June 2019
Text by Burckhardt Rueffer and Miguel Sevener

Image: © nicolasmenijes/123rf.com

Burckhardt Rueffer is the CEO and co-founder of BeatBabel. After spending many years on both sides of the great divide (geographically and professionally on the buyer and supplier side of translation services) he moved from his native Germany to California and co-founded BeatBabel to follow his passion for languages and the communication industry.


Burckhardt[at]beatbabel.com
www.beatbabel.com  


 


Miguel Sevener – a biology graduate turned translator – decided he would rather be a translator with research skills than a scientist with language skills. He joined BeatBabel in 2014 and has spent the past years growing into management responsibilities and attending industry conferences as an active speaker all around the globe. Miguel is currently taking some time off in Spain and is focusing on writing and making music.


Miguels[at]beatbabel.com
www.beatbabel.com  

Is the translation industry suffering fast food syndrome?

Translation quality has been greatly compromised in the attempt to meet budgets and tight timelines. How good is "good enough"? Is it perhaps time to establish quality standards to protect our profession?

Just like businesses in any other sector, language service providers compete for clients from a finite pool. The figures can be quite misguiding: According to market research firm Common Sense Advisory, annual enterprise spending on translation services is expected to reach US$56 billion by 2021. These numbers can give the false impression of a blooming industry with a promising future for all players, big or small, and all approaches, human and machine-based. What the number doesn’t reveal is the reality of an increasingly consolidated market that "cannot meet anticipated volumes" through human translation.[1]

What does this mean? It means that Quality has been brought to the negotiation table that seats LSPs and clients who want to localize their content. Although quality was once a given, it has now joined Speed and Cost as part of the negotiation.

Language is typically defined as a "system", and as such, it is a very strange thing to commodify and sell as a product. Of course, the language itself is not the product, more so the highly qualified work that goes into transferring meaning from one language to another. Even though one would never dare bring the phrase "fit for purpose" into a discussion about air travel (Kamikaze AirTM: "You may not make it alive, but we guarantee you will get there!") or heart surgery (Barebone BypassTM: "Get a couple more years out of the old ticker"), the quality of our translations is now fair play for sales and marketing strategists everywhere.

Levels of service are a legitimate business strategy and a valid product for certain uses. They also offer LSPs a level playing field to compete against the bigger providers. But it is not obvious that "good enough" translations are good for our profession. We would like to make the case for reclaiming the establishment of quality standards, instead of following the demands of a market that is often ignorant of our profession.

 

The negotiation

This article, and the presentation we gave at tcworld conference 2018, were inspired by a real encounter with the "good enough" culture during the preparation phase of a very large project. In late 2018, we received an email from one of our clients with very high-visibility content. The company announced a translation project of 600,000 source words to be translated into twelve target languages.

600,000 words into twelve languages is a big deal for any localization service provider, so, after the initial reaction of hysteria had settled, we took action, rounded up every possible resource, and sent our client a detailed estimate including timelines and special volume-discounted rates.

Their response was quick:

"We reviewed your per-word rates and found them to be about 15% higher than your competitors’ rates."

This won’t come as a surprise; such negotiations are standard when dealing with large projects. What might seem new, however, is what is being used as leverage in this negotiation and what is being negotiated:

  • 100% translation matches: The reason we were proposing a symbolic rate for these is that the sample files shared with us were full of inconsistencies and all types of errors, which we recommended having reviewed by our linguists.
  • Dispensing with editing and proofreading steps: We believe in offering all-inclusive rates, and do not consider these essential steps in the process as "add-ons", just like you wouldn’t consider technical inspections of a plane as an "added service" when buying an air ticket. Either it flies or it doesn’t.
  • PM fees: It is not entirely uncommon to waive this fee for certain projects. However, the justification in this case is a novelty of our times: It is based on the efficiency introduced by new technologies and tools which, in our experience, tend to create more problems than they solve.

 

In summary, the object of this negotiation was Cost, which is nothing new. What was new is that it was being leveraged against the Quality of the product, with the aid of new technologies and an improved efficiency that justifies this trade-off.

We pointed this out and made several compromises, to receive the final email which would set the tone for the entire project:

Thank you for sharing the updated rates. We are looking forward to working together on this."

 

The "good enough" culture

The state of the industry: First, we believe it safe to assume that, given the option, translation buyers would always choose a high-quality translation over a poor one, if speed and cost were not a factor. Therefore, our starting point must necessarily be that – in the absence of deadlines and with no budget constraints – the buyer will always choose quality.

