Charlie Bavington

Professional French to English Translator - Business and I.T.

Bringing a pragmatic eye to meeting your needs

Article: Pros and cons of free tests

September 23rd, 2011 | Categories: agencies, business, My articles

(This was originally an article I posted in my Articles section in April 2009. I recently got involved in another discussion about tests and I haven’t posted since my holiday – yes, thanks, it was splendid – so here it is, tweaked just a tad.)

Ah, indeed, tests, and the payment or otherwise thereof. An old issue, scarcely a week goes by without some translator somewhere passing comment thereupon. Translators being the largely discontented bunch they appear to be, the comment is rarely along the lines of “How topping! I was asked to do a free test and you can imagine the pleasure I took in agreeing to this perfectly reasonable request.”

Actually, it seems to me to be a fairly straightforward issue, regardless of its age. An issue comprising two key points, namely a) why do tests and b) should they be paid for? Plus a third, subsidiary point which irritates yours truly beyond all reasonable measure, so we’ll come to that last.

So, a) “why do tests?”

I confess, I think “why do tests” is probably a little simplistic – it should be “what need drives clients to ask for tests and do tests meet that need?” – but I was looking for a short sub-heading.

Three reasons justifying requests for tests

The obvious first point is that clients ask for tests in order to test (naturally enough) that translators can deliver what they say can deliver. If you claim to be a legal translator, that you can translate a paragraph or two from a contract. If you claim engineering knowledge, that you can translate the description of a cable stay for a bridge, say. And so on. Unfortunately, the translation business is full of people with a misplaced confidence in their abilities, or who deliberately mislead clients, and anything in between. Your exams and credentials may be perceived as not having examined the specific subject area to the depth that a client may need. Or maybe the information about you in the public domain doesn’t really indicate much one way or the other. So they ask for a quick couple of paragraphs to prove capability to deliver.

The second point is that clients may ask for a test to check translators can follow simple instructions. I have limited outsourcing experience, and even I can tell you that some translators will get the document, and just jump into translating that document, before they have read the accompanying email all the way down to the inevitable “Regards….” bit. So asking people to start at the third paragraph (say), can just be used to test how much attention the translator pays.

Third, they may ask for test to see if you are actually able to meet technical or ancillary requirements. Can you handle XML? Provide a TM in TMX format for the client? Giving the translator a handful of HTML pages and receiving a nice Word document back, even if translated perfectly, may not be what is required.

Counter-arguments – valid

A common counter-argument to that first point is that samples demonstrate the same thing. True to an extent, and more so for a specialist. As a counter-counter-argument, I would say that I would expect a sample made available to be as near to perfect as a translation ever gets, and all it demonstrates is the ability to hone that particular text to the nth degree. It does not necessarily demonstrate the ability to deliver the specific type of text the client requires. And it in no way demonstrates the important additional ability to follow instructions. And neither do credentials, certificates, diplomas, membership of professional organisations, or indeed paid membership of popular translation websites. Some of these can indeed be easily forged, faked or presented in a misleading way, and also be a bit of a bugger to check, particularly from another country.

A more reasonable but conversely less universally-applicable counter-argument is that if you are being asked for a test, you probably don’t know the potential client from Adam, and the potential client also does not know you from any other character from the religious text of your choice. So we should be careful. It is certainly safer to acquire new clients by personal recommendation and referral, and the same is partly true of agencies using new translators. I have also heard the viewpoint that a client who is testing several people (not that we usually know how many are being tested at once) is likely to view each of them as interchangeable or disposable, initially at least, until they prove otherwise. Once again, a spot of (demonstrable) specialisation is your friend.

Counter-arguments – invalid

“I have done several tests and have never subsequently received any work. So tests are pointless.” Or you might be incompetent, or out of your depth in the tests you have done. In any event, if you genuinely believe that your personal experience represents a universal truth, there is something amiss with your logic ciruits.

An inconclusive conclusion

In a nutshell, then, if the work is likely to just be general blurb requiring no particular specialist skill or knowledge, then I would agree that testing seems a little excessive when the client ought to be able to meet that need with little effort. The more obscure, specialised and technical the work, the more reasonable it seems to me to check, with a test, unless the translator has a relevant sample or samples or has been recommended. And yet, tests do typically tend to be fairly straightforward, as if testing for adequacy rather than specialisation. But there are vast numbers of self-proclaimed translators out there who struggle even to be adequate. And let us not forget, self-proclamation is all it takes (in most countries, anyway). So for many clients, unfortunately, even adequacy is a step in a welcome direction. They can’t believe a word half of us say, and so they feel the need to test us.

And b) “should tests be paid?”


