Machine translation and human translation are both widely used in international business services today – but, which is inherently better or more valuable? In this 3-part series, we’re showing you an inside look at our latest white paper, Man vs. Machine: A Guide to Translation Technology in Modern Business Use Cases.
This downloadable white paper explores the current and future states of the translation industry in terms of services for organizations, and provides guidance on how companies can better understand and navigate the complicated world of language services. Stay tuned on the blog for the next part of the series, and see a copy of the white paper here.
Machine Translation: Going Fully-Automated in a Digital World
The Pros and Cons of Machine Translation
Pros
- Easy to produce and access online, through many various websites and services.
- Extremely cost-effective, with many free services available on the market.
- Immediate – often instant – turnaround of translations are created using a simple text input/text output format.
- Algorithms are ever-evolving and improving, including the implementation of advancements in artificial intelligence and machine learning.
Cons
- Low accuracy is often an issue, with a 60 – 80% accuracy rate in most cases [7].
- Inability to understand cultural context, despite advances in algorithmic technology.
- Available languages to translate to/from are often not representative of the complete global language footprint.
- High probability of the translation being incorrect in terms of cultural nuances, field terminology, spelling errors, etc.
- High potential for costly mistakes due to syntax errors or cultural inaccuracies.
Costly Mistakes in the Translation World
A recent study indicates that there is a global footprint of 6,909 distinct, officially-recognized languages worldwide. Google Translate, as an example, has a database of just 103 [8]. These limitations, along with the other negatives of machine translation, lead to a myriad of problems. From a business perspective, the most harmful errors that technology-produced translations fall victim to are:
1. Pure Verbatim Translation
For every language and dialect, there are rules related to syntax that must be understood and integrated into the translation process. When this isn’t done properly, the intended meaning can be easily confused.
- Example 1: Pepsi’s slogan “Pepsi Brings You Back to Life” was debuted in China as “Pepsi Brings You Back from the grave” [9].
- Example 2: KFC also made Chinese consumers a bit apprehensive when “finger licking good” was translated as “eat your fingers off ” [9].
2. Inaccurate Interpretations of “Language-Specific” Terminology
Certain words and phrases are incredibly difficult to translate from one language to another, and some cannot be translated at all. So much of language is cultural; every culture has their own phrases, sayings, and words that are a part of their history and heritage.
Machine-based localization solutions will translate these words literally – without understanding the emotion, tone or the true meaning behind the source text. an expert human translator that is not only trained in the grammatical challenges of translation, but also the cultural context, is infinitely less likely to make these mistakes. It is for this reason that highly-qualified translators always translate into their native language (versus out of it).
3. Culturally Inappropriate Translation Outputs
The tone and meaning of the message is a key consideration when translating text from one language to another. Because machine translations cannot process nuance in language, such as implied changes in tone, the essence of each translated message executed by machine translation can become distorted. To convey messages appropriately, a mastery of tone is critical to accuracy [10].
- Example 1: Clairol launched a curling iron called “Mist Stick” in Germany. In German, “mist” is commonly understood slang for “manure” [9].
- Example 2: Puffs tissues marketed under implied changes in tone, the essence that brand name in Germany without realizing that “puff” is German slang for a brothel [9].
Those organizations familiar with the risks of implementing inaccurate translations are then presented with a paradox: “How do we, as an organization, ensure that translations are not negatively impacting business if we do not have any familiarity with the output
language?”
Human Translation: Using People to Build Greater Connection
The Pros and Cons of Human Translation
Pros
- High overall accuracy when compared to machine-generated translations, with most agencies guaranteeing 99% accuracy or higher.
- A rich understanding of both the field-specific terminology and cultural context involved.
- Greater depth of creative knowledge, cultural understanding, and the application of both within translation.
- Inclusive, standardized editing and quality assurance processes to ensure accuracy.
- Fluent language speakers hold skills in providing word and phrasing alternatives that machine translation cannot.
Cons
- Typically longer turnaround time for projects when compared to machine-generated translations.
- Greater costs associated due to a combination of human resources, CAT tools, and a series of quality checks.
- Depending on the resource utilized for language services, there are limitations within the languages that translators speak fluently.
Barriers Broken: Wins for Universal Translation
While machine translations may have people beat in terms of price and turnaround time, human translators and proofreaders will always have the upper-hand when it comes to quality and true cultural understanding. When fluent language-speakers and cultural intelligence are combined, businesses can flourish by being able to reach their target audience in their language, in their voice.
Example 1
Unger and Kowitt, a traffic ticket law firm based Florida, understands that its customers are from all cultural background and languages.
To better serve their customers, they translated their website into Spanish, Portuguese and the commonly unsupported Creole to ensure that their website for visitors have an easier time navigating their pages [11].
Example 2
RPMC, a renowned partner for marketing engagement, events, and sponsorship that deliver amazing experiences for brands (like Disney, shown to the left) around the world, has a deep knowledge of the people and communities they serve. In order to make branded events a more inclusive and authentic space, the organization often utilizes translation and live expert interpretation [12].
Example 3
With over 3,100 stores in 30 countries outside of its native home in the United States, Dunkin Donuts has successfully evolved and fully-translated its menu to satisfy the sweet tooth of its global customers [11].
Stay tuned for the rest of the series on the blog, and see the full white paper here.
Resources
1. https://gengo.com/human-translation
2. https://nilservices.com/machine-translation-vs-human-translation
3. https://www.gala-global.org/what-machine-translation
4. http://www.machinetranslation.net/quick-guide-to-machine-translation/machine-translation-technologies
5. http://blog.ititranslates.com/is-there-a-difference-between-machine-and-human-translation
6. https://www.theguardian.com/education/2014/sep/19/tech-removing-language-barriers-jobs-lost-translation
7. https://www.gala-global.org/publications/intro-machine-translation-understand-when-use-mt-and-when-avoid-it
8. https://blog.memrise.com/2017/04/19/machine-translation-vs-human-translation
9. https://www.inc.com/geoffrey-james/the-20-worst-brand-translations-of-all-time.html
10. http://translatorthoughts.com/2014/01/5-common-translation-mistakes
11. https://blog.hubspot.com/marketing/global-marketing-and-international-business
12. https://languagedepartment.com/blog/global-events-translation-case-study-rpmc
