How Machine Translation is Transforming E-commerce

Last Updated October 9, 2020

Machine translation is lifting the language barrier on global trade. Does your e-commerce business have the automation tools to compete?

Expanding into international markets is a fundamental strategy for retailers looking to increase profits. For e-commerce retailers specifically, the online nature of their business puts them in an ideal position to easily reach into lucrative foreign markets.

While this is a huge opportunity, it also poses a significant challenge: how can product information and other messages be communicated in a way that resonates with each different target market?

For some e-commerce companies, the answer is simply to use English across the globe. However, research has shown that, given the choice, customers prefer to shop in their native language. Any e-commerce retailer unwilling to translate and localize their content will inevitably lose out to their more agile competitors.

Traditionally, this work would have been carried out by skilled human translators. But as technology has advanced, the standardized nature of the text used in most online retail websites has made the e-commerce sector especially suited to automation tools, particularly machine translation.

What Is Machine Translation?

Pioneered in the 1950s, machine translation software (often called automated translation or automatic translation) refers to the use of computer technology to translate a piece of text into a different language with zero, or minimal, human involvement. As technology has advanced, new methods of machine translation have evolved.

Rule-based machine translation (RBMT) was developed in the early 1970s. It essentially used a dictionary and a set of basic rules to convert text from one language into another. However, as translation involves much more than word-for-word replacements, the results were often unreliable and required additional post-translation work carried out by a human.

This led to work on statistical machine translation (SMT) methods, which began to make significant breakthroughs in the 1990s. SMT searches a vast body of existing reference texts and translations for phrases a few words long in the source material and converts them based on the most common previous translation. SMT’s weakness is that it can only translate a phrase if it exists in the existing reference texts.

A more recent development, neural machine translation (NMT), adds a key ingredient to the automated translation process: context. NMT uses deep-learning techniques to recognize textual patterns in the source material then determines a context-based interpretation one word, then one sequence of words, at a time.

NMT is the first instance of an automated translation tool that possesses both speed and accuracy—a combination that could transform e-commerce.

How Can Machine Translation Benefit E-commerce?

The sheer quantity of e-commerce content, its standardized structure, and its constantly changing nature, make it an ideal beneficiary of machine translation software.

To fully reap the rewards of the international markets, every piece of content on an e-commerce website should be translated into the target customers’ language. Product information is the most obvious element, but the same is true for non-product pages, such as About Us, blogs, social media, SEO metadata, marketing messages, and any other messaging.

Managing the rapid translation of this ever-changing content into multiple languages would be a colossal challenge for a team of human translators. The latest machine translation software, however, enables e-commerce companies to optimize the entire translation process to maximize their cross-border trade.

As well as the clear advantages of using machine translation, there is a big disadvantage to not using it: it will leave an otherwise successful e-commerce company vulnerable to competition from around the globe.

Singapore-based sports and outdoor apparel retailer, Sportmaster, for example, has integrated machine translation software into its entire operation. The company already operates in China, Russia, Ukraine, Kazakhstan, and Belarus, and is looking to trade in many other countries. As such, the company uses machine translation to support all aspects of its business—from product descriptions, to billing, to HR—to ensure the speed and quality of communication are never compromised when dealing with customers in their native language.

Chinese e-commerce retailer Alibaba has gone so far as to develop their own machine translation technology: Alibaba Machine Translation. This system is dedicated to delivering rapid and highly accurate translations for Alibaba’s international businesses, including AliExpress, Lazada, ICBU, Taobao Overseas, DingTalk and Tmall Global.

It isn’t just corporate giants like Alibaba, eBay, Amazon, and Google that are piling resources into automated translation. Countless small e-commerce traders are embracing automation tools as a way to enter international markets.

At Summa Linguae Technologies, we recognize there is no one-size-fits-all solution to the diverse range of requirements being made of machine translation.

As Lea Backhurst, our Managing Director Nordics, explains: “We use a number of different machine translation engines, each operating in a different way. This allows us to tailor the automated translation method to suit the specific type of text being translated and deliver the most accurate results.”

The Next Steps for Machine Translation

Despite the evolution of machine translation software into an accurate, workable solution, it has not yet reached its zenith. What forms are future developments likely to take?

Here are three areas where exciting innovations are already starting to take place.

  1. Contextualization is likely to be refined to further increase the sophistication of translations. This would mean evaluating the context of an entire document, rather than individual sentences.
  2. Automated feedback that quantifies the quality of the translation would allow users to filter out the translated text that needs the attention of a post-editor from that that can go direct to publication.
  3. Automated post-editing would streamline an area previously considered to require human intervention. This could involve a second machine translation engine trained to a high level of accuracy using the corrections human post-editors have made to previous texts.

Following the widespread adoption of something that would have been considered science fiction just a few years ago, we can only imagine what the future holds for machine translation. Whatever innovations take place, it is sure to continue to have a huge impact on the e-commerce sector.

Related Posts

Summa Linguae uses cookies to allow us to better understand how the site is used. By continuing to use this site, you consent to this policy.

Learn More