;   Medical Translation Insight: Test translations - suck it up or step aside - ForeignExchange Translations

Test translations - suck it up or step aside

Test translations - suck it up or step aside (medical translation)The typical "Yellow Pages" translation service provider claims to translate anything for anybody. However, pharmaceutical and device companies need linguists who truly understand the jargon, abbreviations, concepts, and turns of phrases that are common in their space. Otherwise, they end up with funny-sounding translations, mistakes, product recalls or patient deaths.

Translation tests are a fact of life up and down the translation service delivery chain. Some clients conduct them as paid "pilots" or as unpaid tests. Most translation companies ask subcontractors to conduct short unpaid tests.

Translation tests are also universally hated and decried as ineffective, just plain wrong, or even unethical.

So, what are clients and translators to do?

  • First of all, evaluate why a translation test is needed. What are you trying to test? How will you evaluate and use the results? Is a test translation to best way to get to these results?
  • Then, determine what the test should cover and how long it should be. I believe that a translation test doesn't have to be long to be effective. I can give somebody three carefully selected sentences and determine whether or not the translator has a sufficient command of the subject matter and language.
  • Finally, decide whether or not to pay for a test (if you are the client) or whether or not you are willing to do an unpaid test (if you are the service provider).
Most linguists who are interested in working with us at ForeignExchange Translations agree to take our short (unpaid) test translation. We do frequently have applicants who refuse to take our tests. Sometimes we convince them to make an exception for us but if we cannot convince, we move on.

The same is true for our clients. We recently participated in an RFP for a large biotech company. One of the steps was a test translation - approximately 3,000 words long. After questioning the unusual length of the text, we were essentially told that the client was "looking for a committed supplier that will work closely with us; if you're not even willing to invest a bit of your time and money, then clearly you are not the right supplier for us".

We did what I believe every translation service provider should do:
  1. We weighted the cost vs. the potential benefits. In this instance, there was no comparison, and we dedicated extensive resources to make sure that we not just submitted translated text but that we produced the best possible translation: we conducted extensive research, we paid our linguists, and we invested in additional QA steps.
  2. We made sure that we could use this test translation as a learning opportunity. We tried to get receive quality expectations from the client up-front (sometimes that's possible and sometimes it isn't), we asked the client to commit to giving us feedback on our translation, we managed the test as a project (i.e., kickoff meetings, regular status updates, delivery confirmation), we proactively followed-up with the client to get feedback and schedule a debrief. If you do test translations, do them right!
Let's face it: When it comes to translation quality in the pharmaceutical and medical device space, you definitely don't just take the vendor's word for it. Clients test, they check and double-check, audit, measure and improve. Service providers either understand that - or they stay away from medical translations.

[Thanks, Jenna, for the topic!]

ForeignExchange Translations provides specialized medical translation services to the world's leading pharmaceutical and medical device companies.


  1. Hynek said...
    I agree that test translations are an investment with a great potential. I did an unpaid test for most of the agencies that became my best clients. I just have to make sure in advance that the agency is willing to accept my rates. I had an unpleasant experience with a large localization agency that praised my test translation and offered me 3 cents per word.

    Just by curiosity, did you win the biotech client?
    Tatiana Neamtu said...
    As a medical translator, I discovered that clients in this field are most demanding. Maybe because of the great implications for all parties ....
    Paulina said...
    Some translators see tests as an offence to their omniscience ;) One sworn translator seriously stated at a forum that verifying sworn translators is absurd because after the exam they are already "verified". Others replied that the exam was introduced rather recently and there are many who had not taken it, moreover, this exam tests mostly translating legal texts.
    I think a test is a great opportunity for a translator if it is short, interesting, challenging and if feedback is given. If it is long - it becomes suspicious. If it does not relate to our fields of interest, it does not allow us to show off, without feedback - it is not instructive.
    ForeignExchange Translations said...
    @Hynek - The contract hasn't been awarded yet. Any day now... :-)
    Tom Roland said...
    Great post, Andres. I agree with your approach to the tests, especially about asking for an opportunity to learn more about the client's needs & requirements in exchange for submitting the test.

    But I hope that the clients keep in mind that the tests only reflect the "best possible work that you could do under the circumstances." That is to say, you did everything you could to ensure that the test was perfect, even taking "extra QA steps" that presumably would not be part of your normal project process. But this does not reflect the "real world" of the real projects you might eventually win as a result of these tests. Your actual performance might be worse once you omit the extra QA steps and additional research efforts. But it should also get better as you gain experience with the client's materials, access to the client's expert staff to answer questions, access to the client's glossaries and TMs, etc.