If you do a free test, does that mean that client gains the impression you have nothing better to do? Some say it does. It is hard to construct a convincing counter-argument to assumptions that other people might hypothetically make. But that risk can be minimised. Do not immediately jump to it. Take a few days, perhaps (unless there is a deadline, of course). Email it at 10 p.m. or midnight. Give the impression (whether true or not) that you had to squeeze it into a busy schedule – although obviously make sure it is well done. Or ask for payment, of course. Ultimately, though, when there is no duress or compulsion or threat of sanction, when the whole process is entirely optional in every conceivable way, I have trouble accepting that anyone is being “exploited”.

… or marketing?

Some argue that a free test is merely marketing (a sprat to catch a mackerel, perhaps). The time spent is merely time spent on marketing. Not only that, it is marketing time spent on a potential client with at least a passing interest in the service you offer. Otherwise why test that service?

Yes, that could be slightly idealistic or even naïve of me. My own experience indeed mirrors the common observation that the more hoops a potential client wants you to jump through, the less likely you are to ever get actual paid work out of them. So if the test comes combined with a questionnaire and application form and bank details form and non-disclosure agreement and a request for a copy of your grandparents’ birth certificates, you can probably just hit the Delete button and move merrily on (without in any way wishing to fall into the trap of equating my experience with universal truth, of course).

Personally, on balance, I can see more merits to the “free tests = marketing” point of view than drawbacks.

Working for nothing

Some say that they just don’t want to do “work” for nothing. Fair enough. I have put “work” in quotes because it can mean many things to many people, yet it is so often the word used by those rejecting the idea of free testing. It is clearly “work” in the sense of taking time and effort for the testee, even if viewed as a marketing task. Whether it is “work” for the tester, in the sense of being of saleable value to them, is debatable. Ideally perhaps it should not be – a test should have a standard (of) translation for the testee to aim for, so the tester can pass/fail testees on a consistent basis, which would imply consistent, if not the self-same, text being used on all testees. But if they are testing you for a particular project…

It must be said that a paid test does tend to demonstrate good faith by the client, if the test is paid promptly. That said, any such relationship is still a new one, and just because they paid your 3 groats for a test within a fortnight, it does not mean they will ever pay your 300 guineas for the entire user guide you subsequently translated.

And a paid test also demonstrates that the potential client recognises that your time is worth something, which is a point in favour of paid tests, without in my view managing to conclusively demonstrate the opposite.

Fraud! Scam!

Others appear not to trust most potential clients as far as they can throw them, and seem to assume that every single person who asks for a free test is somehow going to cobble together a paid-for deliverable for a third party from a disparate collection of 300-word tests. And they are not going to get caught out like that, oh no. Fine. I’ve never actually seen a scintilla of concrete proof that this has ever been done, but that doesn’t mean it hasn’t happened. Even if we accept that it has happened, that does not mean that the purpose of all tests is this kind of deception. This whole idea that free tests are used in this way seems largely spread by those convinced that Beelzebub himself is pulling the strings behind every agency in the known universe. It’s a kind of reverse engineering – “hmmm, I don’t like the idea of free tests, what nefarious purposes can I devise to justify this dislike…. I know, cobbling together an entire document from free tests”.

Whereas I would contend it is more logical to put yourself in the position of an individual determined not to pay for work, and think how best that would be achieved. I would further suggest that simply refusing to pay (optional extra: challenge the quality) and using geography to lessen the chances of any comeback is the simplest route. Those wiser than me have calculated it would probably cost about as much in time for the tester to coordinate such a stunt as it would cost in money to get the job done properly. That said, it could certainly not be ruled out if the tests are longer. A 3,000 word translation, say, pieced together from “tests” of 600-700 words provided by 4 or 5 people would probably be feasible, and much more feasible than attempting to achieve the same result from 9 or 10 different people providing 300 words. Moral: Tests ought to be short.

Some argue that if a free test comes with a deadline, then that is proof of some fiendishly cunning plan to construct a patchwork translation for free. Not impossible. But note that a deadline could be imposed on a test simply to determine the ability to follow instructions (as mentioned above) or because there is a deadline for the project being tested for, hence a translator needs to be booked from date D and so the tests need to be evaluated by date D minus 5 and so all the tests need to be delivered by D minus 10, say. It definitely helps to know the purported purpose of the test, so if the stated aim is to confirm suitability for being added to an agency’s database for projects as yet unspecified, a deadline would appear unnecessary.

On the subject of length, I would tend to agree that a free test should take about an hour of one’s time (otherwise it could indeed give the impression you have nothing better to do). I guess a paid test can be any length at all. Indeed, one could reasonably suggest that those agencies which make a habit of giving short texts as their first job to a new translator are blurring the distinction between tests and paid work. But while there may not be much consensus on the subject of testing in general, there does seem to be consensus that a test, if tests are permitted to exist, should be no more than 300 words, plus or minus 50.

