Amazon Web Services
Any organization that operates in a multilingual environment needs to be able to translate internal and/or external communications in order to be effective. Employees across various regions need to understand critical business processes, customer communications, and technical product information; customers need marketing and other external materials that are translated with particular care to ensure key branding and messaging are not lost. While this work was traditionally done by human translators, as the need for translation at scale has grown and technology has improved, machine translation (MT) software is increasingly being used to help improve speed and consistency while reducing costs in the translation process, particularly for large volumes of content.
Initially, machine translation software was based on handcrafted rules, and then later, it shifted to largely statistical models but still struggled in terms of reaching “human like” translations. As a result, humans still often ended up doing much of this work, increasing costs, timelines, and the risk of inconsistency. Starting around 2016, however, the rise of deep learning marked a shift in the way machine translation is done. Neural machine translation, or artificial intelligence (AI)–based machine translation, uses deep learning (either replacing or in combination with rules-based or statistical models) to train artificial intelligence systems to recognize language patterns and produce increasingly accurate translations. This is the method used by most, if not all, MT software vendors today. During the actual translation process, some systems are fully or semi-supervised by humans, while others perform translation completely unsupervised (note that this may differ from the level of human supervision used in initial training of the neural machine translation models). In some cases, the level of human input for translation workflows can vary based on the language or content type and client-designated accuracy thresholds.
While neural machine translation has been a game changer for the machine translation industry, there still is no universal translation capability, à la Star Trek or other science fiction shows. Most of the neural translation models are good for general language but often don’t handle industry-specific or even company-specific terminology particularly well. In these cases, some vendors offer the capability for organizations to build custom language models that can incorporate and handle industry or even company-specific jargon. However, building custom language models can be time consuming and sometime require large amounts of training data to effectively perform machine translation around very specific types of language. In addition, some vendors are starting to use what they call “adaptive machine translation” models that “learn” and adapt to new words and phrases over time, which lessens the need for custom language models.
Finally, yet another consideration is support and capabilities for a wide variety of languages. Some vendors support translation to and from a wide variety of languages, sometimes ranging into hundreds of language pairs. Again, neural machine translation has been transformative in handling these wide variety of languages, but accuracy and capability will vary substantially based on the popularity of the language and the amount of training data available to the vendor.
As the accuracy of machine translation software has improved, organizations have run into issues with employees using free online translation tools for business use. While these kinds of tools are fast, reasonably accurate, and handy for individual consumer use, they are not intended for business use and are often not secure, resulting in the potential exposure of sensitive information. However, there are a variety of machine translation systems that are designed for enterprise organizations, even for particularly sensitive domains such as finance and healthcare. They are available in a variety of forms, from APIs that can be embedded in various applications to full-service platforms that include the ability to add human validation to the translation workflow. This IDC MarketScape evaluation examines the major enterprise machine translation software products in the market today.
This section briefly explains IDC’s key observations resulting in a vendor’s position in the IDC MarketScape. While every vendor is evaluated against each of the criteria outlined in the Appendix, the description here provides a summary of each vendor’s strengths and challenges.
After a thorough evaluation of Translated’s strategies and capabilities, IDC has positioned the company in the Leaders category in this 2022 IDC MarketScape for worldwide machine translation software.
Translated, founded in 1999, is a language services provider (SP) offering neural machine translation software as well as human translation services. Its core IP is ModernMT, an adaptive machine translation technology developed over the course of two EU-funded research projects since 2010. Translated took full ownership of this technology as of May 2022 following a merger. ModernMT provides commercial neural machine translation software with advanced features including document-level translation and real-time adaptivity, enabling it to work with a wide variety of use cases without the need to customize individual models for each. It supports translation between more than 2,100 language combinations and is available for both real-time and batch translation and human-in-the-loop translation workflows that include Translated’s human translation services for validation and customization of complex/high-value content.
Translated is headquartered in Rome, Italy, and is privately held.
Consider Translated’s ModernMT for machine translation if you are a large, multinational corporation looking for adaptable multidomain support that can be deployed either on premises or via cloud, with related support for areas such as localization and human validation. Translated provides a wide range of related services and products that can assist with translation at scale.
The criteria used for the selection of IT suppliers that were evaluated are the following:
For the purposes of this analysis, IDC divided potential key measures for success into two primary categories: capabilities and strategies.
Positioning on the y-axis reflects the vendor’s current capabilities and menu of services and how well aligned the vendor is to customer needs. The capabilities category focuses on the capabilities of the company and product today, here and now. Under this category, IDC analysts will look at how well a vendor is building/delivering capabilities that enable it to execute its chosen strategy in the market.
Positioning on the x-axis, or strategies axis, indicates how well the vendor’s future strategy aligns with what customers will require in three to five years. The strategies category focuses on high-level decisions and underlying assumptions about offerings, customer segments, and business and go-to-market plans for the next three to five years.
The size of the individual vendor markers in the IDC MarketScape represents the market share of each individual vendor within the specific market segment being assessed.
IDC MarketScape criteria selection, weightings, and vendor scores represent well-researched IDC judgment about the market and specific vendors. IDC analysts tailor the range of standard characteristics by which vendors are measured through structured discussions, surveys, and interviews with market leaders, participants, and end users. Market weightings are based on user interviews, buyer surveys, and the input of IDC experts in each market. IDC analysts base individual vendor scores, and ultimately vendor positions on the IDC MarketScape, on detailed surveys and interviews with the vendors, publicly available information, and end-user experiences in an effort to provide an accurate and consistent assessment of each vendor’s characteristics, behavior, and capability.
Machine translation (MT) software is software that can translate between human languages. This includes neural machine translation (NMT), the method used by most (if not all) MT software vendors today, which leverages reinforcement learning and deep learning to recognize language patterns and produce increasingly accurate translations. More traditional techniques, often used in combination with NMT, include rules-based and statistical methods.