A Smokescreen

L’illusion est le premier plaisir.

Conjuring Games

The Conjurer (Hieronymus Bosch)For technology vendors, success depends more and more on the ability to help customers “unleash the business value” of their investments, especially when presenting a much-feared migration to an as-a-service model as necessary because this implies that those investments yield quantifiable value.

Technology vendors should help customers become more effective at leveraging data and analytics by providing them with relevant services.

Despite the interest from investors, in this respect the bar is still pretty low in the translation industry. Maybe this is due to insufficient information or a lack of real interest from industry players.

Nearly all translation technology vendors, more or less surreptiotiously, more or less closely, integrate their tech offering with a translation service offering.

In a pulverized industry with very little differentiation, both in services—and especially in technology—a tight integration of software and language services can be more easily singled out and it could entice a more discerning clientele interested in turning to one-stop-shops. On the other hand, in the same conditions, any tech-only or service-only company is relegated to being a very small company.

According to industry sources, the license for a typical enterprise TMS would cost between US$ 4,000 to US$ 18,000 per year. With an average base of 1,000 satisfied customers, this corresponds to US$ 4M to US$ 18M potential revenues per year, which is not really enticing for investors. Also, the investments to build, maintain and operate a true enterprise platform—scalable, secure, reliable, feature-rich, with every possible integration—are almost as much. Therefore, either a pool of confident and deep-pocket investors or a well-sustained revenue flow is necessary. With insufficient licenses, services should feed the revenue flow. Tertium non datur. In a service industry like the translation industry, though, a lot of confusion and conflicting interests can arise, either because the vendor base to draw on for services is the same as LSPs, or because tech vendors sell the licenses of their systems to these same LSPs.

The key word here is transparency. There is nothing wrong with integrating technology and services as long as the two remain separate and distinct, i.e., if the technology business unit sells its solutions to the services business unit on the same terms as it offers them to its competitors and vice versa. Also, based on the same principle of transparency, it must be clear to any potential customer that the two business units are part of the same conglomerate.

Yes, you are right, in an ideal world. In the real world, all companies that integrate technologies and services do so surreptitiously, either by undisclosing that they have a services unit in the case of a company with a predominantly technological character and front-end, or by selling licenses to the same companies with which they compete on services, or, finally, by dumping on one or the other side according to convenience.

Anyway, the right combination of software and language (services) cannot be reduced to a matter of pricing only. A translation company is such when it is capable of delivering any kind of translation-related services, from localization engineering and testing to multilingual DTP. For this reason, keeping the software services separate and distinct from the language services is, first of all, respectful to customers who would always be able to look elsewhere for the latter and maybe ask the translation service vendor to use the former.

A Lesson from Big Tech

What lesson big tech can teach to the translation industry? As Chris Alcantara, Kevin Schaul, Gerrit De Vynck and Reed Albergotti noticed in a recent article for The Washington Post, big tech got that big because they “acquired hundreds of companies over decades to propel them to become some of the most powerful tech behemoths in the world. They all followed a similar pattern. First, they became dominant in their original business, then they grew tentacles, making acquisitions in new sectors to add revenue streams and outflank competitors”.

The M&A trend in the translation industry is at least erratic and only a bunch of acquisitions led to larger and wealthier and more profitable entities. Almost the same happened after the injections of VCs, with the landscape remaining largely unchanged and none of the beneficiary companies leaving the industry’s cramped riverbed. In a few cases, the pandemic has hit more than elsewhere in spite of the generally bombastic narrative.

Also, although machine translation has arisen from being a relatively niche solution for sixty years to nearly ubiquity in just over a decade, LSPs still process in general less than one-quarter of their volume with machine translation despite the constantly increasing demand for multilingual content have been pushing corporate customers towards machine translation.

Not only does this indicate that many LSPs are considerably less tech-savvy than people may think, but also that despite the considerable room for machine translation to grow in the industry, a professional use of it is not yet real.

This is partly due to the average size—and revenue stream—of LSPs that makes any serious investment a major problem, even in everyday technology. But the inability to manage machine translation is the manifest proof of the absurdity of chosing to pursue, twenty years ago, a quality assurance model to suit the industry’s alleged needs together with an obsolete working pattern. ISO 17100 takes flexibility away from the LSPs who adopt it, preventing them from differentiating their offerings, especially for basic services. ISO 18587 makes thing worse.

For example, with the increasingly shortening of the shelf time of consumer products and the resulting ever-decreasing value of user documentation, customers are reasonably expecting lower prices for translation of such content and expect machine translation to be used. Therefore, LSPs might cut off one or two steps of the stiff model envisaged in ISO 17100 from the offering of translation services and rather present customers with source (pre-)editing. Yes, provided they are not ISO 17100 norISO 18587 certified.

On the contrary, getting caught unprepared, still many LSPs cannot deal with the legitimate requests for discounts from customers in no other way than by dumping their cost on vendors.

Après le déluge

The pandemic has redisegned the future of work following a different pattern than that we had originally drawn. The new workplace will no longer be the old physical office where people are tethered to desktops, but one where the same people are tethered to their own laptops or mobile devices.

The large-scale spreading of webcams and webchats, smart speakers and personal assistants, alongside with old emails and messaging is supposedly going to advance people to a much more interactive experience.

