Translation Industry Future Perspective: Utopia or Dystopia?

This is a follow-up post by Luigi Muzii on the evolving future of the “professional translation industry”. His last post has already attracted a lot of attention based on Google traffic rankings. In my view, Luigi provides great value to the discussion on “the industry” with his high-level criticism of dubious industry practices, since much of what he points to is clearly observable fact. Bullshit marketing speak is a general problem across industries, but Luigi hones in on some of the terms that are thrown around at localization conferences and in the industry. You may choose to disagree with him, or perhaps possibly see that there are good reasons to start a new, more substantive discussion. Among other things, Luigi challenges the improper usage of the term “agile” in the localization world in this post. The concept of agile comes from the software development world and refers most often, to the rapid prototyping, testing and production implementation of custom software development projects.

Agile software development is a set of principles for software development in which requirements and solutions evolve through collaboration between self-organizing, cross-functional teams. It promotes adaptive planning, evolutionary development, early delivery, and continuous improvement, and it encourages rapid and flexible response to change.

To apply this concept to translation production work is indeed a stretch in my view. (Gabor, can you provide me a list of who uses this word (agile) the most on their websites?) While there is a definite change in the kind of projects and the manner in which translation projects are defined and processed today, using terms from the software industry to describe minor evolutionary changes to very archaic process and production models, is indeed worth raising some questions on. The notion of “augmented translation” is also somewhat silly in a world where only ~50% of translators use the 1990’s technology called translation memory, a database technology that is archaic at best.

It is my feeling that step one to make a big leap forward is to shift the focus from the segment level to the corpus level. Step two is to focus on customer conversation text rather than documentation text that few ever read. Step three is to have proper metadata and build more robustness on this leveragable linguistic asset.

MT is already the dominant means to translate content in the world today, but few LSPs or translators really know how to use it skillfully. Change in the professional translation world is slow, especially if it is evolutionary, and involves the skillful use of new technology (i.e. not DIY MT or DIY TMS). In my long-term observation of how the industry has responded to and misused MT, I can attest to this. Those few who get MT right I think will very likely be the leaders who will define the new agenda as tightly integrated MT and HT work is a key to rapid response and business process agility (not “agile” ) and continuous improvement. Effectiveness is closely related to developing meaningful new technology savvy translation process skills, which few invest in, and thus many are likely to be caught in the cross-hairs of new power players who might enter the market and change the rules.

Given the recent rumors of Amazon developing MT technology services, we should not be too surprised, if in the next few years a new force emerges in “professional translation”, that from the outset properly integrates MT + HT + Outsourced Labor (Super Duper Mechanical Turk) with continuous improvement machine learning and AI infrastructure to deliver equivalent translation product at a fraction of the cost of an LSP like Lionbridge or Transperfect for example. They are already building MT engines across a large number of subject domains, so have deep knowledge of how to do this on the billions-of-words per month scale, and they are also the largest provider of cloud services today. As I pointed out last year the players who make the most money from machine translation are companies outside the translation industry. Amazon has already displaced Sears and several other major retailers and they have the right skills to pull this off if they really wanted to. Check out the chart on retail store closings that is largely driven by AMZN.

Even if they only succeed in taking only 5–10% of the “translation market” it would still make them a multi-billion dollar LSP that could handle 3 word or 3 billion word projects into 10 languages with equal ease, and do this with minimal need to labor through project management meetings and discussions about quality. It might also be in the most automated and continuous improvement modus operandi we have ever seen. So, think of a highly automated and web-based customer interaction language translation service, that has a Google scale MT with output better quality across 50 subject domains and an AI backbone, and the largest global network of human translators and editors who get paid on delivery, and are given a translator workbench that enhances and assists actual translation work at every step of the way. Think of a workbench that integrates corpus analysis, TM, MT, dictionaries, concordance, synonym lookup, and format management all in one single interface, and which makes translation industry visions of “augmented translation” look like toys.

So get your shit together boys and girls cause the times they are a-changing.

The highlights and emphasis in this post and most of the graphics are my choices. I have also added some comments within Luigi’s text in purple italics. The Dante quote below is really hard to translate. I will update it if somebody can suggest a better translation.