    Therefore, I think tests should be seen as only one step in a process of evaluating a localization supplier's performance. The client is probably right to think that, if we're not willing to invest in a test, then we may not be willing to invest in a close relationship with the client. But even if we do turn in a stellar test, this is no guarantee that we will turn in stellar projects. To do so requires close cooperation with, and support from, the client. In effect, after we've passed the test and been awarded the work, the next test is a real or pilot project, which tests both our and the client's abilities to work together.
    √Člise Hendrick said...
    I generally think that test translations are a useful tool, especially for evaluating the abilities of translators whose skills don't fully show through on paper. Generally, I will offer a brief test translation or a sample of prior (non-confidential) work to a prospective client for their evaluation.

    However, there have been cases in which there has at least been the suspicion that an agency was using test translations as a way of backdoor crowdsourcing in order to get a free translation of an entire document.

    I remember a situation a few years ago on ProZ where it appeared that a client was using so many different sections of the text to be translated as sample material that there would be little, if anything, left to translate for pay. It was a screenplay or something of the sort, if I remember correctly.

    The trouble, from the individual translator's point of view, is that it's often hard to know going in whether a sample text is legitimate. On a few occasions, I have been given sample texts to translate that were in my areas of greatest subject-matter expertise and for which I have a wide variety of standard reference works close at hand, and have provided a translation that I could prove, if there were any question, to be excellent, only never to hear another word from the client again.

    In itself, this does not necessarily indicate anything underhanded, but I've found that, often enough, this has happened when the sample I provided was actually sufficient to evaluate whether a particular document needed to be translated in the first place (e.g., a patent abstract). In other words, the short sample may very well have been all these particular clients might really have needed.

    I still offer samples, and cannot recall ever refusing a test translation, but I do understand why there are those who are wary of them. In order to reduce the incentive to play these sorts of games with test translations, I personally would like to see a requirement that job posters on ProZ etc. certify that any sample texts they ask to be translated not be from the text they are actually seeking to have translated. If there is evidence that they're using bogus "tests" to get their documents tranalsated for free, their accounts should be suspended. If a policy like that were in place, and strictly enforced, I think that fewer people would be hesitant to provide samples.

    It would also be nice if these samples were evaluated by people who were at least proficient speakers of the source and target languages with a working knowledge of the subject matter. While it hasn't happened frequently, I have occasionally had situations where my samples were reviewed who marked turns of phrase familiar to any native speaker and expressions specific to the field in question as errors, while the accompanying comments were virtually indecipherable.

    In short, I think these tests are useful, but it would be nice if there were some generally accepted standards about how they are administered and evaluated.
    Andrew Bell said...
    After 10 years in the game, I don't feel the need to jump through hoops particularly. When I first started I would have stood on one leg with a bucket of water on my head if it had got me a job - but now I prefer not to. There are unscrupulous individuals out there who effectively get translations done piecemeal by sending "test" pieces out to unknowing translators. I was only ever stung once, but that made me cautious. So how does the client make sure they're buying good quality services? Word of mouth goes a long way - I think a strong Internet presence, professional website (and this should be meticulously edited by a mother-tongue editor) and an excellent CV - containing only relevant content and references from satisfied clients - is a very good compromise in terms of showcasing your work to the client. The ProZ "Cert Pro" (I'm Cert. Pro) system is also pretty good. Basically you submit a sample of your work which is peer reviewed, along with your membership profile, willingness to help others, answering queries etc., and this may be considered a badge of quality - depending who reviews your work I suppose. I think the bottom line is that time served and experience, coupled with a solid stable of repeat clients, give translators choices. My preference these days is to politely decline test pieces, but instead give the potential clients the means to get to know my profile, take a look at my written work online, and speak with satisfied clients. I think it's worth mentioning that you should avoid giving your regular clients "reference fatigue". I keep a list of over 20 referees, including all contact details, in a Word file and use these on a rotational basis - that way I don't go to the mill too often. I always make a point of asking my referees before I include them in my reference list. Another method that works for me is suggesting to the client that they use me for a smaller job first and see how we like each other - as does (for non-agency clients) offering the client a copy of the ATA's guide Getting it Right - which I think makes a statement that you're an informed professional who takes their business seriously. I should also mention that I also check out the client too to make sure I like them!
    ForeignExchange Translations said...
    Great perspectives, Tom, Elise, and Andrew!

    BTW, Andrew's comment originated on Watercooler. Check it out at http://translationandlanguage.ning.com.

Post a Comment


Services | Resources | Company | Contact Us | Blog | Home

(c) Copyright 2010, ForeignExchange Translations, Inc.