We might learn a few things on the translation industry from survey reports and complaints of the last few weeks.
Surveys and Statistics
Traditionally, people have often looked at statistics as deceptive. In Italy, a sonnet from dialect poet Trilussa is famous for saying that, if you believe in statistics, when someone eats a chicken, and someone else does not, each has eaten half chicken. In the 1960s and 1970s, Darrell Huff’s How to Lie with Statistics became an introduction to statistics for many college students, and it is still one of the best-selling statistics books in history. Scottish novelist Andrew Lang is also known for paraphrasing Alfred Edward Housman in saying that people use statistics as drunkards use lamp-posts, for support rather than for illumination.
In this respect, surveys should always follow a scientific approach as to questionnaire construction and methods for improving the scope and accuracy of responses. In fact, the main challenges are the identification and assembly of representative samples and the evaluation and testing of questions and responses. In turn, the inferences about the population of interest depend on survey questions and the representativeness of the sample.
The slovenliness in running surveys within the translation industry and the sauciness in claiming their results as credible are disheartening.
Another Missed Opportunity
This is not the first survey from these organizations, and it is indeed another missed opportunity. All these trade organizations might have set up a consortium way earlier, with a research department and a statistics unit to collect data, get insights and issue reports that they might also use for lobbying. They rather opted for a DIY approach and results have been as predictable as disappointing: 1,404 random responses from 55 countries despite they can reach out for more than 15,000 industry players of all types and sizes.
To be clear, this is way better than a dozen or two respondents out of 5,600 subscribers or than some rough Fermi estimates from intuitions and innuendo.
On the other hand, the industry’s “I-know-it-all-need-no-advisor” mindset is peculiar.
The report is as predictable as disappointing because there was no need for experts to come to the same conclusions. There was no need for a survey either. In fact, the report’s major outcome is that “expectations are tapering off”.
Another easily predictable outcome is that quality of service is top priority for customers, followed by responsiveness, quality of deliverables and flexibility, while “low cost” is not.
Also, individual professionals report that their clients want a high focus, while companies favor a wide offering. Correspondingly, companies express a concern that freelancers show less quality focus and should follow the specific project instructions provided to them. Indeed, to them, in high-level assignments, the quality of deliverables is too often unsatisfactory.
On the other hand, relationships with PMs is a major concern, for the inadequate reference materials, too short deadlines, assignments that do not always consider the translator’s areas of specialization and long but not always relevant boilerplate instructions. As a consequence, more and more highly qualified and experienced translators are leaving the industry (requires password).
Professional translators also complain about LSPs increasingly including them in tender submissions without assigning any job after winning the contract. This is part of a perceived increase in competition, together with significantly lower growth in sales/revenues, and the tendency of individual language professionals to have less opportunity than a company to significantly increase their own sales volume.
Oddly, or maybe not, in earlier survey, language service companies reported that they were expecting a rate increase, for the first time since 2015. However, 2018 did not meet the expectations and only a few companies report that their rates have increased, while a few more experienced a further price drop. Equally oddly, language service companies report a quite significant increase in new recruitment in 2018. Maybe, everybody lies. In fact, companies still report an increase in certification requirements in 2018 as in earlier surveys, but the increase was not as strong as expected. Consequently, they have also adjusted their expectations for 2019 to a somewhat lower increase.
Interestingly, from an IT perspective, both independent professionals and language service companies are looking for better ease of use and/or integration and lower cost of ownership. Language service companies are also looking for better machine translation quality, and investments in this respect are on top of their priority list.
However, machine translation seems to have a long way yet to become a standard part of the toolbox of translation professionals and companies. In fact, all respondent types still plan solid investments in CAT tools. So, the certification of crappy translation quality systems is an investment, cost of ownership is an issue, but respondents generally still look suspiciously at SaaS. Some might call this schizophrenia…
Quite surprisingly, the transfer of user rights and ownership of language data is still a sensitive issue, despite the very low contractual focus of customers on this area. Indeed, this might be an indirect confirmation of the irrelevance of make-believe translation data marketplaces and an explanation for the bland resonance of the possible application of blockchain technology, regardless of the complete failure of all solutions in the international development sector.
Back to the roots then, where rates are still the strongest challenge and government regulations the weakest one, and where independent professionals show strong concerns about competition and price pressure.
It is comforting to read that all types of respondents agree that “market awareness of the importance of professional translation is slightly improving”.
It is predictable and yet discomforting, to read that most respondents feel the lack of “market data and information about the competition […] despite the existence of surveys run by national associations and the reports published by international market research organizations”.
So, here comes a little tip for the 2020 edition: Sutor ne ultra crepidam. Have the EU fund the research, at least partially, and hire a professional polling organization to work with to pick the sample, prepare the questionnaire and analyze the responses.
The Other Side of Surveys
The main problem of the EU market is the presence of the world’s largest customer, a public institution that plays according to market rules, but that does not apply to vendors the same rules that it applies to employees.