— — — — — — — — -

Vuolsi così colà dove si puote/ciò che si vuole, e più non dimandare.  It is so willed there where is power to do/That which is willed; and farther question not. Merriam-Webster defines technology as “the practical application of knowledge especially in a particular area.” The Oxford Dictionary defines it as “the application of scientific knowledge for practical purposes, especially in industry.” The Cambridge Dictionary defines technology as “(the study and knowledge of) the practical, especially industrial, use of scientific discoveries.” Collins defines technology as the “methods, systems, and devices which are the result of scientific knowledge being used for practical purposes”. More extensively, Wikipedia defines technology as “the collection of techniques, skills, methods, and processes used in the production of goods or services or in the accomplishment of objectives.”

This should be enough to mop away the common misconception that technology is limited to physical devices. In fact, according to Encyclopedia Britannica, hard technology is concerned with physical devices, and soft technology is concerned with human and social factors. Hard technologies cannot do without the corresponding soft technologies, which, however, are hard to acquire because they depend on human knowledge that is obtained through instruction, application, and experience. Technology is also divided in basic and high.

That said, language, to all effects and purposes is a technology. A soft technology, and a basic one, yet highly sophisticated.

Why this long introduction on technology? Because we have been experiencing an exponential technological evolution for over half a century that we can hardly master. We have been adapting fast, as usual, but every day less.

This exponential technological evolution is the daughter of the Apollo program, whose upshot has been universally acknowledged as the greatest achievement in human history. It stimulated practically every area of technology.

Some of the most important technological innovations from the Apollo program were popularized in the ’80s and the ’90s, and even the so-called translation industry is, in some ways, a spinoff of that season.  Indeed, if the birth of the translation profession as we know it today can be traced back to the years between the two world wars of the last century, with the development of world trade, the birth of the translation industry can be set around the late 1980s with the spread of personal computing and office automation. The products in this category were aimed at new customers, SMEs and SOHO, rather than the usual customers, the big companies that had the resources and the staff to handle bulky and complex systems. These products could be sold to a larger public, even overseas, but for worldwide sales to be successful, they had to speak the languages of the target countries. Translation then received a massive boost, and the computer software industry was soon confronted with the problem of adapting its increasingly multifaceted products to local markets. The translation industry as we know it today is then the abrupt evolution of a century-old single person practice into multi-person shops. As a matter of fact, intermediaries (the translation agencies) existed even before tech companies helped translation become a global business, but their scope and ambition were strictly local. They were mostly multiservice centers, and their marketing policy was to essentially renew an ad on the local yellow pages every year. With the large software companies, the use of translation memories (TMs) also burst onto the scene. The software industry saw in the typical TM feature of finding repetitions, a way to cut translation costs.

So far, TMs have been the greatest and possibly the single disruptive innovation in translation. As SDL’s Paul Filkin recently recalled, TMs were the application of the research of Alan Melby and his team at Brigham Young University in the early ’80s. Unable to bear the overhead that the large volumes from big-budget clients were procuring, translation vendors devised a universal way to recover from profit loss by asking for discounts to their vendors, regardless of the nature of jobs. In the late 1990s, Translation Management Systems (TMSs) began to spread; they were the only other innovation, way less important and much less impacting than TMs.

At the end of the first decade of the 2000s, free online machine translation (MT) engines started releasing “good-enough” outputs, and since the surge in demand for global content over the last three decades has resulted in a far greater need to translate content than enough talent available, MT has been growing steadily and exponentially, to the point that today, machines translate over 150 billion words per day, 300 times more than humans, in over 200 combinations, serving more than 500 million users every month. (Actually, I am willing to bet the daily total is in excess of 500 billion words. KV)

We are now on the verge of full automation of translation services. Three main components of the typical workflow might, indeed, be almost fully automatized: production, management, and delivery. Production could be almost fully automatized with MT; TMSs have almost fully automatized translation management and delivery. Why almost? Because the translation industry is not immune to waves and hype, but it is largely and historically very conservative, a little reactive, and therefore a “late” adopter of technology. A manifest evidence is an infatuation with the agile methodology, and the consequent excitement affecting some of the most prominent industry players. Of course, prominence does not necessarily mean competence.

In fact, agile is rather a brand-name, with the associated marketing hype, and as such, is more a management fad, that has a limited lifespan. In fact, localization processes can hardly be suitable for agile methodology, for its typical approach and process. If it is true that no new tricks can be taught to any old dog, for agile to be viable, a century-old teaching and practicing attitude should be profoundly reformed. Also, although agile has become the most popular software project management paradigm, it is understood for not having even really improved software quality, that is generally considered low. (Here is a website called http://agileisbullshit.tumblr.com/ that documents the many problems of this approach. KV) In contrast, the translation industry has always been claiming to be strongly focused on and committed to quality. If quality is the main concern for translation buyers, this possibly means that most vendors are still far from achieving a constant level of appreciable quality. In fact, while lists of security defects for major software companies show high levels of open deficiencies, the complaints of translation users and customers around the world say that the industry works poorly.