In reality, however, it appears that speed and cost are the two factors lessening the importance of quality in the client’s decision-making. Those two factors cannot be understood without taking a look at recent trends in the translation market and the global economy in general.

It is almost impossible to obtain an exact number for how many words were translated by language service providers (LSPs) in any given year. The exponential growth of content is obvious, however, and we can get an indirect glimpse of it through other numbers:

  • In the past two years alone, 90 percent of the data in the world was generated.[2]
  • In 2017, Netflix users streamed 1 billion hours of video per week.[3]
  • Google Translate processed 143 billion words per day in 2016, in 100 languages[4]
  • Wikipedia users publish 600 new page edits per minute.
  • Total worldwide data will swell to 163 Zettabytes (ZB – that is a 1 followed by 21 zeros) by 2025, one order of magnitude higher than it is today. Most of the data (60 percent) will be generated by enterprises, and 20 percent of it will be critical to people’s daily lives.[5]

Now, these numbers can seem a bit foreign to our industry, especially as many of them refer to "data" and not words. Not all this data will be text-based, of course. But it is safe to assume that the data that will be “critical to people’s daily lives” will be verbalized and, necessarily, translated. In any case, these numbers are solely to put into context the immense growth in content and the inescapable nature of the market that will cover this need for translation: It will be Fast and it will be Cheap.

 

The age of comfort

On the flip side of the coin – leaving the realm of the theoretical – we have the actual professional practitioners that are surrounding all this content. It is crucial to understand that we are still working with people (thankfully), and that their decision to work with you or not will ultimately boil down to one thing: their comfort. Your client contacts want to be comfortable, they want to be making the right decision when they buy your services. Their bosses, procurement, accountants and management want to be comfortable: They want the process to go well and to deliver their own services or products within budget and on time.

As language service providers, we can cater to the comfort of our clients if we understand that it is articulated, just as the overall trends we reviewed, in terms of Speed, Cost and Quality.

Speed

Fast means that our point of contact (POC) is happy because they can meet the fast pace set by their direct supervisor < business unit lead < management < market < end user.

But how fast is fast? It is almost impossible to estimate an average turnaround time across the industry. There is no hard data to go by besides what is being advertised, which ranges from 60-minute turnarounds based on crowdsourced solutions, Netflix and their 48-hour complete localization cycle, to more traditional per-project estimates. Speed is being redefined with each new technological milestone and all players are forced to keep up. In any case, speed is an easily measurable factor and we can simply define "fast" as "faster than the competition" or "well within the buyer’s timeline".

Cost

Cheap means our POC is happy because they don’t have to justify every detail of the expense or navigate the budgeting constraints of their unit < department < management < end user (who will not think twice about shifting to the competition’s better price point).

But how cheap is cheap? It really depends on who you’re competing with. For general competition, in the U.S., a 2016 article from Slator found the average translation rate per word to be US$0.21. [6] In Germany, that number was 0.15 EUR[MOU1] .[7] So, cheap is anything cheaper than those industry averages, and of course, within budget. It must be said that cheap isn’t always within budget, and vice versa. This factor is more difficult to define than speed, although it is also easily measurable.

Quality

Finally, quality is the only factor without a direct impact on our POC’s happiness. Rather, its impact is indirect: Good quality means our POC is happy because they avoid negative feedback from their reviewers or from the end users after the service/product goes live, and the dip in customer satisfaction and revenue this entails (or used to entail). But mostly, good quality equals comfort because it avoids rework (preventing delays and additional cost). Industry research lists rework as one of the main reasons for a translation buyer to drop their translation provider. It is therefore evident that quality becomes important insofar as it affects speed and cost.

Bearing this in mind, our initial question “How good is good enough?” takes on a less optimistic tone. Of the three cornerstones of comfort, quality is clearly the most difficult factor to measure. Also, with comfort as the ultimate value, it is easy to see how quality just has to be good enough to avoid the negative impacts of slower speed and higher cost.

 

The fast-foodifying of translation

If you ignore the idea of quality for the sake of quality, and remember that we are working with people, you can ask "What makes them happy?" And you might conclude, as we did, that "good enough" is the minimum level of quality needed to ensure the comfort of all the parties involved in the process. But that still leaves some variation on what the minimum threshold is. And it isn’t clear who makes that decision.

Is it the…

 

  • End users: They decide on quality indirectly, through the purchasing decisions of the collective. Their vote is cast in the form of dollars, clicks, views, time per page, and attention. But do end users care about quality anymore? For them, a translation either "flies or does not fly".
  • Clients: They decide on quality directly, by what they are willing to buy. In the "good enough" culture, comfort is the supreme value. The email exchange we described justifies this very controversial question: Do they care about quality anymore?
  • Translation providers: We decide on quality directly, by what we are willing to offer.

