In part 1 of this article, István Lengyel, localisation industry expert and founder of memoQ and continuous localisation platform BeLazy, describes the global collaboration network built for localisation services and explains the value added by each partner in the network.
In part 2, he’ll examine the role of technology at each stage of service provision and discuss how it ties in with the tasks performed at that stage.
Same old story: buyers, MLVs, SLVs, freelancers
When I was younger and more revolutionary, I loved to question everything in localisation. I wanted to replace translation memories with bilingual documents, because it seemed so logical. I wanted to push LSPs towards reusing translations for multiple clients, because it also seemed logical.
Some practices remain unchanged because the benefits seem smaller than the effort required to bring about the change.
Neither is a reality today, and I learned to live with the path of least resistance: some practices remain unchanged because the benefits seem smaller than the effort required to bring about the change. Very few concepts in the translation industry have changed over the last two decades, whereas the technology that translation buyers use has moved on significantly.
When I joined this industry over 15 years ago, the parties in the localisation chain were translation buyers, multi-language vendors (MLVs), single-language vendors (SLVs) and freelance translators – and that remains the same today. It’s interesting to see what function each of these parties has had in the process, and how technology has changed – and continues to change – their roles.
At the advent of localisation, large IT companies started pushing out multilingual content, which gave rise to large MLVs like Lionbridge, Moravia and SDL. The majority of the content came from desktop software that was delivered on CDs, with full documentation and often a manual. Desktop software followed the waterfall model – with specification, development, testing and bug fixing, localisation and distribution being very distinct steps. The timing of a release was crucial, as large press conferences or user conferences attracted many target users.
Remember visual localisation tools like Catalyst and Passolo? There were plenty of them: Multilizer, Sisulizer, RC Wintrans… strings had to be positioned and resized, scrolling was often not an option, Asian characters and bidirectional writing for Hebrew and Arabic was a challenge. Consistency between the interface and the help was hard to achieve.
Today, internationalisation has a very distinct set of challenges. The prevalence of Unicode made the support for writing systems easy. Today’s challenge lies in mobile-ready, responsive interfaces of web pages. Cloud-based systems support remote working, and instead of large ERP systems, many companies work with a variety of systems to store their content.
Agile development is practised by most enterprises, not only IT, but also those where IT supports the core activity. Continuous integration/continuous delivery drives the need for quick-turnaround localisation. With the explosion in the volume of content, machine translation has become part of the workflow.
Buyers continue to face the problem of isolation and silos, not so much from the technology perspective but from the ignorance perspective. By necessity, their managers continue to collaborate with people, both within and outside their company, with various levels of understanding of the bigger picture in multilingual content management.
The local power plant or food manufacturer also needs translation services, though perhaps on a smaller scale.
When people in the translation industry speak of buyers in general terms, they tend to think of IT companies. In reality, the local power plant or food manufacturer also needs translation services, though perhaps on a smaller scale.
The research on the fragmentation of the translation market indicates that not every buyer needs the sophisticated expertise of the largest translation service providers. You only need the services of auditing firms like KPMG if you start optimising tax cross-border, and the same applies to buying localisation services.
Multi-language vendors (MLVs)
MLVs emerged because they had the technical knowledge to understand and support a variety of clients. They employed engineers who were able to help the buyers in the IT sector. Most of today’s successful MLVs have grown on the back of their clients’ growth. They were the ones who started providing staffing for their clients and they built their service on two cornerstones: dedicated customer teams and a customisable technology stack.
Buyers’ content repositories have changed significantly and the rise of APIs and the concept of platforms has enabled software to play better together. MLVs always pursued the know-how and control of the technology stack, and over the years, technology became the most important driver for LSP growth. Large MLVs partnered with content technology providers to create efficient localisation solutions for their clients. Technology is a great lock-in because changing it requires investment in time and money.
[MLVs] understand the bigger business picture better than most smaller translator-led shops do.
In the footsteps of continuous integration came continuous localisation: buyers sending increasingly small chunks of text for translation. A handful of large MLVs handles the bulk of the software and website translations for large enterprise clients, and they do it under time and price pressure. The role of the MLVs has not changed much: they understand the bigger business picture better than most smaller translator-led shops do.
And then there is the ripple. Large tech companies like Facebook believe they can build better technology for their localisation needs than the MLVs can provide and they start contracting directly to SLVs. At the same time, Spotify moves from an SLV outsourcing model to an MLV outsourcing model. Other giants, like Netflix, build their own technology but do not change their outsourcing model.
