Google Takes Pride in Not Paying Translators
08/01/2008, 4 weeks ago
Sometimes Google simply puzzles me. It is a multi-billion dollar company, right? It has all the resources to venture into absolutely any new field - by either investing in internal development or buying all the companies it thinks could complement its other products or even launching a venture fund of their own if they so choose.
And now I look at the latest post on the Official Google Blog and one thing comes to my mind: Google actually takes pride in its ability to crowdsource parts of its own work and never pay for them. The post describes how satisfied Google is about the fact that Google homepage is now available in more than 100 languages thanks to the efforts of volunteer translators around the world.
The latest addition to the family is Google's homepage and search interface in Maori language that has recently been celebrated in New Zealand with Google taking part in the event. This addition is entirely the result of volunteer work and Google's only participation seems to only be in encouraging the effort and participating in celebration.
Since I actually started my professional life by working as a freelance translator in my university years, I could not help but think it is not entirely fair to crowdsource the localization process completely without compensating the people for their efforts. I know that life is getting tough for translators in many language pairs with English growing to be a more and more accepted international language and many companies never bothering to translate their manuals or agreements into the language their customers have as their native ones.
And I don't understand why it is that Google - of all companies in the world - can not afford actually hiring professional translators to do the job. After all, they don't choose to cut costs by offering their engineers and coders to work for free, right?
Sure, they have done the job quite well in providing the tools that volunteers can use to translate various Google services. But can this replace payment? I really don't think so.
What's more, I am not only concerned about Google not helping keep other industries in a healthy state, what's more important is that unfortunately - and naturally - the best translators rarely have much time (or any at all) for volunteer work: they usually have a stable group of customers that they work with constantly, they already have a reputation and a brand not to need Google for another line in their CVs and they also earn enough to afford themselves some good time away from their computers when they have spare time at all. So people who engage in this type of volunteer activity normally just want to do it for extra practice and some experience to show off in their resumes. Are they good translators? I think rarely.
So even if we ignore the ethical concerns of Google keeping the money only for the internet industry, unfortunately there's a bigger problem that could damage Google itself - and that problem is quality of translation. I know Google is fascinated with machine translation and probably they really don't think it's a big problem to have a few defects in translation - after all, if the main products work as they are supposed to do it should be fine. Right? Wrong. To a user it really means irritation and probably even temptation to move to a local competitor that won't disturb them with poorly translated phrases in the interface of the products.
I have no idea how good Google is in your native language (if it is not English, of course, so do leave your comments below) but in Russian it is definitely far from perfect. This imperfection makes me to constantly try to persuade Google that my preferred language is English though it still switches me back to Russian constantly on a number of products for an unknown reason. The worst example is Google Reader (it has only recently been localized into Russian) where the translated links and text in buttons rarely make sense at all and it takes me translating them back in my mind into English first to figure out what this or that link is intended for. And don't forget that Google actually has a pretty big development office here in Russia so the situation should be better here than in other countries where Google does not have any local staff at all.
Another disappointment is that Google acts as an example for many startups to crowdsource localization and as a result they launch local versions that are not even imperfect but are totally incomprehensible. I am sure that crowdsourcing may be perfect for some things and in certain cases it can really help cut costs but I don't think user experience with a product deserves such an approach.
The image of poor translation example (the text should be "Slippery shoreline") by gardenghelle used under Creative Commons.
If you enjoyed this post, make sure you subscribe to profy RSS feed!
Yeah, mate live sucks. Unfortunately, Google is in it for the profit, not to improve the job prospects of any particular professional category. They probably spent more money making the translation tools than it would cost to translate their interface into a couple of languages, but now they can get more languages almost for free. On the other hand, the celebrations themselves probably cost more than the translation would have. But translations done by paid professionals don’t get much media attention, so that’s still win for them.
But don’t worry, things are going to get much worse for translators. Computer translation is getting better by the minute, and that’s gone suck-up a good chunk of work from professionals. Until computers get so good in doing it that only automated translations will be accepted for official documents. Don’t be fooled. Translation is one of those tasks computer can be much better at than humans. It’s just a matter of time. And I bet that Google will be one of the first ones to come up with it.
What’s worse is that there are cases of volunteers having done all the work to translate Google products into their language and Google IGNORED the translation. This happened for example with Google Mail in Esperanto. A 100% translation has been available for several years, but Google never bothered to make it official. And it’s not a grudge against the language, as Google is available in Esperanto and offers searches of Esperanto websites.
