It was announced. We were seeing the trend (or should I say the threat?). It's starting to happen.
No strange internet web pages, no short strings from an indie mobile application.
No: a long, hard, financial report, received by a partner/client that has been preprocessed by a TM (Translation Memory) plus, for no matches, gone through a Machine Translation engine.
The request: Please provide a quote for editing this material.
"We don't do this", "Editing this to get something decent would take us more time than translating from scratch"…
But wait: that's not what the client is asking... The client asks for a quote, a proposal.
Why not providing the client with what he needs? Wouldn't it be a matter of establishing a productivity based on the documents to edit, calculate the amount you want to charge and make a proposal?
The only switch would be to quote on a project basis (based on the number of hours the project takes: don't we do it already for many proofreading human material projects?). Why is it so different from quoting on a per word basis? The per word rate is not the same (or shouldn't be) for all materials, so why quoting on this particular project be an issue? Just establish how much you want to charge. It is very possible that the final pay would not be less than it was on a per word basis, the job would just be different.
The translation and localization industry is living a period of change. And with change come opportunity :-) (sorry for the cliché).
Maybe it is the time to establish a new relationship with clients. Maybe these changes allow service providers to get deeper into the process. You shouldn't be charging less for an editing job (no matter how the material you have to work with was generated), you have to charge what you have to charge. If the effort is finally more: charge more; if the effort is finally less…than maybe charge less? It may be that the amount of content you will work with, will increase. It may happen that the amount of content to edit will grow exponentially. It may happen that your revenue will increase too. It's just a switch, it's just an adaptation.
Some will argue that the intention of the client is to get cheaper and cheaper translation and spend much less on your services. Well: you know what? We are not going to stop this. You may chose to look for different clients, but you may also chose to start looking at your process, your way of doing business and adapt. The GILT industry is a perfect market after all: lots of demand and lots of offer.
In this sense, nothing has changed: position yourself where you want to be. If there is a new service to offer, then offer it and charge what you find fair to charge. There will always be someone that will be charging less, don't even worry about this (there is always, always, always, someone cheaper), but that's not the league we are in, or are we?
You are absolutely right that it is worth seeing opportunities instead of just accepting the continuous worse conditions that many agencies offer.
ReplyDeleteAny agency will of course ask for as low rates as possible - just like anyone buying anything, so that's just business.
Previously machine translation has been very costly in means of training the translation engine, but I guess that costs will be more and more reasonable within a few years.
Within a few years it will make sense for a group of freelancers to share the costs and invest in a machine translation solution for their specific subject area and then do editing instead of translation/proofing.
Previously agencies had a lot to offer in terms of experienced staff, processes for handling lots of different file formats etc., but as everyone outsources, trained staff no longer exists in the agencies, and as more and more material are in XML format the only file conversion required is to and from the translation tool that each agency prefers.
It will probably already make sense for end customers to go directly to a group of freelancers. Currently the customer sends projects to larger agencies, that sends it to smaller agencies etc., and it is often a long way to the translators. Every link in the chain tend to just do file handling and often mess up query handling, so most of the parties involved doesn't add value compared to the costs they inflict.
In addition the quality of TM content has become very low in the last few years, so doing fuzzy matches becomes more and more expensive for the translator as he/she is still responsible for 100% quality. Sharing TMs between freelancers are often also a problem as not everyone trust each other to deliver the required quality and messing up TMs even more. Using a shared MT solution might give better results in the long run as training the MT engine with bad translation probably doesn't affect the end result as bad as populating a TM with bad content which can overwrite an existing good translation.
So, you are absolutely right that machine translation is not all bad. It will most likely be bad for smaller agencies that will have a hard time adding value to the translation chain, but it might also make it more difficult for larger agencies.
So the result might even be that the freelancers will get closer to the end customer so that the end customer can save money and that the freelancers can make more money as editors than as translators/proofreaders.
Excellent article! Thank you!
ReplyDeleteI think MT is both a threat and an opportunity. As the amount of content increases global enterprises have to seek new production models. If used with skill MT can provide production efficiency. But it still needs talented humans to be successful.
ReplyDeleteFor those who are involved with translating highly repetitive content, I think it will be a threat as this is the material that MT will do best.
I have blogged about this at http://kv-emptypages.blogspot.com