Pages in topic:   [1 2] >
Fuzzy matches - the effects of percentage reductions.
Thread poster: Simon Taylor
Simon Taylor
United Kingdom
Local time: 23:44
Italian to English
Aug 24, 2016

Hello,

First, a confession. I am very new to the world of translation, so this (and subsequent questions) may seem, at best, naive and, at worst, downright silly. But I'm going to ask anyway. You may judge me accordingly.

Fuzzy matches/repetitions and so on.

I understand that there is an expectation in certain quarters for reductions in fees for fuzzy matches/repetitions when using CAT tools and TMs. I've seen various topics covering what percentages to use for the varying closeness of match. However, I've seen nothing on how much these reductions affect what's charged as a result of agreeing to reductions for matches. I've used the ProZ calculator to work out a base rate per source word, looked at the prevailing information on the market, and set my rate accordingly. Reductions for matches are new to me.

E.g. Assume a 2,000 word project and a base rate of 0.08 per source word. An agency requests a reduction of fees based on a sliding scale of match from, say, -30% to -90%. What is the definition of a match? A word? A sentence? Or an entire TU in the CAT? Given the fact that two translators may translate the same word or sentence in two completely different yet correct ways, defining what is and isn't a match seems subjective.

I don't want to price myself out of the market but I do expect to make *some* income. As I wrote at the beginning, this is likely a very naive question but one that is causing me a fair bit of concern. Everything I know about the professional, business side of translation has come form being self taught. I do learn quickly but some lessons are proving a little problematic. Please could anyone give me a pointer or some advice on this topic?

Thanks for taking the time to read this. I hope you can help me.

Simon.


Direct link Reply with quote
 

Yael Ramon  Identity Verified
Israel
Local time: 00:44
German to Hebrew
+ ...
fuzzy matches Aug 24, 2016

Hi Simon

first of all, welcome to the world of words.

to your question: percentage is calculated in a very simple and mathematical manner:
words in a source segment vs. words in a target segment that match them.
Accordingly, a source segment with one word that matches a target segment of one word will be of course 100% match.
A source segment with 4 words and three matching words in the target segment will score 75%.
But this is on the simple side of the issue. For example, if the segment appears several times in the translation memory (TM) and shows different results each time, Trados will choose the higher score, but will reduce the score as penalty for inconsistency.
in short - agencies love it because it reduces their costs. Translators hate it because some times just searching for the difference and correcting it takes longer than simply re-translate the whole segment from scratch.
I do not know your source and target languages, but generally my advice, do not say YES before receiving the files and checking them to see what kind of fuzzies you are dealing with, i.e. do they require a significant effort for corrections (rearranging, adding tags), or just a minor change (comma, dot, Initials...).
Also, if you're dealing with proofreading, it is sometimes better to charge by the hour!!! and not by word-count.

Good luck.
Yael


Direct link Reply with quote
 

Mirko Mainardi  Identity Verified
Italy
Local time: 00:44
Member
English to Italian
+ ...
"Critical business secret" Aug 24, 2016

Simon Taylor wrote:

What is the definition of a match?


I don't know whether this has changed, but see: http://www.proz.com/forum/sdl_trados_support/26239-how_are_trados_fuzzy_matches_defined.html#180525

Basically, what we are paid when working with a CAT tool and a fuzzy matrix depends on some CAT tool developer's "critical business secret". Funny, isn't it?

Also note that increasingly often fuzzies/repetitions are calculated even when there is no TM, based on the internal analysis of a document/project.


Direct link Reply with quote
 

Robert Rietvelt  Identity Verified
Local time: 00:44
Member (2006)
Spanish to Dutch
+ ...
Agree Aug 24, 2016

Yael Ramon wrote:

Hi Simon

first of all, welcome to the world of words.

to your question: percentage is calculated in a very simple and mathematical manner:
words in a source segment vs. words in a target segment that match them.
Accordingly, a source segment with one word that matches a target segment of one word will be of course 100% match.
A source segment with 4 words and three matching words in the target segment will score 75%.
But this is on the simple side of the issue. For example, if the segment appears several times in the translation memory (TM) and shows different results each time, Trados will choose the higher score, but will reduce the score as penalty for inconsistency.
in short - agencies love it because it reduces their costs. Translators hate it because some times just searching for the difference and correcting it takes longer than simply re-translate the whole segment from scratch.
I do not know your source and target languages, but generally my advice, do not say YES before receiving the files and checking them to see what kind of fuzzies you are dealing with, i.e. do they require a significant effort for corrections (rearranging, adding tags), or just a minor change (comma, dot, Initials...).
Also, if you're dealing with proofreading, it is sometimes better to charge by the hour!!! and not by word-count.

