The Content Translation Challenge

Machine Translation

As the amount of content generated by marketers continues to grow, so do the challenges faced by organisations to effectively translate it into native languages that customers understand around the world. Consumers are hungry for relevant content and a deeper experience with their favorite brands. Google’s statistics show that its total number of indexed pages grew from one trillion in 2008 to 30 trillion in 2013.

Add in the influx of user generated content from social media channels, and organisations face a major translation dilemma. According to the Economist Intelligence Unit, 64% of executives say language barriers make it difficult to gain a competitive foothold in international markets and 49% say that a language barrier has stood in the way of a major international business deal.

Today’s language translation solutions must not only translate content, but also help deliver informative, relevant and consistent customer/partner experiences that increase sales and drive brand loyalty – across multiple channels, cultures and devices. The volume of text is growing at such a rate that no human translator can keep up with this growth

However, solutions such as new machine translation technologies provide scalable and cost-effective ways to deliver high-quality translations and expand the types of content that can be translated, including chat, email and social media.

Machine translation is the translation of text by a computer, with no human involvement, also referred to as automated translation, automatic translation or instant translation. There are two main types of machine translation systems: rule-based and statistical:

  • Rule-based systems use a combination of language and grammar rules, plus dictionaries for common words to generate a translation.
  • Statistical systems learn rules automatically by analysing large amounts of data for each language combination. These systems can be trained for specific industries or companies using additional data relevant to the domain.

The statistical machine (computer) translation model uses mathematical probabilities to determine the most likely interpretation of chunks of text between foreign languages. Computer algorithms no longer analyse sentences word by word but instead base their translations on whole sequences of words (phrases) to determine the most likely interpretation of a given text.

Phrase-based translation reduces the restrictions imposed by the old word-based system. While a word may have several potential meanings (computers unfortunately cannot understand context), phrases usually have only one. This invention has made some of the most obscure languages in the world understandable to anyone with Internet access.

Machine translation is ideal for gaining quick cross-language understanding. It allows organisations to communicate and support customers across languages and channels like chat and email, in a cost effective manner – enabling worldwide self-service. Machine translation technologies impact across five key areas:

1. Self-service translation for anyone, anywhere, anytime publishing

One of the critical requirements for simplifying translation is to enable anyone and everyone to self-serve so they can submit content for translation without going through dozens of steps or evaluating individual translators or vendors. Marketers, analysts, execs should all be able to quickly select the language, the quality they need and the deadline, upload the content and supporting information and receive the results – on time and on budget.

Additionally, by providing multilingual content via the corporate site through a knowledge base, FAQs, chat or other channels, companies can significantly reduce their overhead costs and put the power in the user’s hands. For example, the cost of self-service is typically 15 times cheaper than handling an inquiry at a call centre.

2. Sentiment analysis/predictive analytics

Brand makers need to know what customers are saying about products by region and language, and use that data when designing new product rollouts, adding value to existing products and solutions, or improving the customer experience.

Machine translation enables the prediction of a customer journey based on localised, hard facts about behaviors and buying patterns. It provides the best method for translating user-generated data from multiple languages into a single language and offers a more efficient and accurate analysis of global customer sentiment. This improves the customer experience, helps build loyalty and brand advocates, and assists the planning and execution of brand strategies.

3. Enterprise Search

Adding machine translation to the enterprise search process can deliver more relevant results because when keywords are translated, documents can be searched in their native language and more accurate results are returned.

4. Invaluable collaboration

With a machine translation technology, all employees can communicate immediately, inexpensively and globally even via email and chat in other languages – previously difficult to implement due to the real-time nature of collaboration.

5. Address multiple translation needs

Machine translation solves every translation challenge – from translations of real-time content like chat and email, to dynamic content like user reviews, knowledge-based articles – to more static documents like legal agreements and product manuals (as part of a human translation process).

Conclusion

Deciding what content to translate can be difficult as translation is often seen as a complex and costly process. This becomes even harder at times of strained budgets and short content shelf life, but this is exactly why savvy organisations are using machine translation. It is a cost effective solution to translate all types of content, in a short amount of time regardless of the shelf life.

Customers, including those inside the enterprise, prefer to interact and consume content in their native language. The rise of machine translation technology is invaluable in helping eliminate global language communication barriers.

SHARETweet about this on TwitterShare on LinkedInShare on FacebookShare on Google+Pin on PinterestDigg thisShare on RedditShare on TumblrShare on StumbleUponEmail this to someone
Keith Laska

As CEO, Keith leads the effort to drive significant growth and distribution reach of SDL’s innovative automated translation and language management technology to customers across the globe. Keith started his career with SDL in North America, and subsequently spent seven years on an international assignment as an executive in SDL's headquarters, building up global strategy and operations for SDL Language Technologies in Europe and Asia. He recently moved back to the US to head up the division’s cloud computing and online community strategy. Prior to SDL, Keith worked in senior positions in sales and engineering including technology and services startup Translations.com in New York. He started his career as a French Instructor at Phillips Exeter Academy in Exeter, New Hampshire.