Raising the bar, increasing the stakes, pushing the boundary always a little further, are all motives for the adoption of a new working methodology like agile. These motives translate into more, faster and cheaper, but not necessarily better. Indeed, higher speed, greater agility, and lower cost of processes are supposed to make reworks and retrofitting expedient.

Anyway, flooding websites, blogs, presentations, and events with paeans praising the wonders of a methodology that is supposed to be fashionable is not just ludicrous, it is of no help. Mouthing trendy words without knowing much about their meaning and the underlying concepts may seem like an effective marketing strategy, but, in the end, it is going to hurt when the recipients realize that this only disguises the actual backwardness and ignorance.

The explosion of content has been posing serious translation issues to emerging global companies. The relentless relocation of businesses on the Web made DevOps and continuous delivery the new paradigms, pushing the demand for translation automation even further. Many in the translation community speak and act as if they were and will be living in imaginary and indefinitely remote place that possesses highly desirable or nearly perfect qualities for its inhabitants. They see the future, whatever it is depicted, as an imaginary place where people lead dehumanized and often fearful lives.

In the meanwhile, a survey presented a few weeks ago in an article in the MIT Technology Review reports, there’s a 50% chance of AI outperforming humans in all tasks in 45 years and of automating all human jobs in 120 years. Specifically, researchers predict AI will outperform humans in translating languages by 2024, writing high-school essays by 2026, writing a bestselling book by 2049, and working as a surgeon by 2053.

After all, innovation and translation have always been strange bedfellows. Innovations come from answering new questions, while the translation community has been struggling with the same old issues for centuries. Not surprisingly, any innovation in the translation industry is and will most certainly be sustaining innovation, perpetuating the current dimensions of performance. Nevertheless, despite the fear that robots will destroy jobs and leave people unemployed, the market for translation technologies is increasing, but translation Luddites are convinced that translation technologies will not endanger translation jobs anytime soon, and point rather to the lack of skilled professionals.

Indeed, the translation industry resembles a still life painting, with every part of it seemingly immutable. A typical part of this painting is quality assessment, still following the costly and inefficient inspection-based error-counting approach and the related red-pen syndrome.

In this condition of increasing automation and disintermediation, a tradeoff on quality seems the most easily predictable scenario. As for the software industry, increasing speed and agility, while controlling costs could make reworks and retrofits acceptable. MT will be spreading more and more and post-editing will be the ferry to the other banks of global communication, allowing direct transit between points at a capital cost much lower than bridges or tunnels. MT is not Charon, though.

Charon as depicted by Michelangelo in his fresco The Last Judgment in the Sistine Chapel

The key question regarding post-editing is how much longer it will be necessary or even requested. Right now, most translation jobs are small, rushed, basic, and unpredictable in frequency, and yet production models are still basically the same as fifty years ago, despite the growing popularity of TMS systems and other automation tools. This means that the same amount of time is required for managing big and tiny projects, as translation project management still hinges on the same rigid paradigms borrowed from tradition.

The most outstanding forthcoming innovations in this area will be confidence scoring and data-driven resource allocation. They have already been implemented and will be further improved when enough quality data is going to be available. In fact, confidence scoring is almost useless if scores cannot be first compared with benchmarks and later with actual results. Benchmarks can only come from project history while results must be properly measured, and measures must be known to be read and then classified.

This is not yet in the skillset of most LSPs and is far, very far to be taught in translation courses or in translator training programs.

However, this is where human judgment will remain valuable for a long time. Not quality assessment, which is still today not yet objective enough. Information asymmetry will remain a major issue, as there will always be a language pair totally outside the scope of any customer, who has no way of knowing if the product would match the promises made to the customer. Indeed, human assessment of translation quality, if based on the current universal approach, implies the use of a reference model, although implicit. In other words, everyone who is requested to evaluate a translation does it based on his/her own ideal.

MT capability will be integrated into all forms of digital communication, and MT itself will soon become a commodity. This will further make post-editing replace translation memory leveraging as the primary production environment in industrial translation in the next few years. This also means that, in the immediate future, the urge for post-editing of MT could escalate and find translation industry players unprepared.