Good luck.
Yael



Couldn't have said it better.


Direct link Reply with quote
 

Angela Rimmer  Identity Verified
United Kingdom
Local time: 23:44
Member (2014)
German to English
+ ...
Also depends on the type of translation Aug 24, 2016

I do technical translation, and most of the time a match really is a match, and repetitions are truly repetitions. I don't really understand Yael's complaint about "searching for the difference" as if it's that difficult -- a decent CAT tool will have a clear way of indicating changes so you don't have to go trawling through the segments yourself.

I also disagree that agencies love TM matches but translators don't. Again, perhaps it depends on the type of translation you do, but working with translation memories for me is a must, it makes my work a whole lot easier and faster, and it also gives me more flexibility to negotiate rates and deadlines on projects where the client has indicated that the budget or timeframe is particularly tight.

In my field, at least, it is very possible to offer discounts for repetitions and fuzzies and make a very decent living. But Yael is right in advising that you should always look at the file first before agreeing to anything, just in case there are some nasty surprises there. And also, it is always a good idea to be able to make your own file analysis so you don't just trust what the client is telling you about the word counts. Not that the client is out to cheat you, necessarily, just good to be able to verify the statistics in case the wrong translation memory was used or the client has made some very weird assumptions about what discounts they are entitled to.


Direct link Reply with quote
 

Samuel Murray  Identity Verified
Netherlands
Local time: 00:44
Member (2006)
English to Afrikaans
+ ...
@Simon Aug 24, 2016

Simon Taylor wrote:
What is the definition of a match? A word? A sentence? Or an entire TU in the CAT?


Unless I've missed something, all CAT tools evaluate matches on a per-segment basis (i.e. "TU" as you call it).

Given the fact that ... , defining what is and isn't a match seems subjective.


It's not subjective at all. It's mathematical.


Direct link Reply with quote
 

Robert Rietvelt  Identity Verified
Local time: 00:44
Member (2006)
Spanish to Dutch
+ ...
Disagree Aug 24, 2016

Angela Rimmer wrote:

I do technical translation, and most of the time a match really is a match, and repetitions are truly repetitions. I don't really understand Yael's complaint about "searching for the difference" as if it's that difficult -- a decent CAT tool will have a clear way of indicating changes so you don't have to go trawling through the segments yourself.

I also disagree that agencies love TM matches but translators don't. Again, perhaps it depends on the type of translation you do, but working with translation memories for me is a must, it makes my work a whole lot easier and faster, and it also gives me more flexibility to negotiate rates and deadlines on projects where the client has indicated that the budget or timeframe is particularly tight.

In my field, at least, it is very possible to offer discounts for repetitions and fuzzies and make a very decent living. But Yael is right in advising that you should always look at the file first before agreeing to anything, just in case there are some nasty surprises there. And also, it is always a good idea to be able to make your own file analysis so you don't just trust what the client is telling you about the word counts. Not that the client is out to cheat you, necessarily, just good to be able to verify the statistics in case the wrong translation memory was used or the client has made some very weird assumptions about what discounts they are entitled to.


You wrote: '....most of the time a match really is a match, and repetitions are truly repetitions'.
I dare the difference. I also do technical translations, and my TM's are working perfectly as a dictionary (read Concordance), but after all those years of translating (and thinking you translated it all), I still have to translate from scratch. And if there is a fuzzie match, it takes me a lot of time to figure out what is right and what is wrong. Translating the sentence from scratch would take less time, but anyhow I loose a part of my income because of the grids agencies are using.

Furthermore: 'I also disagree that agencies love TM matches but translators don't.'
Maybe that is because you are standing on the other side of the line. For the rest, see above.

