In today’s global economy, rapid advances in technology enable companies to create customized experiences for customers all around the globe at scale. “In an increasingly digital world, organizations that can leverage language to better improve customer interaction and operations are ahead of those that can’t,” says Mike Veronis, chief revenue officer at AppTek. “While previous versions of machine translation (MT) existed in the form of phrase-based translation systems, organizations like call centers, news outlets, and even online retailers have a greater demand for dynamic and flexible translation services in real time,” Veronis adds.

The fact is, “There is no single region in the world where the amount of content is decreasing,” says former digital globalization leader at Nielsen, Bruno Herrmann. Stakeholders in the translation and localization industry don’t expect the demand for services and technologies to taper off any time soon.


The Translation and Localization Year in Review

While demand is increasing, most experts agree that there weren’t any earth-shattering developments in translation and localization in 2018. “There has been no revolution, but there was an evolution,” says Herrmann. “But it’s actually been even more interesting to see the growth in the industry. The first part of the evolution is the acceleration and innovation of translation and localization tools due to the progress made in artificial intelligence [AI] and machine translation,” he says.

Large organizations with massive amounts of content are adopting these tools because they see a huge opportunity in using AI for translation purposes. In previous years, it was challenging to perform translations when a large number of examples of translations from one language to another didn’t already exist for a language. In 2018, this all changed when Facebook unveiled unsupervised neural machine translation (UNMT), a system that can translate without sentence pairings. This advancement could enable devices that can translate between rare languages in real time or even allow the translation of documents written in lost languages. Facebook’s former AI translation director joined Alibaba in 2018 to launch the first ecommerce translation tool, using AI to bridge the language barrier between buyers and sellers.

For marketers, there is an increased focus on omnichannel experiences. It’s not just about “translating their website, but also looking at all of their major channels of communication and making sure that the experience is consistent and specifically localized across the board,” says Craig Witt, EVP of global sales, marketing, and go-to-market at MotionPoint.

When it comes to seeing a clear ROI from localized content, conversion rates are only a small piece of the puzzle. “Companies spend hundreds of thousands of dollars on search engine optimization for their origin flagship site, but they typically don’t make the same investment in their translated websites,” says Witt.

Herrmann says there has been some progress in the way data-driven localization is used. “Due to the progress of machine learning, data-driven localization is becoming the standard because internationally, customer journeys are becoming more fragmented. If organizations are not using data to deliver on translation and localization, they will likely miss key parts of those fragmented customer journeys.”

Much like last year and 10 years ago, there is no silver bullet solution to manage data effectively. The trend for all global organizations is to consider localization effectiveness (which is objective and tangible) rather than localization quality (which is subjective). “Data is not subjective, although it can be interpreted in different ways,” explains Herrmann. He argues that organizations waste a lot of time and money around data and key performance indicators (KPIs) because of subjective discussions.

“On the one hand, traditional KPIs are still around and used to capture and measure the successful impact or the pain points of localized content. They include the alignment of terminology and the accuracy of delivery. However, these traditional KPIs should now be combined and refined with metrics for snackable content to be more granular and more closely linked to the micro-experiences of the customers,” Herrmann says. “If the localized content is not transferable enough, companies may see great sales in one market, but poor sales in another. A new mindset—free of subjectivity—must be adopted to manage data effectively.”

A Look Ahead at Translation and Localization

In 2019, Witt predicts that the need for translation and localization won’t stop once the sale is closed, but extend to include the full customer experience from beginning to end. “We’ll see more companies focusing on the full experience—translating not just through the buyer’s journey and engagement, but all the way through to the customer-support side.”

Experts in the field expect to see growth among language services suppliers in 2019 and beyond. “Some languages will certainly be in high demand, including Asian and Western European languages. New opportunities will drive the market, including African languages,” says Herrmann. “One global initiative that will have a huge impact on customer experiences and digital localization is the new Silk Road—the initiative to connect China and other parts of Asia more closely to Europe and Africa. It will create more products, more content, and more need for translation and localization services,” Herrmann explains.

“In the coming years, AI is going to touch every product and every piece of content that we have on this planet,” Herrmann predicts. However, translators should have no fear of robots stealing their jobs. “AI tools for translation and localization will have to be combined—for a time, maybe forever—with human resources, including human translators and linguists.”

Witt agrees that in 2019 we’ll see more requests for hybrid translation and localization, starting with MT and then adding human editing elements to it.A lot of the initial translation is going to be done by machines, and the human piece is going to come in more from an editorial perspective,” says Witt. “However, the technology has to advance.”

As we look to the future, we can expect to see the continuing increase of machine translation models relying on artificial neural networks, as they easily outperform the previous statistical phrase-based systems,” says Veronis. “As time and technology progress in terms of processing power and time for AI to learn, we can expect to see neural machine translation (NMTs) fully integrate with automatic speech recognition (ASR) systems allowing a complete spectrum of translation, but also the benefit of joint training and learning,” he explains.

“This combination will further increase the gap between previous MT models and NMTs, as NMTs will become increasingly more proficient at tasks other than pure translation,” Veronis adds. “This looks like automatically inserting punctuation marks, restoring the correct word case, and even converting numbers to digits from their spoken form. All of this is done with a quality level that is much higher than what could be achieved just 5 years ago, ultimately allowing organizations to improve the customer experience by not being inhibited by linguistic boundaries and providing their services across the globe.”