In late August, UK-based RWS Holdings announced that it would acquire fellow UK supplier SDL in an all-share deal. We analyzed the importance to the market of that deal in “RWS Absorbs SDL – The Business Story.” Here we analyze the impact of the acquisition of SDL’s language and content management technology – and recommend some next steps for the soon-to-be new owners of storied brands like Trados and WorldServer.
First, An Inventory of RWS and SDL Technology
Acquiring SDL immediately transforms RWS into one of the leading suppliers of translation technology and a player in structured content management. Language services accounted for 70% of SDL’s £376.3 million in 2019 revenue, while its two technology units – language and content – earned 14% and 16% of the total, respectively.
- SDL technology offerings. The langtech offerings include well-known and some lesser known brands in translation management (SDL TMS, WorldServer, and MultiTrans), translation productivity software (Trados and MultiTerm), machine translation (BeGlobal and SLATE), and artificial intelligence (Linguistic AI). The content technologies are web content management (Tridion) and structured and technical content management software (Contenta).
- SDL financial performance by business unit. Performance at the mid-year 2020 point was varied – language services revenue was down 2% compared to the same period in 2018, while langtech was on par with 2019 results and content technologies grew by 2%. Of course, COVID-infected 2020 revenue will be off the previous growth patterns – its year-over-year growth from 2018 was 20% for language services, 8% for langtech, and 10% for content technology.
On the acquiring side, RWS brings technology to the table, but the company doesn’t break out its revenue for most offerings. The exception is its PatBase global patent search platform which earned it £4.5 million in 2019. Most of its other software works in support of RWS life science and patent offerings, such as AOP Connect (repository search) and the Inovia online platform for foreign patent filings.
However, three RWS offerings stand out: 1) neural MT provider Iconic Translation, acquired earlier this year; 2) AURORA, a TMS optimized for life sciences with asset management including translation memory, glossary and terminology, and a style guide; and 3) Global Artificial Intelligence (AI) Services, which provides a wide range of AI program support, data provisioning, data evaluation and enrichment, machine learning model validation, and AI model deployment and evaluation. Why do these three stand out? The first two arguably overlap with SDL products, while the third – the AI services – dovetails nicely with SDL.
The Mission: Rationalize the SDL Portfolio
Over the last 15 years SDL has acquired an array of translation management systems, such that today it offers SDL TMS, WorldServer, and MultiTrans to enterprise buyers. It has also been selling lesser-powered business process management and workflow solutions to LSPs. And it will sell its translation memory and terminology tools to any of the above. At the same time, it’s been working on various replacements designed to unify the TMS and WorldServer universe on its modern Language Cloud, which also underpins its latest Trados version.
All of this technology will now shift to RWS along with the attendant urgency to update those three legacy TMSes. And RWS will complicate matters with AURORA, its own TMS. Another major offering of SDL is its machine translation solution, which will collide with RWS’ recently acquired Iconic Translation. RWS has to solve this surfeit of TMSes and MT engines, de-dupe its offerings, and continue fighting against strong competition in each of its technology areas from faster-moving, smaller ISVs such as Memsource, Smartcat, Wordbee, and XTM on the TMS front, along with hybrid LSP-ISVs such as LILT and Smartling.
Note: We have to wonder about the timing of the Iconic Translation purchase in early June for €17.8 million. Unless its specialization in intellectual property fills a specific gap in the broader picture that SDL does not meet, the dual acquisitions speak to the velocity with which the all-share deal was negotiated – a couple of months at most, noteworthy for a deal of this size.
Then throw into this mix Plunet, XTRF, and various and sundry other competitors for specialized functions and the RWS SDL-derived technology division will be fending off attacks from all sides. And it’s not just about competitors, but also supporting long-term partners such as Kaleidoscope that add functionality to SDL products.
We’ll call this technology division “RWSDL” because it will be quite a while before RWS can re-brand the langtech and content management parts of this acquisition.
- It’s time to get tough on TMSes. Our recommendation to SDL when it had its first TMS fork with SDL TMS and WorldServer in 2008 was to bite the bullet and subsidize SDL TMS users’ conversions to WorldServer. That didn’t happen and the pain has only gotten worse. Twelve years and another TMS acquisition later, RWSDL management will have to make the hard decision to declare end-of-life on the older systems. This move will anger some buyers, but they’ve been expecting this shoe to drop for years. An accelerated transition to the Language Cloud program would help. SDL CEO Adolfo Hernandez had been pushing for rationalization, but was hindered by corporate inertia. The danger is that with him out of the picture, the technology division may become a zombie – this will affect everybody in the language industry, not just the internal players in the new combined entity.
- MT duplication should be easier to resolve. With Iconic still fresh in RWS management’s mind, SDL’s MT team will now have to win mindshare from their new bosses. So much of the machine translation effort happens behind the scenes or under the covers which means there’s room for different streams of AI and natural language processing (NLP) to co-exist, thrive, and intermarry. RWSDL must move quickly to integrate the teams, which likely means treating Iconic as a pricey acqui-hire. That said, SDL has had a hard time capitalizing on BeGlobal and the other grandchildren of its 2010 purchase of statistical MT pioneer Language Weaver. A more aggressive RWS Holdings applying the MT portfolio to its clients could be a huge gain for MT and productivity.