Last but no least: 'it also gives me more flexibility to negotiate rates and deadlines on projects where the client has indicated that the budget or timeframe is particularly tight'.
Good for you, but you could or stick to your rates and tell the client to find another agency or let the translator pay for the difference. A CAT = a tool for translators, and agencies are taking the profit (no offence). I would agree if an agency would pay for my CAT-subscription, but as long as that is not the case, I agree totally with Yael. Do you know the rates they are offering us? It is a disgrace.

[Edited at 2016-08-24 22:14 GMT]

[Edited at 2016-08-24 22:16 GMT]


Direct link Reply with quote
 

Mirko Mainardi  Identity Verified
Italy
Local time: 00:44
Member
English to Italian
+ ...
Legacy TMs Aug 24, 2016

Robert Rietvelt wrote:

Angela Rimmer wrote:

I do technical translation, and most of the time a match really is a match, and repetitions are truly repetitions. I don't really understand Yael's complaint about "searching for the difference" as if it's that difficult -- a decent CAT tool will have a clear way of indicating changes so you don't have to go trawling through the segments yourself.


I also do technical translations, and my TM's are working perfectly as a dictionary (read Concordance), but after all those years of translating (and thinking you translated it all), I still have to translate from scratch. And if there is a fuzzie match, it takes me a lot of time to figure out what is right and what is wrong. Translating the sentence from scratch would take less time, but anyhow I loose a part of my income because of the grids agencies are using.


Just imagine if you're working with legacy TMs and/or on a project with other n translators populating the same TM you're using... :-/


Direct link Reply with quote
 

Robert Rietvelt  Identity Verified
Local time: 00:44
Member (2006)
Spanish to Dutch
+ ...
I did Aug 24, 2016

Mirko Mainardi wrote:

Robert Rietvelt wrote:

Angela Rimmer wrote:

I do technical translation, and most of the time a match really is a match, and repetitions are truly repetitions. I don't really understand Yael's complaint about "searching for the difference" as if it's that difficult -- a decent CAT tool will have a clear way of indicating changes so you don't have to go trawling through the segments yourself.


I also do technical translations, and my TM's are working perfectly as a dictionary (read Concordance), but after all those years of translating (and thinking you translated it all), I still have to translate from scratch. And if there is a fuzzie match, it takes me a lot of time to figure out what is right and what is wrong. Translating the sentence from scratch would take less time, but anyhow I loose a part of my income because of the grids agencies are using.


Just imagine if you're working with legacy TMs and/or on a project with other n translators populating the same TM you're using... :-/


Did you? Did you ever had a look at those TM's? Half or more of the content is mud! You know why? A lot of agencies simply don't care for quality, they only look at their wallet. That is my honest experience.


Direct link Reply with quote
 

Mirko Mainardi  Identity Verified
Italy
Local time: 00:44
Member
English to Italian
+ ...
Me too Aug 24, 2016

Robert Rietvelt wrote:

Mirko Mainardi wrote:

Robert Rietvelt wrote:

Angela Rimmer wrote:

I do technical translation, and most of the time a match really is a match, and repetitions are truly repetitions. I don't really understand Yael's complaint about "searching for the difference" as if it's that difficult -- a decent CAT tool will have a clear way of indicating changes so you don't have to go trawling through the segments yourself.


I also do technical translations, and my TM's are working perfectly as a dictionary (read Concordance), but after all those years of translating (and thinking you translated it all), I still have to translate from scratch. And if there is a fuzzie match, it takes me a lot of time to figure out what is right and what is wrong. Translating the sentence from scratch would take less time, but anyhow I loose a part of my income because of the grids agencies are using.


Just imagine if you're working with legacy TMs and/or on a project with other n translators populating the same TM you're using... :-/


Did you? Did you ever had a look at those TM's? Half or more of the content is mud! You know why? A lot of agencies simply don't care for quality, they only look at their wallet. That is my honest experience.


Agreed. That's why I mentioned it...

Not always, but often, if you really care about what you're doing and the quality of the work you're delivering, you'll end up spending more time double checking those TUs while being paid less for it...


Direct link Reply with quote
 

Soonthon LUPKITARO(Ph.D.)  Identity Verified
Thailand
Local time: 06:44
Member (2004)
English to Thai
+ ...
Joking expeience for JP>TH pair Aug 25, 2016

Simon Taylor wrote:

I understand that there is an expectation in certain quarters for reductions in fees for fuzzy matches/repetitions when using CAT tools and TMs. I've seen various topics covering what percentages to use for the varying closeness of match. However, I've seen nothing on how much these reductions affect what's charged as a result of agreeing to reductions for matches. I've used the ProZ calculator to work out a base rate per source word, looked at the prevailing information on the market, and set my rate accordingly. Reductions for matches are new to me.