Another inconclusive conculsion

So, a whole set of pros and cons and conflicting logic. Tests should, logically, test specialist knowledge, but in reality usually don’t, and specialists usually get new clients by word of mouth and recommendation – no testing required. Tests do not logically need to test adequacy, since even a basic qualification should guarantee that… but there are no barriers to entry to this profession. Upshot – some people do tests, and some don’t, and adopting either position, be it permanently or case-by-case, is perfectly valid and logical.

That is, of course, merely my opinion. It would be remisss of me not to point out that while the Chartered Institute of Linguists in the UK has no such restriction at this time (and neither does the ITI), the Code of Conduct for the ATA in the USA in fact says members will not require translators or interpreters to do unpaid work for the prospect of a paid assignment.
(Sept 2011 update: since this was first written in April 2009, that link no longer leads where it once did, and the page it now leads to makes no such explicit demand, although one could argue it is still implied.)


c) The irritating (beyond all reasonable measure) subsidiary point.

Certain translators seem to be constantly comparing the translation industry to plumbers and builders and lawyers and architects and doctors and taxi-drivers in an attempt to justify their position. When it comes to tests, they will reel off a litany of professions, make the bald statement that “they don’t do free tests/work” and then wrap up with “and neither do I”. Or perhaps the would-be opinion-former will endeavour to create an amusing scenario whereby they ask a builder to build some stump of a wall as a test, or a plumber to install one tap as a test, or an architect to sketch out a shed before awarding a contract for a tower block, or a taxi driver to convey the translator a few hundred yards as a test prior to booking a trip to an airport, before asking the reader to compare such a scenario with being asked to do a free translation test and then share hollow laughter with the author of the said opinion.

Now, analogies and comparisons that do not work well really do tend to irritate me more than a thistle-lined jock-strap.

So, perhaps we could stop comparing ourselves to professions with strict entry requirements, such as the law and medicine, where there are equally strict sanctions for charlatans and chicanery and a qualification is generally likely to be both genuine and proof of reasonable competence (or adequacy).

Perhaps we could stop comparing ourselves to professions where testing is indeed highly impractical.

Perhaps we could stop comparing ourselves to professions where asking for a test genuinely is tantamount to (part-)performance of the actual job required, demonstrably and self-evidently so.

And perhaps we could stop pretending that, as consumers or as businesses, if there were hypothetically a way to test builders, plumbers, lawyers, mechanics, and other providers of goods and services we use, we wouldn’t do so. Of course we would. We would test everything, if we could (wouldn’t we? Or am I projecting?). But we can’t. Which is why some critical professions are strictly regulated, and why we have trading standards authorities (and consumer magazines/websites such as “Which?” in the UK or “Que Choisir” in France) for the others. Oh, and contract law… but don’t get me started on translators and contracts here.

And let us not forget, in our enthusiasm to paint translators as the only profession in the world where some practitioners do “free work”, that lawyers (oh yes, them again) often offer free consultation, the actors do auditions, that graphic designers submit designs for competition without payment, that advertising agencies submit ideas ditto, that some software publishers offer a trial period or an evaluation version, car dealers will offer test drives for entire weekends, and so on.

Not all such “free work” takes the form of free testing as such, granted. But I genuinely believe one reason translators are asked to do tests is because it is a service which readily lends itself to the concept of testing. If other services did the same, they too would be tested, and some in fact are.

Now if you, personally, do not wish to “work” for nothing, that is absolutely your prerogative. It is an entirely reasonable stance to adopt.

But there are plenty of sound and economically-rational reasons for doing tests, even free ones, so perhaps those who have opted not to go down that path could treat those who have with a modicum of respect. Argue the case from a perspective related to the business of translation or even language service provision more broadly, and please keep the incessant and irrelevant drivel comparing translation to other occupations out of the discussion.

  1. September 23rd, 2011 at 11:55
    Quote | #1

    Thanks for an excellent piece! I agree wholeheartedly with the points you make. I have no issues with doing short free tests, as I wouldn’t expect to be given permanent employment without sitting through an unpaid interview either. I’ve found some of my best clients after doing free tests.
    One thing I would add to your piece is that the translator could also use the test to assess the tester. When an agency advertises for a “patent translator”, the number of sub-specialisations that could end up in your inbox is still vast. Doing a test will help you decide whether or not you’re the kind of patent translator they need. I recently agreed to do a ‘technical’ test, only to receive a text on chemistry while my CV clearly states that physics is my specialty. I didn’t understand a word, but when I tried to get the agency to send me something I at least had a chance of completing, I got an automated reply to say I had failed the test. Would I have wanted to work with that agency? Not on your nelly.