Therefore, every research in this area requires a deep knowledge and understanding of the many regional markets, which is hard to achieve without a constant and active presence.
Nonetheless, running a continent-wide survey is easier that running a world-wide one, especially when working in, from and mostly on the world’s purportedly largest market. In this respect, hiring a former director-general from the largest EU customer in a blatant conflict of interest is pointless, especially for a very short time, unless it is a payback.
Also, periodically querying language service providers is a sterile exercise, without having accurately shaped the pool of respondents in a representative sample beforehand, at least for a specific market segment.
This might be the case of CSA’s latest report. Maybe, the focus on the 195 largest companies in CSA’s Top LSPs rankings might make it more reasonable, but these rankings might also be highly questionable as they follow mostly the same approach as the 2019 language industry survey report.
However, it is quite comforting to notice that rather than taking only revenues into consideration, CSA’s report finally contemplates also profits and rates, thus, although indirectly, volumes.
According to this report, most ‘market leaders’ registered growth in revenues and profitability without raising their prices. In other words, profits have increased or remained unchanged.
This means that volumes have increased, although maybe less than expected. Also, if, rightly, “profit equals revenue minus cost”, to keep profit margins stable, “reducing operating costs” has been necessary, and, with larger volumes and, therefore, higher job complexity, this obviously means cutting compensations of vendors and low-ranking employees.
Therefore, the strong concerns of European independent professionals about competition and price pressure might be sound.
On April 11, 2019, TransPerfect Global (TPG) filed a complaint in Manhattan federal court against H.I.G. and Lionbridge Technologies for misappropriation of trade secrets under the Defend Trade Secrets Act (DTSA).
Allegedly, H.I.G. obtained confidential information in the context of a possible takeover bid and used it to bolster Lionbridge’s business instead. The possible takeover would have followed the decision of the Delaware Court of Chancery to put TPG at auction, after a deadlock between the company’s shareholders.
The complaint is an enlightening read. In it, TPG alleged that H.I.G. acted as a “bad-faith” participant in the auction to gain access to confidential information: “H.I.G. submitted one or more bids and entered into a written agreement it never intended to honor so that it could gain increased access to TPG’s trade secrets”.
The proprietary information H.I.G. allegedly accessed was mostly trade secrets regarding TPG’s business processes. According to TPG, by using this information, Lionbridge has “revamped its sales strategy, product offerings and pricing to mirror that of TPG and continues to use TPG’s trade secrets to poach TPG’s clients”.
The allegedly misappropriated documents concerned TPG’s employee files, top client lists, sales models, marketing strategies, unique cost and pricing structures, compensation models and commission schedules.
The most intriguing allegation of the complaint consists in H.I.G. devising a strategy in 2016 to buy Lionbridge and TPG and create the world’s largest language translation service provider. In this perspective, TPG would have enabled Lionbridge to diversify its customer base, achieve double digit revenue growth, and increase its margins given TPG’s profitability. Also, according to the complaint, the likelihood of Lionbridge repaying its loans would increase significantly with the acquisition of TPG.
The complaint explicitly reports that Lionbridge’s CEO John Fennelly “announced in words or substance (after having access to TPG’s trade secrets and confidential information) that «Lionbridge is going to be a dramatically different company regardless of whether they do TPG or not». Fennelly continued that he thought «TPG’s sales model was pretty cool» and recognized that the «compensation model was part of the secret sauce». Lionbridge CFO Clemente Cohen agreed, commenting that he «loves seeing commission plans go up and up»”.
Finally, according to the complaint, Lionbridge sales people falsely told TPG’s customers that Lionbridge was purchasing TPG and that they should contract with Lionbridge directly before the sale. Allegedly, they also contacted TPG’s existing and prospective clients, and both misrepresented the nature of the underlying litigation and introduced doubt about the stability of TPG in bad faith for the purpose of damaging TPG and advantaging Lionbridge.
Beyond any easy and pettily rude comments, one thing looks plainly clear: Technology is out of the industry’s scope. On the other hand, none of the top industry players runs an R&D department. They have bought virtually all the technology they use. Recently, Lionbridge reached an agreement with AWS to exploit Amazon Translate and reduce its costs by 20%. Previously, TPG took over the Spanish language technology company Tauyou.
Despite the persistent claims that quality is a top priority for customers, price is still the leading factor, like the CSA report and TGP’s complaint show.
In the last few years, machine translation has been infiltrating the modus operandi of many translation professionals and companies out of necessity. Therefore, a huge new technology leap from any of industry players is largely unexpected.
Twenty-five years have passed since Jeff Bezos founded Amazon, twenty since his interview to Charlie Rose in which he outlined “a place where people can come to find and discover anything that they might want to buy online”.
People—and companies—are going to buy all kind of translations through an online shop very soon. So, a drastic change in the traditional business model is way more necessary than ever. And if the question is when, the answer is now.