Volume of Multilingual Content Increases According to Netcraft Ltd1, there were 717 million websites in August 2013 compared with 176 million at the end of April 2008. As a result, the ever-increasing volume of multilingual web content makes machine translation software a fundamental application for everyday use. Without machine translation software most of these websites would never be translated. YearNumber of websites 1995 19,000 1997 1,000,000 2000 10,000,000 2004 57,000,000 2005 74,000,000 2006 101,000,000 2007 155,000,000 2008 186,000,000 2009 234,000,000 2010 255,000,000 2011 555,000,000 2012 634,000,000 User Demand for Multilingual Content Soars Each year the number of Internet users increases dramatically. According to Internet World Stats2, the global Internet audience today has surpassed 2.4 billion internet users, out of which Asia represents by far the largest internet audience. In particular, China accounts for more than 22% of the global Internet audience. These users seek information in their native language and rely on machine translation to understand foreign language information. Top 15 Countries by Internet Audience Ranked by Total Users (000) 30-June 2012 Source: Internet World Stats Country Internet Users (000) Share of Total Worldwide Internet Audience (%) Worldwide 2,405,518 100.0% China 538,000 22.4% United States 245,203 10.2% India 137,000 5.7% Japan 101,228 4.2% Germany 67,483 3.7% Indonesia 55,000 2.3% United Kingdom 52,731 3.6% France 52,228 2.3% Mexico 42,000 1.7% South Korea 40,329 1.7% Turkey 36,455 1.5% Italy 35,800 1.5% Philippines 33,600 1.4% 1 https://news.netcraft.com/archives/2013/08/09/august-2013-web-server-survey.html 2 http://www.internetworldstats.com/stats.htm Use of Free Web Translation Services Increase as Corporate Interests Face New Risks Millions of pages and documents are translated for free every day on the many online translation services, including the most commonly used search engines. This in turn creates confidentiality and security concerns for corporations. Enterprise Content Expands The amount of content created and maintained across the enterprise for different initiatives is massive and continues to grow each day. Because human translation is expensive, most of this content remains in one language. SYSTRAN solutions, which can be customized for domain-specific subjects and specific business goals, help the enterprise translate and deliver this content to users worldwide in over 60 languages. Simply put, every company large or small can leverage and monetize existing information and achieve additional return on investment by creating multilingual content from one source document with a SYSTRAN solution. Gisting or Publishing? Initial use of machine translation was limited to foreign language information access to get the gist or general understanding of content. As the technology evolved, it created better quality translation results that were suitable for publishing multilingual content. Built-in features to customize specific domains and business objectives along with open APIs for easy integration with third-party applications triggered new corporate deployments of language translation software solutions for publishing within enterprise content management systems. What Do Customers Want? MT customers want a cost-effective solution that delivers “good enough” translation quality and saves time. The expression “good enough”, commonly used in enterprise environments, means that translations must reach the company’s defined quality threshold. This can be achieved in a number of ways, such as customizing SYSTRAN’s software or imposing language consistency and terminology control. The initial cost to reach the quality threshold for every target domain leverages all available linguistic assets such as glossaries, translation memories and existing translated texts. But companies also need to consider related ongoing costs to improve translation quality of existing domains, or the quality for new target domains. In summary, these investments must be cost-effective. MT must be accurate, fast on current hardware, easy to maintain, and deliver savings on existing translation costs.