Simon.


Regarding my own experience, my JP>TH project paragraphs were broken wrongly. I got report of so many fuzzy matches but they were useless for a translator. It just a fair pricing [reduced price] for the client while the translator needs to rework from scratches. Example [In English so that ones who do not read JP can understand]

[Source text line breaks]
[Original source texts] My revered professor who came to this university recently urged me to further my study in English literature since I have mastered in basic linguistic subjects.
[Source texts with line breaks for Trados Analysis and Report]
My revered professor who
came to this university recently
urged me to
further my study in English literature
since I have mastered
in basic linguistic subjects.

These source texts give many hits, but useless to reuse in translation.

Such the broken segments are troublesome and translators should be beware of!!!

Soonthon L.


Direct link Reply with quote
 

Andy Watkinson
Spain
Local time: 00:44
Member
Catalan to English
+ ...
Gotta luv those Fuzzies Aug 25, 2016

Angela Rimmer wrote:

I also disagree that agencies love TM matches but translators don't.


I think, Angela, what's being referred to is that agencies rely on the fact that most clients are not generally aware of what a translation actually involves, let alone CAT tools, pricing practices, etc...

They'll tell the client that the text actually has a certain number of repetitions and will offer, say, an 8% discount.

Agency imposes 28% discount on translator.

Client happy thinking it's working with an efficient, honest agency.

Agency happy because on top of its usual mark-up it's made an extra 20%

Translator.....not so happy.

At least I think that's the idea.


Direct link Reply with quote
 

Angela Rimmer  Identity Verified
United Kingdom
Local time: 23:44
Member (2014)
German to English
+ ...
Flexible mindset allows you to see the opportunities Aug 25, 2016

Robert Rietvelt wrote:

You wrote: '....most of the time a match really is a match, and repetitions are truly repetitions'.
I dare the difference. I also do technical translations, and my TM's are working perfectly as a dictionary (read Concordance), but after all those years of translating (and thinking you translated it all), I still have to translate from scratch. And if there is a fuzzie match, it takes me a lot of time to figure out what is right and what is wrong. Translating the sentence from scratch would take less time, but anyhow I loose a part of my income because of the grids agencies are using.


Perhaps you are talking about those "fuzzies" that aren't actually fuzzies (and for which I personally do not offer much in the way of discounts. I agree with you that translating from scratch is better if the match is something like only 50%.


Furthermore: 'I also disagree that agencies love TM matches but translators don't.'
Maybe that is because you are standing on the other side of the line. For the rest, see above.

Last but no least: 'it also gives me more flexibility to negotiate rates and deadlines on projects where the client has indicated that the budget or timeframe is particularly tight'.
Good for you, but you could or stick to your rates and tell the client to find another agency or let the translator pay for the difference. A CAT = a tool for translators, and agencies are taking the profit (no offence). I would agree if an agency would pay for my CAT-subscription, but as long as that is not the case, I agree totally with Yael. Do you know the rates they are offering us? It is a disgrace.


I am speaking as a translator here, not as an outsourcer. I DO stick to my rates, but you really are being unreasonable if you are saying that every job is exactly the same and there is absolutely no flexibility to negotiate on anything. If your CAT tool is working for you, then using TMs etc (especially in technical work) should ultimately reduce your workload, which then gives you flexibility to negotiate rates and deadlines. Any potential income you would have had without accepting discounts for matches you regain because you have more free time to accept more work, which compensates for that. I may be an outsourcer, but I'm still a full-time translator, and as a translator, I am telling you that I am making a decent living while also offering discounts for matches. Of course there are conditions attached to that -- I don't offer discounts across the board based on the analysis alone. I offer discounts based on things like whether the TM is trustworthy, whether the matches really are matches, etc. If my workload is reduced as a result of the matches, I will give discounts because it makes sense to do so and ultimately improves my income.