More often than not, procurement at large enterprise buyers doesn’t allow for single-sourcing translation, thus the buyer’s partner MLVs have to collaborate with each other in the service provision. Many buyers have their own shared service centre for their translation needs which then engages a mix of MLVs, SLVs and freelancer translators.
In a nutshell: MLVs have always been the guardians of process and technology. They have the knowledge of internationalisation, workflows for dozens of languages and the ability to test and quantify localisation outcomes. MLVs are also able to specialise in niche workflows for regulated industries or computer gaming and can work with large procurement teams at multinationals.
Single-language vendors (SLVs)
Single-language vendors are defined by the fact that they translate into one target language or a group of regional languages. If the MLV is the tech shop, the SLV is the language quality shop. Interestingly, the same company often wears both hats: for example, in Spain, most of the local translation orders are from English, French, Italian, German or Spanish into the same EFIGS and into Portuguese, Polish, Russian, Turkish, Chinese. Any other language combination is in low demand. Any LSP in the region can function as an MLV by managing the work between the aforementioned languages and outsourcing the occasional extra languages to other SLVs. At the same time, they act as an SLV by offering translation services into Spanish and Portuguese to large global MLVs.
Freelance translators turn into small SLVs, defeating the original purpose of the disintermediation.
What makes SLVs attractive partners is affordable quality: they have the staff to review and assess translations efficiently, which in turn helps them keep vendor costs relatively low. Their operational excellence – focused resources, standardised processes and streamlined configuration of translation technology – helps them build cost-effective solutions. And then there is the ripple: MLVs like the idea of disintermediation and seek ways to work with freelance translators directly. There are three major challenges in this:
- The MLV’s inability to assess the quality of all language pairs in-house.
- The fact that freelancers quote higher rates to foreign clients and for ad-hoc work, which means that only the largest MLVs with offices in multiple countries achieve favourable rates.
- The admin and management overhead increases with more suppliers.
I’ve seen many MLVs making a push towards working with freelancers, and then going back to working with SLV partners. Typically, an MLV builds an outsourcing platform and creates an open marketplace for registered suppliers to grab the jobs on offer. In response, some freelancers start taking more jobs than they can possibly translate themselves and pass them on to peers or mentees. The vendor managers may notice but won’t interfere as long as the delivered quality is acceptable. Thus the freelance translators turns into small SLVs, defeating the original purpose of the disintermediation. And the MLV is back with SLV outsourcing.
In a nutshell: SLVs are the best solution for finding, maintaining and training talent in a specific target language. Their unique added value lies in their efficient vendor management for a limited set of languages.
Until recently, one of the greatest concerns for freelance translators was language technology: how to install, maintain and pay for all different pieces of software. They tended to master one translation tool better than the others and wanted to perform most of their jobs in that environment for improved productivity.
Today, translators work in a smaller variety of desktop-based tools and with more online tools. Online tools are simple and very similar to each other. The freelance translators’ main concern now is the size of jobs: they receive many 5–100-word tasks every day. The best translators are the ones who can communicate efficiently with their project managers and clients.
Translators are burdened with numerous email notifications about tiny jobs and can become frustrated if clients think they are easily replaceable. Job auctions, which from the clients’ point of view facilitate the speedy placing of a job, can be viewed by a freelance translator as an insult to their uniqueness and expertise.
Turnaround times have decreased with project sizes, and no LSP really knows their freelance translators’ availability because they all work for multiple clients. Translators are asked to update their availability in a portal, but many say that if they were to do that for all their clients, they would hardly have time to work on any translations.
[Freelancers] prefer clients who see them as a valuable individual, take them seriously and listen to their feedback.
And then there is the ripple: due to the stress of having to place jobs quickly with a vendor who is contracted to never turn them down, clients choose SLV or MLV partners over a vast pool of translators – let them take care of the scheduling and translator selection problem.
In a nutshell: freelance translators have always done the bulk of the actual translation work. Today, they work with translation, revision, review, post-editing and transcreation. They prefer clients who see them as a valuable individual, take them seriously and listen to their feedback.
There are fewer in-house translator positions available now than in the past, but there are still reasons why all types of companies employ them: data security, restrictive translation environments, quality control, their guaranteed quantifiable availability and deep subject-matter expertise, the return on investment in terms of their training and continued professional development. In-house translators can be employed at the buyer organisation, MLV or SLV.
In part 2 of this article, István will examine the role of technology at each stage of service provision and discuss how it ties in with the tasks performed at that stage.