I agree with you on the importance of having a professional human translator translate company web pages and any kind of company-to-customer communication. Studies have shown that customers are 4 times more likely to buy something on your site if you speak their language and feel local. The opposite example are of course those recent spam e-mails which offer illegal localized copies of Microsoft products “in your language” and which are hardly understandable at all, seemingly coming straight from a machine translator.
Jonathan:
>> Until computers get so good in doing it that only automated translations will be accepted for official documents.
>> Translation is one of those tasks computer can be much better at than humans.
I’m a professional computational linguist currently developing a computer translation system, and believe me that there are a lot of things a computer will NEVER be able to translate, simply because it’s not human. I mean things requiring world knowledge and sophisticated cultural / transcultural knowledge. Eventually computer translators will be good enough for average people to understand texts written in a foreign language, but computers won’t ever be able to replace human translators when it comes to important information, such as business communication, negotiations, legal or medical texts.
@Jonathan: Sure, Google is in it for profit but all the businesses in the world are and they don’t make distinctions between third-party contractors - be it coding, translation or design maybe. You have a very valid point that they must have invested a lot in creating the translation tools to ensure that crowdsourcing process. But unfortunately in this case crowdsourcing does not really work (my reasons are in the post).
As for the machine translation, I can never believe it will be ever as good (or nearly as good) as human - and not because I used to work as a translator myself but simply because I see poor results. Yes, it is improving and it will continue to improve but I don’t think it will ever be enough for anything but some basic simple texts and I fully support Judith’s comment above.
@Judith: Thank you for the Esperanto example, I have not heard of it and it is definitely beyond my understanding. I can imagine how disappointed I could have been myself if it were my time and efforts invested!
Glad you support me in the necessity of human translation, I am a firm believer in it myself. Sure, some simple things can be processed good enough to be understandable. But important things should be done right - and normally “right” should mean “professional” as well. I also invite you to take a look at my older post where I tested Google’s machine translation, you may be interested as well http://profy.com/2007/05/24/google-search-translation-service-tested/.
> it will ever be enough for anything but some basic simple texts
> but computers won’t ever be able to replace human translators when it comes to important
> information, such as business communication, negotiations, legal or medical texts.
Famous last words. Never is a long time. Well, only time will tell. Best of luck!
Another potential problem with this crowd-sourcing translation could potentially be the quality of the translations, I seem to recall there was some problem with this (mentioned on this blog) about Facebook, and some of the ads that were translated into questionable things.
At some point it was amusing to observe the localization of proZ Web site. Even with supposedly real full-time or freelance, but still translators working on it, the result of this crowd-sourced effort was quite inferior. Giving the amateurs a chance to participate usually brings disaster.
Somewhere on ff there was an article about Digg drawbacks, I think, where the author concluded (and I concur) that crowd-sourcing is good when you need vetting, not creativity.
If I could get people to do less-than-perfect work for free rather than pay professionals for (usually also less than) perfect translations, why wouldn’t I?
@Judith, that is the same reason that machine translation will win the day. While the quality of the translation may be inferior, MT will become the standard because its translations will be sufficient to meet most consumer needs. (That’s why, yeah, sure, you can get your clothes still made by a seamstress or tailor in a small boutique shop, but the vast majority of clothes are being mass produced.) For marketing work, where creativity may matter more, actual translators may remain important. But legal language? The less variety, the better.
I wouldn’t bet my long-term career success on human translation failing. No way.
@Stephen: Very true, outsourcing translation can be a very tricky question and Google definitely does not show the best example to startups venturing in the same field. As I’ve mentioned, localization is an important part of user experience and user experience is definitely not something to sacrifice.
@Alex: Same here, ProZ localization was an interesting thing to watch. But I guess it only proves what I have pointed out in the post: the experienced translators simply have no time or need to do volunteer work and that partly explained the problems with the results. And by the way, I have never switched ProZ to Russian - same as I do with the vast majority of crowdsource-translated sites.
@Tammi: Simply because when you pay you usually still have a decent quality of results, not perfect maybe but still very decent. And if you bother to do the translation process right (hire a translator and an editor) you can end up with a very professionally-looking and sounding site that users won’t feel the urge to switch back to English.
It’s really nice to find someone with a matching view on localization, with the same quirks (fighting Google to stay in English, and so on). Makes one feel less insane =)
Alex, same here, I suspected I was not the only one trying to make Google stop switching me to Russian
I enjoy our conversation here and hope to see you on Profy again!