From Kickstarter projects like the “Fidget Cube,” to Starbucks’ “White Cup Contest,” crowdsourcing — or collectively contributing ideas, money, time, or expertise to a cause or a product — is shaping the future. The translation industry is no exception.
But in reference to the translation industry, crowdsourcing has particular nuances and jargon, and you might also hear about “collaborative translation.” So, is there any difference? While the two terms are often used interchangeably, crowdsourced translation and collaborative translation are divergent. Before we talk about more specific crowd-based translation classifications and the effectiveness (or perhaps the ineffectiveness) of the technique, let’s explore those term differences, first.
Collaborative Translation vs. Crowdsourced Translation
In the translation industry, crowdsourcing concerns assigning translation projects to individuals or groups of people, often through what’s called a “flexible open call.” This invitation seeks individuals, mostly amateurs, who want to work on a smaller piece of a larger project or document.
However, there’s a fine line between crowdsourced and collaborative translation. While crowd-sourced projects and documents often divide work between many different workers, crowdsourced translation cannot be considered collaborative translation, unless there is actual collaboration happening simultaneously — such as between managers, proofreaders, and subject experts.
This translation technique gathers numerous people who use shared resources in a collaborative work space, like via cloud computing, often referred to as, “the cloud.” Cloud computing includes, “… the delivery of on-demand computing resources — everything from applications to data centers — over the internet on a pay-for-use-basis,” according to IBM’s site.
One of the most popular cloud computing services is Google Cloud Computing, which includes everything from data analytics to management tools. By using these services collaboratively, companies and businesses can communicate more effectively and cut the time it takes to translate a document or project.
But, there isn’t a single type of crowdsourced or collaborative translation. And while businesses enjoy some benefits by using these crowd-based techniques, such as lower costs, there are also plentiful disadvantages the internet has swept under the rug. Inconsistent quality and inaccuracy are the most common crowdsourcing issues. But, different projects call for different methods. Learn more about the different types of collaborative translation in the section that follows to see if any of the methods might be right for your company.
Collaborative Translation Types
Crowd-based translation can be broken down into even more specific classifications that are helpful when deciding on which type is best is for your project, if at all. Many of these collaboration techniques are used alone, but more often than not, they’re paired with other types to meet different project needs.
Take a look at some of the most common types that follow:
- Traditional translation crowdsourcing — The most common type of crowd-based translation, traditional crowdsourcing uses the former-described “open-call process” to call on a variety of people, mostly amateurs, to translate a project. From translating technical documents to providing video transcripts, traditional translation crowdsourcing is widely-used for a variety of project types. Two words of caution: results vary greatly and ownership of intellectual property of translated content is questionable.
- Terminology resources — Some platforms, such as Wikipedia, allow multiple users to create and edit massive-scale resources. These resources can then be used to supply information to be used on other projects. Other well-known resources of this type include Urban Dictionary and Wiktionary.
- Agile teamware — Agile teamware processes let experts of different topics, such as domain experts, managerial experts, and editing experts, work together on large translation projects. Instead of using a hierarchical workflow, agile translation teamware techniques use a parallelized process, so turnaround time is generally quicker.
- Memory sharing — Some examples of translation memory sharing include Google Translator Toolkit and MyMemory. These platforms allow the sharing of a collection of translated texts, which can be pooled for multiple people or organizations to use or edit, into multiple languages.
- Crowd-based post-editing — Instead of hiring professionals to edit translation work, crowd-based post-editing calls on mostly non-specialists to improve and correct machine-translated documents and projects. Google Translate and Bing Microsoft Translator are two of the most popular examples of post-editing by the crowd. However, there are legal limitations to using Machine Translation: according to the DOJ regulations machine translation can’t be utilized for commercial use without being post edited by a professional human editor. So, in general, it ends up being more costly than having it done by professional from the beginning.
Although crowd-based translation is fairly contemporary, the approaches and systems used, such as workflow systems and databases, have been around for a while. In fact, the Oxford English Dictionary editors used crowdsourcing over 100 years ago to gather language data. But how effective is crowd-based translation today?
The Success and Decline of Crowd-Based Translation
When the Oxford English Dictionary editors collected language data long ago, they probably used pencil and paper, not web-based programs. However, similar dictionary-compilation work continues today online, such as for the Oromo language, a lesser-known, lesser-documented language spoken in the Horn of Africa. Some programs such as WeSay even, “…help non-linguists build a dictionary in their own language,” without the use of complicated code and expensive software.
Crowd-based translation has also been used to create scientific and mathematical terms, such as “light year,” “organism,” and “photosynthesis,” for American Sign Language, according to the University of Washington.
Even larger companies, like $15-billion company Facebook, recognize how important crowdsourcing is in the translation industry. In fact, Facebook has asked users to translate for free, to “…better serve the 60 percent of its 69 million users who live outside the United States,” and aid in the company’s global strategy.
The Disadvantages of Crowd-Based Translation
But, one question remains: Can crowdsourcing lead to high-quality translations, if it’s done by amateurs? The NBC News article comments that, “Critics complain of sloppiness and skimping, even as Facebook says it is improving service in an innovative way.”
While crowdsourcing is practical for some project-types, especially for projects in which accuracy isn’t so important, it isn’t fitting in most cases. Before jumping on the crowd-based translation train, let’s take a look at some of the disadvantages that come with these translation techniques:
- Decreased quality and accuracy — Crowd-based translation is usually of better quality than automatic machine-based translation services. However, while the “open call” technique used in crowdsourcing doesn’t prevent professionals from participating, the technique usually attracts inexperienced amateurs. For amateurs in the translation industry, this is fantastic news; who doesn’t want a chance to get their foot in the door? But for companies looking for high-quality translation, crowdsourced finished products can be disappointingly inaccurate.In addition, “… [amateur] translators are also not used to using linguistic assets … in order to leverage that already translated content and terminology,” said an article on the comparison of crowdsourcing translation with machine translation in the Journal of Information Science.
- Poor consistency — Can you imagine if 100 chefs worked on the same dish? A few people filleted a fish, a few people seasoned it, a few people cooked it, and so on. Even though all 100 chefs are professionals, it’s likely the finished project wouldn’t taste as good as if one or a handful of professionals worked on it. Now, imagine that all 100 people are amateur cooks. What a disaster that meal would be!When too many people, especially amateurs, dip their hands into the same project, consistency wavers. This is true for all industries, including translation.
- Difficult to oversee and manage — Most often, the lead or management team of a crowd-based translation project lacks organization. And because a lead isn’t always appointed, the resulting management strategy might be unambiguous, or worse, non-existent. For the best translation results, there should be an in-house manager or managerial team working on the project. But even when a manager is appointed, quality can slip. Controlling such a large group of people isn’t only costly, it’s time-consuming, too.
- Uncertain Project Deadlines — By not using a professional language service provider, you will be relying on a crowdsourcing team that may be inexperienced.
Using crowdsourced translation and other collaborative techniques for your businesses can be tempting, especially if money is tight. But in the long run, these crowd-based techniques are risky at best and provide inconsistent results. Perhaps, the best way to ensure high-quality translation is to entrust language solution companies like Bromberg & Associates to combine their own collaborative translation techniques with their team of qualified professionals for your translation needs and projects.