E.g., 11/12/2019
E.g., 11/12/2019

Machine Translation Means New Business Models for LSPs

By: Diego Bartolome, Director of Artificial Intelligence

The Spanish version of this article was post-edited from a machine translation prepared using tauyou <language technology>. Read the Spanish version here.

No doubt, you have used one of the many online machine translators (MT) to translate a short text or a web page. If you didn’t speak the target language, you probably looked at the result and thought, “Well, that French looks good enough to me!” On the other hand, if you did speak the target language, you probably thought, “MT doesn’t work. This isn’t the quality I want to offer my customers. MT produces awful results. MT is the enemy that will put me out of business.”

In both cases, you are right. There is a risk in using MT, just as there is a risk in not using it. But your customers know best what they need, so listen to them and consider differentiating your offerings into multiple quality levels – with and without MT – to suit their needs. This is a much better way to deal with today’s downward price pressure on Language Service Providers (LSPs).

Some customers want top quality, and for them you should implement your optimal process to produce traditional high-quality translation. That is, give them what they need. Although even here, MT may be useful to reduce your internal costs and turnaround times. For example, in my experience, some LSPs working in the (challenging!) medical domain are reporting a 38% cost reduction by using MT in certain language pairs such as English-French.

Other customers might require less quality because the content they are translating is intended for readers that don't really value the quality of the text, for example online travel reservation systems or certain e-commerce sites. Even in the original, native language the texts are far from perfect. So here, optimal language quality in translation would be overkill. And keep in mind that these customers have a limited budget, too, so offer them a service level below what you would consider optimal.

Nowadays, we see many more types of content than we used to – e-mail, multilingual chat, social networking, document filtering and selection, etc. – and more are sure to come in the future. Some of the customers who come to LSPs with translation needs in these areas would be quite happy with a customized MT solution that produces “good enough” quality without any human intervention. Your best option is not to try to convince them they need better quality (and refuse to use MT at all), but rather to embrace the chance to diversify your portfolio with MT-based services that bring new, recurring revenues for your company.

The bottom line is this: review your value proposition and look for ways to add and extend customer relationships with MT. You cannot do everything well, but you can surely do more to satisfy the needs of your customers, rather than focusing on your own. Solve your clients' pain points and they will be happy, whether you use MT or not.

NOTE: The views expressed here are those of the authors and do not necessarily represent or reflect the views of GALA.

 

Diego Bartolome

Diego (GALA Board member in 2013-2015) is the Director of Artificial Intelligence at TransPerfect, where he oversees the R&D and the deployment of AI solutions for clients, including neural machine translation, natural language generation, recommender systems, natural language processing techniques, and many others. He holds a PhD in Electrical Engineering and an MS in Management and Business Administration.