Saving Money Using Translation 2.0

As part of our work we do a lot of translation for clients but with things (i.e research budgets, the economy, phase of the moon, et al...) being the way they are, we're seeing some clients (who shall remain nameless) either opt for or suggest using online or machine translation.

Let me just say that not only have I dedicated an entire book to it; Beyond Translation; The Marketer's Field Guide to Understanding Today's Transcultural Consumer, stating the obvious; that this is a bad idea and not just for general or multicultural market research, as Jiri Stejskal, Ph.D., mentions in his article, Stir-fried Wikipedia with Pimientos;

"Machine translation, in which text is translated by a software program without human involvement, has opened up the floodgates on potential translation errors. In China, a restaurateur eager to attract an international clientele decided to display the restaurant’s English name on the storefront next to its Chinese name. Unfortunately, the machine translation application he chose to perform the task was not working at the moment, and his restaurant now bears the English name “Translate server error.”

Using the Internet has also thwarted several Chinese restaurants’ attempts to get accurate translations at a low price. Menus have featured “Stir-fried Wikipedia with pimientos” and “Barbequed congo eel with Wikipedia and fermented bean curd.” Apparently, these would-be translators confused the name of the Web site with the name of the item they were trying to find."

But I cannot emphasize enough that "The more technical your subject, the more important it is to have a translator who knows it inside out", especially when doing usability testing or ethnographic interviews, these translations, before, during and after, cannot be left to a machine, or even any of the online collaborative translation services we're seeing spring up.

No it's not cheap but it's imperative to work with real human beings especially bilingual researchers that will translate your screeners, guides, etc... At the very least if you are _ _ _ _ bent on using a cheaper alternative such as online or machine translation or claim to already have the translated material or "are taking care of it internally" have it reviewed by a research professional before you go out in the field when inevitably either a respondent or participant will bring it to your attention, because by then it will be too late.

If you are lucky enough to be bilingual and are curious to see just how many ways machine translation can go wrong, even with this simple post, just click on the "Translate This Blog" link to the right.

Below are some more tips to avoid translation blunders:

• Does it really need to be translated? Get rid of unnecessary information before translating.
• Use pictures instead of text whenever possible.
• Think international from the start. Avoid cultural clichés, literary references and sports metaphors that do not make sense in other countries.
• Differentiate between translation needed for information only and translation for publication. Will an accurate but unpolished translation be sufficient, or are you trying to persuade or convey an image?
• Tell the translators what it’s for. Make sure they know the type of publication and the target audience.
• Typographical conventions vary from one language to the next. For example, neither months nor days of the week are capitalized in French and Spanish. Do not be tempted to “correct” translated text to follow an English convention.
• The more technical your subject, the more important it is to have a translator who knows it inside out.