- Content management is important enough an issue for another blog. SDL’s website and structured content management technologies were acquisitions that reflect an earlier epoch in its strategy, but today account for more revenue than language technology. Why? The value of content management for customer experience management and the core of omnichannel marketing, commerce, and information publishing is immense in global enterprises. Tighter integration between a singular TMS offering and its CMS solutions would benefit everyone. But even that will require some hard thinking. Global content management is a different game than langtech – RWSDL will have to learn the rules that SDL flubbed when it pushed heavily on its Tridion product several years back and in the process alienated some big content management business partners.
- Spoken-language technology is missing in action. Neither RWS nor SDL have a presence in this sector but building a robust interpreting delivery platform (IDP) is essential to a global content service provider (GCSP) story about managing omni-channel experiences. Doing so will require bringing in new management and cultivating talent. This area, where competitors such as thebigword are doing well, will be a challenge and they will be learning from the ground up (“Interpreting in the COVID-19 Business Climate”). Expect another acquisition to address this lacuna in the combined company’s capabilities.
Job #1: Satisfying SDL – and Future RWSDL – Langtech Buyers
SDL sells its langtech software to both end-buyers and other LSPs. Enterprises and government agencies typically buy TMS and MT solutions as part of their information publishing, marketing, and commercial efforts. LSPs buy translation productivity tools and business project management software, while freelancers make up the bulk of SDL Trados sales. Therein lies a challenge for RWSDL on solving software-agility, competitor-supplier, and restricted-usage problems that SDL didn’t manage well:
- Software agility. SDL has been more agile in acquiring technology than developing its own, and not very good about making hard decisions about when to pull the plug. SDL under Hernandez had started this far-from-complete-process although we do see some promising signs of success in the form of SLATE and cloud-based Trados. But supporting the multiple code bases of TMS acquisitions while developing a replacement over the last decade has sucked momentum, energy, and cash from the business. Overcoming software development inertia once and for all represents the number one langtech challenge for RWS CEO Richard Thompson and his team.
- Competitor-supplier tension. SDL supplies essential software such as Trados to rival LSPs and many of the world’s freelancers. Paying a competitor for software sticks in the craw of some LSPs and freelancers, especially since their purchase subsidizes SDL’s in-house use of the same technology as it competes with other LSPs for the same language services business.
- Restricted usage. SDL wouldn’t sell its aging crown-jewel TMS technology to its LSP customers, instead providing them with less-performant software. MT and Linguistic AI will likely become a similar issue for this competitive dynamic. RWSDL has to resolve this one quickly as faster-moving software rivals target the LSP and freelance communities. That opens doors for TMS-sans-services vendors willing to sell to anyone with a translation management or machine translation requirement. The tens of thousands of profitable businesses that depend on these technologies make this a desirable market.
What Should RWSDL Do Next?
RWSDL will have to work hard to convince SDL technology users that it will do a better job evolving and replacing the software with newer solutions than SDL has. In the face of this challenge, it has a few choices:
- Double down. RWSDL management could decide to get more aggressive on the software front, investing more cash and resources to move Language Cloud more quickly into accounts and retire older technology. However, it would also have to resolve the tensions inherent in the competitor-supplier dynamic. Nonetheless, if it chose this path and somehow mitigated those tensions, RWSDL could stress greater development agility and leverage the “small AI” benefits that could come from RWS’ and SDL’s free cash flow (£90 million in 2019, or about US$120 million) combined with the rich, reliable data and content flows to train its software.
- Spin off the technology division as an independent company. Selling off the technology division would resolve the competitor-supplier issues. However, it would also deprive RWS Holdings access to core technology, input into the development process, subsidized prices, and ultimately control of its technology stack. With langtech essential to its service business success, such a solution wouldn’t fly.
An alternative to a spin-off would be creating an independent strategic business unit (à la GE’s long-term practice). Such a BU that has a dedicated technology focus could open up all the RWS and SDL APIs and create a common platform for language services – a de facto nod to Trados’ market position 10 years ago, and an opening for Language Cloud that could result in a rapidly growing market.
Think of the Roku streaming media player, initially a Netflix product, which was spun off as an open platform for any streaming service. That decision allowed – for better or for worse – the massive deployment of streaming media to the world on a single box rather than on separate hardware for each service. Translation could follow the same path with a common specification for files, content, code, and other elements of translation and localization (“The Interoperability Dilemma”). If successful in bringing competitors to a common approach to data, RWSDL could do what dozens of international translation standards committees have failed to do. RWSDL could enhance its play for a common platform by coordinating with more successful groups like the XLIFF Technical Committee.
Crossing the Chasm – Buckinghamshire to Berkshire
As executives make the 13-mile journey for transition meetings between RWS headquarters in Chalfont St Peter and SDL HQ in Maidenhead, they will ponder how to maximize the value of SDL’s technology portfolio. SDL has been engaged for a few years in a huge and long-running pivot to cloud-based technology for all of its core products. If RWS can continue what Hernandez started, and position the technology offerings so they aren’t caught between competing business interests, the acquisition could actually be transformative for the tech sector.
Additional Author Credit: Dr. Arle Lommel, Senior Analyst, CSA Research