This is no news for the language industry, at least for most translators, localizers, etc., even project and vendor managers. It would be worthwhile to take advantage of it in anticipation of the expected post-pandemic recovery as many predict that this year is going to mark an inflection point.

And yet, having rounded the one-year mark of the pandemic, with the advanced world getting ready to return to normality (whatever that is), the translation industry is still struggling with the same old issues as ever.

One of these issue concerns the alleged peculiarity, uniqueness, extraordinariness of localization that would make it look like a black box to laymen. These types of positions are not at all good for the entire sector.

On the other hand, and to make things worse, stating that project management is a bit limiting might mistakenly lead people to think of localization as “merely managing projects”. In fact, the premise itself is flawed here. Managing any kind of projects is anything but simple and there must be more than one reason for so many projects, including localization projects, to fail, miss deadlines, exceed budgets, and so on.

Truth is that most companies would like to view localization teams as strategic partners, rather than service providers, but most LSPs do everything they can to make them see things differently. Take requirements. They are the foundations of any projects whatsoever. When collecting requirements, questions are essential, and they should always be neutral, i.e., unbiased—or at least look like that—specific, detailed, and stringent. For example, while the expected user and usage are requirements, the deadline is not; while quality expectations is a requirement, at least for budgetary reasons, domain knowledge should be a given, at least for specialized content, so it is not exactly a requirement. Finally, outbound format specifications is a requirement while inbound format is not.

With respect to questions and requirements, there is always at least one question being left open: Who is supposed to collect project requirements? The project manager? Uhm… The account manager? Possibly. The sales manager? Definitely, at least on first instance. And who does it normally? And how?

To be partners, partners should first understand each other; to understand each other, partners should speak the same language, or at least one should speak the counterpart’s language, that is, as Willy Brandt is alleged to have said, “if I’m selling to you, I speak your language; if I’m buying dann mussen Sie Deutsch sprechen”. And yet, the industry still lives under the presumption that it must educate the customer, so much for Marshall Field.

Money Makes the World Go Round

During her presentation at the last Gobal Ready Conference in April, Lynn Nguyễn reminded that “92.2 percent [of consumers] prefer to shop and purchase in their local currency” and that “33 percent [are] likely to “abandon a purchase” if pricing is listed in USD only”. However, if, as Nguyễn said, “Nothing is more personal than speaking someone’s language”, localization should be far more than simply dealing with language issues.

In fact, to effectively do business across the world the ability is required to move money as fast and smoothly as possible. Unfortunately, with more than 150 world currencies, smoothness is unrealistic. Even the so-called strong currencies are a dozen or so. This variety entails sometimes heavy burdens for exchange, often made worse by the additional banking charges on transactions.

In the original vision, cryptocurrencies were idealistic, anarchical, and revolutionary, following the idea of everyone being their own sovereign bank. Cryptocurrencies would be the future of money that did not require centralized authorities like banks, and the basis of a better, fairer global financial system. This should also stop the egemony of banks and ease global trades with the obsolescence of foreign exchange and currency markets. That has not happened, at least not yet. Also, over a decade later, there is still no proved evidence that cryptocurrencies are useful. In fact, cryptocurrencies are higlhy volatile and cannot be used to buy almost anything. Nevertheless, they have become a huge business, mainly as a vehicle for financial speculation. This is presumably going to keep growing, further feeding the frenzy around blockchain that, like it or not, is all about the apparent success of cryptocurrencies.

And yet, the interest surrounding Ethereum—and its recent upsurge—seems to be due essentially to the technology behind the cryptocurrency, allowing for third-party applications to use blockchain to perform functions other than money transactions as in the case of NFTs.

Over a decade later, we can now say that the dream of a single world currency free from the constraints of banks and money changers has failed miserably. This means that currencies will remain the permanent challenge to global trade. The inequalities that have burst out with the pandemic will sour and deepen across and within industries, with the translation industry being affected possibly more than others.

The funny thing, if it were not dramatically stupid, is that the weakest point of certain “futuristic” language data exchange platforms, especially those leveraging blockchain, is precisely the monetization of transactions. As if that were a negligible thing.

Payments, on the other hand, are a key aspect of any job performance and, in the translation industry, are all too often poor and late. The inequality in payments comes from the inherent lack of transparency of translation industry players, which is the main cause of the typical information asymmetry in the industry.

In recent years, the gimmicks for paying less and less, worse and worse have multiplied, in spite of the significant advances in technology and finance. The widespread practice of continuous localization has been going on for over two decades now exacerbating the picture up to making it absurd. So much for the new fair pay fad and for as simple and effective practices as SLAs.

It is not by chance that the estimates of the industry’s size have always been unreliable, and most probably inflated. After all, any relevant research has always been lacking the necessary depth and detail, with samples being at the very least poorly representative. Therefore these estimates have been the results of projections based, at best, on sales data—when not earsay—from the larger LSPs, who have never excelled for timeliness and generosity of payments.

So, among the countless innovations the industry needs, the first and most important one is about payments. Nor does it require any special technology (ever heard of APA?). Sure, exchange rates can still be an issue in some cases, but that is something that falls under what is known as “business risk”. Innit?

And yet, a possible solution at least for reducing the costs in connection with money exchange is still far to come. So, until then, the one real chance for overcome language barriers will remain free, online machine translation. So much for the survival on the translation industry, at least as long as greed rules.