Michael Roberts was a programmer in a past life, but after the 2001 recession found that software development is too difficult to manage as a sole proprietor, and switched to technical translation, which has clear specs, very little debugging, and is replete with full-time agencies so he never has to interact with end customers. In the meantime, he translates over a million words a year and sometimes manages to automate some of his workflow.
How did you get started in freelance translation?
When Proz.com came online, I realized that I could get self-contained jobs for very little expenditure, and I never looked back.
What have you done to increase your rates over time?
I haven't. I've just learned to translate more efficiently. Heavy use of typing accelerators increased my output and thus my income by about 20-30% when I started using them, for example. More recently, I've started quoting higher word rates to new agencies. Sometimes this works.
What has been the single most effective sales strategy you've used?
Doing a good job on the first job I get from a new agency. Repeat business is *everything*.
Do you have a favourite 'type' of client?
I like people who can be less formal with me. As an American working almost entirely with Europeans, this can be a little jarring for them at first. But in the end, I think a good personal relationship is number one for working with agencies - if they can feel that you're not going to care whether the invoice was 672 Euros or 673 Euros, and that you'll reschedule other things to work in an extra file that got forgotten, you'll have a customer for life. This is the kind of relationship that I like to have with people, and there are agencies I've worked with for nearly a decade now where they're practically family.
What was your most successful project ever, and why?
I translated an entire book for SAP Press once. I can actually link to it on their Web-store. That was a good feeling of accomplishment.
Do you ever negotiate on rates?
If somebody I know and like can't quite swing the usual rate, sure, I'll negotiate. It's all about the relationship, and honestly, I do fine even at the lower rates. I'm not going to quibble over a penny per word with Hungarians - they have to fight for every bit already.
What would you ideally invest in next in order to grow your business?
Time. Time to work out automation of some of my processes, and especially time to look a lot more closely at machine translation. I know it's the Holy Grail, but I have a strong feeling that I have sufficient proficiency with tool use and translation in general now to make really profitable use of MT - but I just don't have the time to sit down with it. I have started to nibble on the edges, though.
Which tools have most impacted your profitability?
Well, let's not even count TRADOS - I don't work without it (or some other TM tool) because it's just a waste of my time and energy. So honestly, AutoHotkey for typing acceleration. It was truly an incredible boost. I have no real ground-breaking technique - if I find myself typing a word or phrase repeatedly, I make a shortcut and try to remember to use it. It really pays off for text where you type a phrase a lot, like patents, or words that you find you just can't type correctly the first time, like (for me) portfolio, which I type as "pfl".
What I'd do if I had time is to run a terminology extractor on my source document as soon as I had it, identify common terms, find translations for them immediately, then define shortcuts for them. Then, each time I typed the full word/phrase, a little popup flag would tell me that I have a shortcut for that (the little popup would be defined as a shortcut for the full word/phrase at the same time - that would be done automatically for me). This would train me to use the shortcuts defined.
Also, for the segment currently open, a small window would show me the identified terminology, the translations (so far, this is a normal tool), and the shortcuts for each. With some practice, I'd type the segment much more quickly.
But since none of that exists except in planning, it's a little underwhelming. Still, it increased my income 20-30%. That's pretty useful for something that simple. And if you do a lot of patents you'll see a bigger bump than that. You just have to remember to think about shortcuts all the time, which is sometimes difficult to do.
Do you have any advice for others looking to raise their rates?
I've never raised a rate. I *have*, however, "cheated". If I can do a job for a given rate, but use a tool or another technique to decrease the amount of work I'm actually doing, I will do that and never look back.
Case in point: I had a rather lengthy project once that couldn't be done with a TM, because the source was in PDF. The material was lab procedures, though, in a tabular format, and each step had an ID number - turned out the ID number actually indexed a step database, and each ID number's text was always the same (well, *almost* always, because it looked like maybe the texts changed over time - but not by much).
So I did what came natural: I wrote a Word macro that would allow me to type in just the ID number in the table cell for the next step - the macro would then consult my private step database (another Word document), and if the ID number was found, it would automagically fill in the table row for that step. It worked like a charm, and allowed me to grind through the entire job at least three times faster than it would have taken to type it all by hand - and it was all paid at full rate.
Unfortunately, in my experience, this kind of trick doesn't actually occur very often. But I hope to be able to exploit text regularity more efficiently in the future, at a sub-segment level, once I have time to work it out. It's not a rate increase, but better tools certainly do have a real impact on my bottom line.