In fact, the quality of the MT output has been steadily improving, and now it is quite impressive. This is what most translators should be afraid of, that expectations on professional translators will be increasing.

With machines being soon better at almost everything humans do, translation companies will have to rethink their business. Following the exponential pace of evolution, MT will soon leave little room for translation business. This does not mean that human translations will not be necessary any longer. Simply that today’s 1 percent will shrink even further, much further. Humans will most possibly be required where critical decisions must be made. This is precisely the kind of situation where information asymmetry plays a central role, in those cases where one party has no way of knowing if the product received from the other party would match the promises, for example when a translation should be handled as evidence in court.

With technology making it possible to match resources, find the most suitable MT engine for a particular content, predict quality, etc. human skills will have to change. Already today, no single set of criteria guarantees an excellent translation, and the quality of people alone has little to do with the services they render and the fees they charge.

This implies that vendor management (VM) will be an increasingly crucial function. Assessing vendors, of all kinds, will require skills and tools that have never been developed in translation courses. Today, vendor managers are mostly linguists who have developed VM management competence on their own, and most of the time cannot dedicate all their time and efforts to vendor assessment and management and are forced to do their best with spreadsheets, without having the chance to attend HRM or negotiation courses. Vendor management systems (VMSs) have been around for quite some time now, but they are still unknown to most LSPs. And yet, translation follows a typical outsourcing supply chain, down to freelancers.

So, translation industry experts, authorities, and players, should stop bullshitting. True, the industry has been growing more or less steadily, even in a time of general crisis, but the translation business still only counts for a meager 1 percent of the total. In other words, when translation buyers are deciding to waive the zero-translation option and have all or most content translated, the growth is still linear.

Agile in translation is not the only mystification via marketing-speak being used in the localization business. Now it is the turn of “augmented translation” and “lights-out project management.” (Lights Out Management (LOM) is the ability for a system administrator to monitor and manage servers by remote control.) Borrowing terms (not concepts) from other fields is clearly meant to disguise crap, look cool, and astonish the audience, but, trying to look cool does not necessarily mean being cool. In the end, it can make one seem she/he does not really know what she/he is talking about. Even trendy models are shaped by precise rules and roles: using them only as magic words may backfire.

Nonetheless, this bad habit does not seem to decline even a bit. Indeed, it still dominates industry events. Localization World, for example, is supposed to be the world’s premiere conference when it comes to unveiling new translation technology and trends. Anyway, most of the over 400 participants gathered in Barcelona seemed to have spent their time in parties and social activities, while room topics strayed quite far away from the conference theme of continuous delivery and the associated technologies and trends, despite the fact that the demand for better automation and more advanced tools are growing steadily. Maybe it is true that social aspect in conferences is what conferences are for, but then why pick a theme and layout presentations and discussions?

Presentations revolve around the usual arguments, widely and repeatedly dealt with before, and after the event, and are often slavish repetitions of commercial propositions. Questions and comments are usually not meant to be challenging or to generate debate, although stimulating and enriching it would be. Triviality rules, because no one is willing to burn his/her stuff that is intended to be presented in other times to different audiences.

Anyway, change is coming fast and, once again, the translation industry is about to be found unprepared when the effects of the next innovation will mess it up. So, it is time for LSPs — and their customers — to rethink their translation business and awaken from the drowsiness in which they have always received innovations. Also, jobs are changing quickly and radically too, and the gap to bridge between education and business would be even wider than it is now, which is already large. It is making less and less sense to imagine for one’s own children a future in translation as a profession, and this is going to make it harder and harder to find young talents who are willing and able to work with the abundance of technology, data and solutions available in the industry, however fantastic. This said it won’t be long before “skilled in machine learning” becomes the new “proficient in Excel”. And now very few in the translation community are concretely doing something about this. Choosing an ML algorithm will soon be as simple as selecting a template in Microsoft Word, but so far, very few translation graduates and even professional translators seem that proficient. In Word, of course.

Luigi Muzii has been in the “translation business” since 1982 and has been a business consultant since 2002, in the translation and localization industry through his firm. He focuses on helping customers choose and implement best-suited technologies and redesign their business processes for the greatest effectiveness of translation and localization related work.

This link provides access to his other blog posts.

Originally published at kv-emptypages.blogspot.com on July 19, 2017.