Machine Translation (MT) research is 70 years old. However, the technology became mainstream only within the past decade, after tech companies started offering MT services on the internet. Over the years, MT has significantly improved. This has opened unprecedented opportunities for companies to further connect with customers around the world and enhance their internal communications with global employees.

Lionbridge now offers the Smaⁱrt MT solution, which enables customers to use best-in-class MT for the entire customer experience. In fact, for the first time, companies can Locaⁱlize everything™.

What Are the Machine Translation Types?

Statistical Machine Translation (SMT) and Neural Machine Translation (NMT) are the most common types of MT. Statistical Machine Translation relies on statistical models and large amounts of language data to predict the best translation. Neural Machine Translation also leverages large amounts of language data, but it uses different types of neural network models to render MT. Typically, Neural Machine Translation generates better results than Statistical Machine Translation, which is slowly being phased out.

Typically, however, human translators produce the best localization results when compared to machine-only translations. When companies need superior quality, they will often incorporate the services of human translators, who will refine MT output through post-editing services.

Just getting acclimated to MT and its terminology? Then, look at our cheat sheet to familiarize yourself with some of the most common terms.

What Strategies Are Used To Implement Machine Translation?

Companies can incorporate MT into their content lifecycle in a variety of ways. They typically use one of the following strategies:

Free MT services — This approach involves the utilization of free services from companies like Google Translate or Bing Translator. These services enable organizations to get quick translations for a small amount of content. The customer will get the gist of what the content means. However, this approach comes with a risk. In exchange for providing free services, the company uses all the content it receives to train its engines. This is not an ideal option for companies that want privacy and data security.
Cloud MT — This offering involves cloud services by large technology companies. It is a good choice for companies that require large scale deployments and who have in-house computational scientists or MT specialists who can train MT engines.
Best-of-breed MT — This service relies on integration. It incorporates multiple MT engines from different tech companies into one platform. It allows the user to select the best option based on content type or language pair. This is the most flexible deployment strategy that increases MT quality and reduces MT costs. Learn more about Lionbridge’s best-of-breed Smaⁱrt MT offering and how glossaries for Machine Translation used in Smaⁱrt MT bolster quality.
On-premises MT — This strategy puts the MT technology on the premises of the company using it. This is a convenient solution for organizations that want complete control of the data going into the MT engines. To execute this type of MT, companies need to set up an in-house MT server, manage it, and train it. This is the costliest deployment strategy since it requires technology and R&D teams to manage it.

What Impact is Machine Translation Having on Localization?

MT will continue to advance in the foreseeable future. Improvements to quality will drive more adoption. This will increase localization cost efficiencies. It will also help meet the growing demand for translation services that the finite pool of human translators cannot handle. MT will enable companies to localize more content into more languages. As such, they will be able to enter more markets and offer enhanced digital experiences to audiences in existing and new markets. Lionbridge MT experts track and analyze the latest MT trends. Visit the Lionbridge Machine Translation Hub to learn more.

What Are the Best Uses for Machine Translation?

There are a variety of ways to apply MT. You can use the technology to translate your website or other product-related collateral. You can communicate with people inside and outside of your company without hiring native speakers. The technology can be used to staff help desks to better serve customers in more markets. It can increase office productivity by enabling employees to translate internal documents. Use MT in the following ways:

Localization — Effortlessly translate large amounts of product, marketing, or user-generated content into many languages. By making MT part of your content journey and integrating it into your content workflows, you’ll achieve greater transactional efficiency. While human translators excel in quality, MT achieves unsurpassed cost-effectiveness. Today’s MT renders much better and more accurate translation quality than in the past, increasing its usefulness for more types of documents and texts.
Customer support — Augment your support teams in ticketing and real-time chat systems to handle more customer interactions in more languages. Look for a solution that integrates with leading ticketing and chat systems. Localization must happen in real time as the content is created. This greatly enhances the customer experience and enables companies to build strong customer relationships.
Office productivity — Enable your global employees to interact with content in languages they may not understand. But beware of free MT solutions that have the potential to put your company data, security, and privacy at risk. Companies can eliminate these risks by choosing to use a Machine Translation as a Service (MTaaS) solution. Lionbridge’s Smaⁱrt MT solution enables companies to use MT in a secure manner and achieve brand voice consistency for all translated content.

Why Should Companies Implement Machine Translation?

MT enables companies to handle a growing amount of content and increase communications across digital channels. It helps them scale the dissemination of their content in both existing and new markets despite fixed marketing and localization budgets. MT enables companies to compete effectively.

Yet, some leaders hesitate to use MT at scale because they fear their customers will have poor experiences. Extensive research counters those concerns with findings that indicate end users are willing to accept less-than-perfect quality.

A survey of 8,709 consumers in 29 countries found:

65% of buyers prefer to purchase a product when content is in their local language, even if that content contains some errors
66% of users turn to online MT when evaluating products available in other languages
40% of consumers will not buy products that are presented in other languages

(Source: “Can’t Read, Won’t Buy – B2C,” CSA Research, June 2020)

As the quality of MT significantly improves, a growing number of companies are turning to the technology to meet the needs of end user who expect more personalized and localized digital experiences. It is imperative for companies to utilize MT to compete with and surpass competitors.

By making MT part of the content lifecycle, companies can improve the experiences of non-native audiences. When MT is executed properly, it enables companies to achieve a consistent brand voice across languages and different content types. There is a caveat. It is not possible to achieve a consistent brand voice when using free MT services. This approach can, indeed, result in poor customer experiences because these free services frequently deliver average or poor-quality translations.

MT’s cost efficiency unlocks extraordinary opportunities. Companies can now make their products and digital experiences available in many more markets. The ability to Locaⁱlize everything is here. With ever-growing improvements in MT quality, companies that implement the technology effectively can expect to grow their presence in multiple markets without incurring additional costs.