    • Charlie
      September 23rd, 2011 at 12:35
      Quote | #2

      Good point. I also had an experience of testing the tester, when I needed to know whether the target was UK or US English (for “dual carriageway”, IIRC). I got no answer. Test remains unsubmitted to this day. Maybe I failed some aspect of their own test by not approaching the issue in the way they wanted, who knows?

      Thanks for the tweet link, by the way!

  2. Rob Grayson
    September 25th, 2011 at 18:25
    Quote | #3

    I agree to do free tests from time to time, though I make no commitment to return them quickly at busy times, and I tend to keep them to 300-500 words max (i.e. I have been known to refuse longer ones). Whether or not I agree to do a free test translation largely depends on the job and/or the agency in question. As a rule, I’m more likely to do it if it’s in connection with a specific potential job, or failing that, if what I can find out about the agency suggests that it is reputable, professional and good to work with. I have gained valuable work off the back of free tests in the past – though to a very significantly lessser degree that what I have gained from personal referrals.

  3. October 5th, 2011 at 12:57
    Quote | #4

    An excellent post, as usual, Charlie.
    I don’t tend to do (unpaid) tests as I usually don’t have the time or inclination, but I have made exceptions for very specialised texts for promising-looking clients (and obviously, in my earlier days as a freelancer I was much more willing to do unpaid tests).
    I’ve never been asked to do a paid test: I guess most clients are happy to let the first paid job itself be the test.

    My experience, from which I believe no generalisation is possible:

    * Failed one test (and justly so, to be honest – it was right at the beginning of my freelancing career and way outside my comfort zone. I wasn’t very impressed having “organisation” marked up as a spelling error when no instructions were given as to which variety of English should be used, mind you, but fair dos to the agency for sending me their revision and comments).

    * Never got any feedback from another, despite being promised it and chasing up. Never heard from them again.

    * Passed one test (very specialised), got the job from which the text had been taken (take note, all you doubting Thomases!), got paid on time, and don’t think I ever heard from them again.

    Erm, that’s about all I can remember. The vast majority of my clients have never asked me for an unpaid test.

    • Charlie
      October 6th, 2011 at 09:26
      Quote | #5

      Actually, I’m starting to think I might need to sharpen up a little here. A couple of times in the past couple of years, I’ve agreed in principle to do a test, then not found the time to do it. And I have gained decent work from doing them in the past.

      I should probably add (I had forgotten) that I did in fact once have the joy of “setting” a test, for a job I posted on a well-known website (!). Just a couple of sentences – piece of cake if you knew the field (insurance). Such variety in the standard of replies, I would hardly have believed it possible.

      Anyway, thanks to all for their comments – we’re all being jolly reaonable about this whole topic.

  4. Louise
    April 2nd, 2012 at 12:16
    Quote | #6

    It’s a good idea, before agreeing to do any test, to find out whether or not the company is prepared to pay your rates. Saves a lot of time…

    • Charlie
      April 3rd, 2012 at 10:54
      Quote | #7

      It certainly can, if you can get them to disclose. It’s been so long since I did a test, I can’t remember if ballpark rates, at least, were established beforehand. I suppose they must have been – I’m not the sort of person who would take a punt like that.

      I can also imagine (not happened to me, though) that some agencies might reply along lines of “let’s see your work and we’ll decide if you are worth it”, but in those cases, I can also imagine that the relationship might not be completely smooth anyway, and could be best avoided.

  5. bob
    June 29th, 2012 at 19:56
    Quote | #8

    Yes you are projecting. Basically all this means that experience + diplomas + references = memmbership in prof. associations … all this is not enough … Never enough – more productive, more this more that – Nothing and that obstacles to work increase constantly. Apart from that, many of those companies bosses make money doing much less work than we do and pay themselves much better.

    The other part of the reality is that for most jobs, 85 p. 100 of the time people get work because they KNOW someone. When I see translators boasting around that they have many customers, that they get the best pay : i just laugh loudly. Many “translation companies” are now backed financially by other companies which can belong to the IT, the financial, and even the oil sector. Easy to keep a job in “your company”when you do not even finance it.

    Except in governments, the translation sector is a jungle and a sector full of ignorant and crookish “managers”.

    • Charlie
      July 4th, 2012 at 12:15
      Quote | #9

      Er, thanks…. I tihnk. The “projecting” comment was not about my tolerance of free tests in translation; it was about my presumption that most buyers would test the product or service they are buying given an opportunity to do so. But I admit that presumption, being my own attitude, naturally colours my tolerance of tests in translation.

      I don’t have access to your figures, but I agree it appears that “people you know” are a major source of new clients for the more established freelancers. Still, no harm in tests for the other 15%, is there? Or for those who are not established, or who have high client turnover.

Comments are closed.