yndigo

Crowdsourcing and Translation

Even though the term is new as of 2006, and the internet has allowed the practice to take off, crowdsourcing is not new. Think of the Oxford English Dictionary, which included word usage and history researched and contributed by hundreds of volunteers, including thousands of entries by a man in a lunatic asylum.

Now not only is crowdsourced translation being adopted by popular sites like Google and Social networking site Hi5, the translation industry itself is beginning to embrace it. Jeff Howe, a leading expert on crowdsourcing, and the keynote speaker at the upcoming Localization World conference in Madison Wiscosin, claims, “[i]f there is one industry where crowdsourcing can turn things upside down, it is the translation industry.” Howe’s reasoning is based simply on supply and demand, asserting that a few hundred thousand translators worldwide are only a fraction of the number needed to meet the ever-growing need for translation.

However, as Chris Satullo points out in his article on the subject, the strength of crowdsourcing is not so much in achieving huge volumes of work as it is in arriving at a more accurate answer by averaging responses from a wide audience. “Picture a glass jar packed with jelly beans at a county fair. Then imagine that 500 fairgoers try to guess the number of jelly beans. If you add up the guesses and divide by 500, the resulting average will be very close to the accurate jelly bean count - and likely much closer than any individual guess.” Satullo goes on to say that only 10 percent of the crowd’s ideas will be gems, the other 90 percent junk.

But crowdsourcing combined with machine translation is bringing creative solutions to bear on the increasing demand for translation. Take Google, whose Google Translation Center — a translation service based on crowdsourcing that is currently in the testing phase — is likely to help its machine translation tool, Google Translate achieve better and better results.

But before we talk about amateur translators, in my experience not even all “professional translators” are cabable of producing adequate translations. And having reviewed many translations by non-professionals (which is the source of the majority of crowsourced translations) the percentage of usable results is far far smaller yet. Thus, just as with machine translation, businesses can see improvements on the horizon and will continue to gauge what solution will produce the best and most cost effective results and shift their strategy accordingly.