Until now, we have neglected this responsibility as an industry. If every stakeholder is happy, then it’s "good enough". And this shouldn’t be too alarming, if it weren’t for the fact that the quality threshold seems to be decreasing by the minute, due to what we call the fast-foodifying of translation.

Image 1: When global content meets mass production

 

The term "fast-foodifying" is useful in this context because the fast food industry is a clear example of an industry at the service of comfort. With comfort as the driving goal, we describe the "fast-foodifying" of our own industry through three distinct trends:

 

  • The separation from the original purpose of the occupation and the incursion into tech. This transition has been a clear trend with the biggest LSPs in the industry, most of whom own and commercialize proprietary translation management systems. Today, technology in our industry is much less likely to replace the doers (translators) than it is to replace the coordinators (project managers). 
  • A decrease in average project size, as well as the shelf-life of translations. The former benefits speed-centric models built around automatization of the management process and the latter relieves the burden of quality.
  • "Would you like fries with that?" We are noticing lower and lower per word rates, made possible by crowdsourcing, machine translation (MT) + post-editing (PE), and TM leveraging –regardless of the quality of the TMs. These do not include additional services, which were traditionally a part of an all-inclusive rate (file preparation, editing, review, etc.).

These three trends build on the realization, by LSPs, that the content creator and the end user don’t care about translation, they want information ("good enough"). And they have given rise to a business model that takes the decision-making out of the hands of the practitioners and into the hands of the market. The cornerstone of this business model is the "levels of service" approach, which sees small and medium LSPs atomizing their services into an ever-increasing number of categories with ever-decreasing levels of quality.

There have always been implicit quality expectations associated with different types of projects, degrees of visibility, target audience, regulatory compliance, etc. A good example of this is the difference between pre-sale and post-sale content, where quality is valued higher in the pre-sale scenario, as it has a direct impact on revenue, whereas post-sale documentation such as maintenance manuals typically does not produce any additional revenue. On the LSP side, these differing quality expectations traditionally had matching levels of service "disguised" under acronyms such as TEP, TE, MT+PE, etc. Or simply under the assumption that client review will become a billable item rather than a free fix.

But this model of implicit quality expectations has given way to explicitly stated levels of quality that clients can choose from, at differing price points. We see this simply as a response to the "fast-foodifying" of our industry, which inevitably encourages the industrialization of every player in the chain, including the suppliers to big LSPs and their competitors.

If quality is the one resource we share as an industry, as well as a crucial element in the control over our profession and the conditions under which we work, then the levels-of-service approach is undermining our credibility as the guarantors of said resource, and is de-professionalizing our industry.

 

Looking into the future

The truth is, there IS one level of service that would cover any of the translation needs in the market: the best quality. We can all agree that the perfect translation would certainly fit any purpose and make all stakeholders happy. It would also serve the language itself, the professionalization of our industry, and the shielding of working conditions. However, it is not always possible to provide this level of quality as it comes at the expense of speed and cost. This is where disruptors play a crucial role.

Each industry needs disruptors to keep innovating. Disruptors evaluate typical processes and identify bottlenecks, redundancies or inefficiencies. By using a non-traditional approach to content management, they can offer tailored solutions and "square" the traditional triangle of SCQ. But in order to disrupt, you need an understanding of what you are disrupting and if you are adding value or not. The industry has seen a dramatic increase in mergers, consolidation and acquisition of small and medium LSPs by players from other industries, especially finance. This is not the kind of disruption we need, as it is not people-focused, it doesn’t treat languages as the extension of cultures, and it treats all content alike. We, the practitioners, should reclaim our industry by setting our own quality standards and erasing the term “good enough” from our playbooks.

 

References

[1] Europe’s leading role in Machine Translation: How Europe is driving the shift to MT. June 2016. Arle R. Lommel and Donald A. DePalma. Cracker Project, EU.

[2] 10 Key Marketing Trends for 2017 Report. IBM Marketing Cloud.

[3] 2017, a year in bingeing. Netflix Media Center

[4] Google I/O 2016

[5] The digitalization of the world – From edge to core. IDC White Paper. November 2018 

[6] https://slator.com/deal-wins/usd-0-21-per-word-americas-translation-rate/

[7] https://slator.com/industry-news/the-magic-number-is-eur-0-15-translator-rate-survey-released-in-germany/