. However, if there is a lot of repetition in the document, or it is really similar to previous documents, with identical segments, I usually give a discount and inform them of this. I
Chris Durban pointed out well the importance of knowing how much you are earning per hour. In that sense, if the job is easier, I think it’s only fair to grant a discount; if for some reason it is more complex (scanned pdfs, the source has been translated from another language, and so on), I usually charge a bit more.
I worked with discounts a couple of times, and will never do so again. I am totally against the concept of TM discounts. Even to accept 100% matches (I mean the action of accepting them [highlighted or not])… takes your time (in a long document, you are forced to wait more time, I am sure you have all experienced that). If you proofread on paper (like I do), who pays for the sheets of paper/tonner you use to print, but for which you would receive no money? (May be there is a way to exclude those parts before printing, but then again… that will mean time!). However, the underlying concept I reject has to do with a more respectful attitude towards our own profession.
Comparing T&I with other professions is like an old cliche, I know, but is usually very enlightening. Why doesn’t society give you the power to demand discounts from your dentist if s/he uses composite now instead of the old systems? Isn’t s/he using less time? (When I was a kid, root canals required 3 visits to the dentist, while today they only requiere one. Are dentists paid less?). Why do we consider such an approach unthinkable, but accept this intrussion in our profession?
Even if TM was born in a company to help that company reduce costs, CAT tools are now in the market as a tool for us, translators. (Enough cons has our profession not to profit from this pro!) I see no reason why I should be thinking about my client’s P&L. Does my client care about mine? I am here to deliver quality translations, and my client is here to pay. CAT tools are another tool we have to improve our own work. We should not allow others interfere with our business. Allowing others (clients, TM vendors who happen to be clients, he!) to do so helps consolidate a lethal concept: translation as a commodity. I say translation for translators!
Thanks for opening the can, Connie!
Au
Corinne, Feel free to edit my words and correct your name, please. (Sorry!) Au
Perhaps the simplest solution to dealing with the bugbear of TM discounts (if you’re finding many of your clients insisting on them) is simply to raise your rates. You make the same as before, and the client gets their ‘discount.’ Win-win.
Another idea is to charge a premium to clients who insist on TMs, which can be justified by the points laid out in Corinne’s second bullet above.
I see two possible outcomes, and both are actually positive:
1) The client agrees to the premium/higher rate and you get paid a reasonable fee for your work.
2) The client balks at the fee, looks elsewhere for a translator willing to work for less, and you get to focus on better work for better clients.
If you deliver quality work, clients will continue to buy your services. If not, then maybe they are not the kind of client you want to be working for in the first place.
As usual, Chris Durban hits the nail on the head. Figure out what you want/need to make per hour, and charge accordingly. Seek out clients that are willing and able to pay the rates you require to earn a living.
A TM program is just another tool in the translator’s kit. But it is an expensive tool, requiring costly updates and training. Like Aurora mentioned above, it should be a tool we control, not one that controls us.
We’re not at the mercy of our robot masters.
It has been my experience that TM discounts are usually appropriate, because more often than not having a big and properly maintained TM actually saves time and helps you make more per hour. I provide TM discounts to both direct clients and agency customers.
I also provided my opinion on this subject in my blog last year.