(BOSTON) May 25, 2013 -- Language services consumers constitute an integral part of the global economy. Translation buyers need more words delivered at a lower price and a faster rate. “Human-Enhanced Machine Translation,” a report released in May 2013 by independent market research firm Common Sense Advisory, shows that while human translators still produce most of the commercial translation used by global brands, the demand for more languages and higher volumes have set the stage for greater reliance on machine translation (MT) in the future. 

Common Sense Advisory conducted a survey of 239 translation buyers in February and March 2013. The firm surveyed respondents about four use cases: 1) post-edited machine translation (PEMT) as an “everything included” managed service that is delivered much like human translation; 2) the use of MT software inside their companies; 3) the case where both PEMT is purchased from third parties and in-house MT is used; and 4) the decision not to use machine translation in any form.

The survey revealed that lower costs, higher volume handling, and faster turnaround times are the top reasons for buying PEMT. These benefits of PEMT fall in line with the purchase decision bottom-line narrated by language services buyers. PEMT significantly increases the utility of final text for information consumers beyond the raw output from even the most sophisticated translation applications.

Additional key findings from the May 2013 report include contrasting models for translation performed solely by humans, post-edited by third parties, and processed in-house:

• Human translation dominates today’s investment mix. Traditional translation by linguists remains the most popular method overall, both in terms of the maximum number of languages and the type of business or customer engagement content being rendered.
• Cost and turnaround time underpin the decision to buy post-edited MT. Eighty-two percent of those using their own MT claimed lower cost as a major benefit versus just 38% of those buying PEMT as an all-in managed service from third parties. 
• Pricing models bother one in 10 of the service-only buyers, but nearly one in four (24%) of those using MT software in-house.
• Post-edited machine translation is most often targeted to information consumers. Product documentation, FAQs, and knowledge bases are the most frequently served types of content processed via PEMT.
• Those buying PEMT as a managed service said that that they use some form of MT to translate about 18% of their source materials today and expect that to rise to 40% in three years. Those with in-house software said they translate about 33% of their content with some form of MT today and plan for that to rise to 57% by 2016.

Limitations of Machine Translation

The survey comprised of respondents belonging to 24 different industries from 30 countries. Eighty-three percent of the respondents have more than five years of involvement in translation services, with 34% claiming to have been using some form of MT for more than five years. However, consumers of MT have a variety of concerns, given as follows:

• 36% of respondents using in-house MT are anxious about its technical complexity
• 30% are uncomfortable with MT’s limited integration with other systems
• 24% are bothered with the pricing models
• 24% consider the language coverage by MT to be insufficient

These numbers flag issues for MT developers to address in efforts to unleash widespread use of PEMT

Comments Don DePalma, Chief Strategy Officer at Common Sense Advisory, “While post-edited machine translation provides less costly and quicker turnaround, it will only go so far in translating the enormous volumes of information that companies and government need in other languages. Post-editing is a process available today to accelerate content flow. Looking ahead, most organizations will find themselves complementing that work with highly trained machine translation software – that won’t require human intervention before publishing the output.” 

What the Future Holds

The future holds greater prospects for smart translation firms − 70% of the respondents using PEMT expecting to translate more in the categories of business and customer engagement by 2016. Most of the respondents believe that MT will be the translation solution for their ballooning storehouses of content. According to them, MT will enable more content for more people in more languages. However, unsatisfactory quality tops the list of concerns, with the pricing model posing another hazard to the growth of MT usage.

Adds DePalma, “If the level of quality assurance is maintained at the consumer’s preference, post-edited MT can be a resource-friendly replacement for a traditional human translator. With the right type of post editing, the volume of output can be increased up to eight times more than what a linguist can produce on their own. MT developers have their marching orders to improve quality for both post-edited and unedited output.”

For more information about the firm’s research services, visit www.commonsenseadvisory.com. 

About Common Sense Advisory

Common Sense Advisory, Inc. is an independent research and analysis firm specializing in the on- and offline operations driving business globalization, internationalization, localization, translation, and interpretation. Its research, consulting, and training help organizations improve the quality of their global business operations. For more information, visit: www.commonsenseadvisory.com or http://www.twitter.com/CSA_Research.