For example, I have a client for which I translated their entire technical knowledgebase. Everything in the TM is mine, I know that for sure. The knowledgebase is huge and we are now at the point where I'm not translating huge chunks of old material anymore, just updates and additions for new products. So now, we have a situation where the very high matches and 99% fuzzies require little to no change. So when that client asked me if I would be willing to reduce costs for jobs that have 10,000 words or more, I was able instead to explain to them that large volumes don't take less time to do and so I cannot offer discounts just because the jobs are big, BUT if they send texts for the Knowledgebase, I CAN reduce the amount I charge to review high fuzzies and complete matches (because I know that personally I am not actually spending much time on those segments anyway) -- as a result, the client felt they were getting a good deal and were happy that I was willing to be flexible, and I was able to avoid a situation where a client thinks that large volumes = discounts. It's win-win.

Let's say a client tells you that have 10,000 words and need the translation in 1 day, and there are seemingly no repetitions. But when you look at the file, you see that there are actually loads of repetitions that the CAT tool didn't recognise because of misplaced commas or new text with just one or two words different each time. Or you did a similar text on that topic recently and have other TMs you can rely on for research and matches. So you know from looking that although the original analysis says that there are 9800 "new" words and 200 reps, you will probably only have about 3000 new words and getting the translation finished within a day is feasible. Other translators may have looked at the original request and thought, "10,000 words in a day, you really are crazy, that is a big NO" but with a flexible attitude, your mind is open to seeing the opportunities for making it happen without having to go out of your way and make big sacrifices in the process.

These are examples of how TMs allow for more flexibility when negotiating with clients, even agency clients.

Forgive me but I wonder if the real problem here is your attitude. You say "the rates they are offering us" and "I lose part of my income because of the grids they are using" as if YOU have no say in the matter, as if the agencies are imposing these things as laws. You have the right to refuse. I get crazy offers, too -- I just tell them, simply, NO. There are other clients out there who do treat translators fairly and pay decent rates.


Direct link Reply with quote
 
philgoddard
United States
Member (2009)
German to English
+ ...
. Aug 25, 2016

I politely tell customers that I give discounts only for substantial quantities of 100% repetition. Nine times out of ten, they accept this.

I've never understood why anyone should buy software not to improve their productivity by doing less work for the same money, but to charge their customers less.


Direct link Reply with quote
 

LEXpert  Identity Verified
United States
Local time: 17:44
Member (2008)
Croatian to English
+ ...
CAT tool marketing Aug 25, 2016

philgoddard wrote:

I've never understood why anyone should buy software not to improve their productivity by doing less work for the same money, but to charge their customers less.


Well, brilliant marketing on the part of CAT makers, for one. Translators initially believe that it will improve their productvity but ultimately cave in to pressure from agencies, who are sold the same software with the promise that it will allow them to pay their translators less.

Though even SDL seems to have abandoned trying to convince now-jaded translators that there are any net productivity gains to be had. At a Trados 2015 presentation I attended last year while in Europe, the reseller didn't even try to convince us we could make more money by upgrading. Instead, they drove home the point - not once but several times - that if we upgraded we would get more work because we could charge less.

CAT tool making is an odd business to be in, for sure. Basically it's a zero-sum game. Anything they come up to benefit one type of customer (translator or agency), results in "money left on the table" for the respective other customer, resulting in permanent conflict between the two. Increasingly it seems like the only way to "win" is not to play at all.





[Edited at 2016-08-25 11:21 GMT]


Direct link Reply with quote
 
Pages in topic:   [1 2] >


To report site rules violations or get help, contact a site moderator:


You can also contact site staff by submitting a support request »

Fuzzy matches - the effects of percentage reductions.

Advanced search







SDL Trados Studio 2017 Freelance
The leading translation software used by over 250,000 translators.

SDL Trados Studio 2017 helps translators increase translation productivity whilst ensuring quality. Combining translation memory, terminology management and machine translation in one simple and easy-to-use environment.

More info »
TM-Town
Manage your TMs and Terms ... and boost your translation business

Are you ready for something fresh in the industry? TM-Town is a unique new site for you -- the freelance translator -- to store, manage and share translation memories (TMs) and glossaries...and potentially meet new clients on the basis of your prior work.

More info »



Forums
  • All of ProZ.com
  • Term search
  • Jobs
  • Forums
  • Multiple search