Web 2.0 Dogma: Is Any Translation Better Than No Translation?

One of the unspoken dogmas of the crowdsourced post-editing (CPE) model is that if you can’t afford to translate a website or document, the next best thing is to McLocalize it. In other words, use the CPE model at pennies a word. Privately, the vendor will tell you that “of course, the translation is lower quality, but at least you have access to some approximation of the original.” The recurrent mantra is that “bad translation is better than no translation.”

Listen, for example, to David Grunwald in a comment to “The Hamster Who Translated 10,000 Words a Day” explaining the deployment of his automatic translation plug-in (emphasis added):

I have a much more noble vision in mind: tons of content is being published by millions of bloggers. Some of the content, such as your own Financial Translation blog, is very good. But it remains local due to high translation costs. Our system is designed to allow publishers like yourself to translate their content at low costs. And make more money from their content while doing so. Our system supports an integrated post-editing process which enables human quality translation. I am using my own GTS blog as a test system but lack the resources to post edit in all the languages. Having said that, I can report that the blog is getting LOTS of traffic in other languages even though the quality is not very good. Our plugin has been installed in over 300 websites and some of them have reported a significant boost in international traffic. And about 30 websites have used our plugin to create their own translation community for the post-editing, which yields good quality for them.

Grunwald is so taken by his business model that he doesn’t pause to devote even a second to thinking through even the most basic questions. Mainly, must every single piece of text be automatically reproduced infinitely like some horrifying mutant cancer cell? Why can’t the CPE people even pause for a second to think about how tacky that is? The idea of my words being automatically translated into 100 languages makes me nauseous.

But I know this type of argument sounds a little too lofty. I know that this is preaching vegetarianism to the cannibals. The MT Crowd wants its meat and it wants it red.

Okay, how about a weaker argument? Try this: McLocalization is simply bad for business.

Did that catch your attention? Check out this piece from the BBC, “Spelling mistakes 'cost millions' in lost online sales.” Some theories about technology in the seventies and eighties talked about a second orality. These models were entranced by the apparent shift from written culture to visual and aural media such as radio, television and telephony. Flash forward thirty years and guess what? The gurus were wrong. The cutting edge technology is the Internet and it depends much more on the written word:

Mr. Duncombe, who runs travel, mobile phones and clothing websites, says that poor spelling is a serious problem for the online economy.

"Often these cutting-edge companies depend upon old-fashioned skills," says Mr. Duncombe.

Very old fashioned. You know, like reading and writing. Add this little tidbit: "This is because when you sell or communicate on the internet, 99% of the time it is done by the written word."

This is not some provincial school marm pining for the days when gals were gals, men were men and a quarter could buy you a three-course meal and a bottle of “sody pop.” No, this is a representative of the British private sector calling for better educational standards. Bad writing is bad business. And he has empirical data to back that up: "He [Duncombe] says he measured the revenue per visitor to the tightsplease.co.uk website and found that the revenue was twice as high after an error was corrected."

Wow. Sloppy writing can cut your revenue per visitor by half. Please note: That is one single mistake. One silly little typo and your sales get squished.

If bad writing is bad business, can you imagine what an error-filled translation can do to your revenue per visitor? The McLocalization vendor will retort that any sale is better than no sale, but what if your sales could be multiplied by ten if you took the time to properly localize your material? And what if for every sale using cheap CPE you are sending a signal to one hundred or one thousand potential customers that you are a shady company? Typos tend to reduce a retailer’s credibility, apparently:

However, there are other aspects, such as a home page or commercial offering… which raise concerns over trust and credibility," said Professor Dutton. "In these instances, when a consumer might be wary of spam or phishing efforts, a misspelt word could be a killer issue."

Once again, crowdsourced post-editing might seem like a cheap option, but the costs may be hidden and much (much) higher.

No. “Any translation” is not better than “no translation.” In fact, “any translation” can be downright hurtful if it transmits any of the following messages about your brand: you aren’t actually willing to invest any money in communicating with new clients in new markets; your product is as second-rate as your translations; you are in it for a quick buck and not interested in establishing long-standing relationships with new markets, etc. Increasingly, we are coming to realize that globalization doesn’t mean homogenization. Globalization is becoming a synonym of how local knowledge trumps a mediocre website translated into 100 languages.

Isn’t it ironic that the localization industry doesn’t actually get the local Web?

Miguel Llorens is a freelance financial translator based in Madrid who works from Spanish into English. He is specialized in equity research, economics, accounting, and investment strategy. He has worked as a translator for Goldman Sachs, the US Government's Open Source Center and H.B.O. International, as well as many small-and-medium-sized brokerages and asset management companies operating in Spain.