TMSs for Crowdsourcing
The following is a professional reflection that was written as a part of my Translation Management Systems class in Fall of 2017.
Now, I am not quite sure that crowdsourcing is as hot as it used to be. I think its place in the buzz-word-ity rankings has been usurped by newcomers like “blockchain.” However, I would argue that it hasn’t completely disappeared. Instead, crowdsourcing has simply become so natural to us (the generation that saw crowdsourcing becoming commonplace) that we have stopped considering it buzz-worthy. Now we crowdsource start-ups, health insurance, and sometimes (unintentionally) criminal investigations and manhunts. Maybe it’s no longer a hot topic, but I’m sure that the appeal of crowdsourcing — of it being the ‘better’ way of doing something — is niggling away at many a localization manager’s mind.
The key takeaway from any insight into crowdsourcing is that there is no such thing as a free translation. Setting up a TMS that allows for crowdsourced translations (that, let’s be honest, must later be checked by professional editors) takes a lot of manpower and money. It needs to be taken as seriously as a foray into machine translation.
I have to say, however, that when I was reading the articles assigned to us in class, my most notable lightbulb moment was that speed is a crucial factor in crowdsourcing’s appeal. This method of translation is appealing if your revenue stems from increasing your audience and you can leverage crowdsourcing to increase it quickly while also creating a sense of community within your target language audiences. It’s an ingenious internet and cloud-based answer to a market that has been sped up and turned upside down by the internet and The Cloud.
You would need a cloud-based TMS for this feature, capable of doling out small bits of text (with context) that people could work on. It would need to be able to handle a fluctuating load of traffic and you would need to be able to input and retire crowdsourced phrases. You would need to be able to set the terms for how the crowd-sourced voting should run and what qualities the system would use to determine that a term is done being crowdsourced. Somehow, and this is the tricky bit, you would need to integrate a living TM to work on this text for consistency’s sake, although I’m still at a loss whether this would need to be done crowd-side or editor-side.
That said, brand loyalty is a major factor in the appeal of crowdsourcing. I don’t really think this is an option for most companies, especially if they’re not well-established or cater to a particularly motivated audience. I also don’t believe that paid crowdsourcing can be a thing. Once you add money into the equation, what you’re really doing is exploiting specialists, inducing competitveness. Eventually, you would be in a situation where poorly-paid rivals would start downvoting each other’s translations in order to maintain control over monetary reward. That’s not the future I want to imagine. (It’d be like the Uber of translation. But it wouldn’t be a disruptor, it would be a disaster.)
