On July 2, 2020, translation productivity and management software provider Memsource announced that one of the world’s largest private equity firms, US-based Carlyle Group, had acquired a majority stake in the company. Carlyle Group is one of the world’s largest private equity firms with over USD 80bn in assets under management and currently 181 portfolio companies. The transaction is notable as one the largest ever in the translation and localization technology space; and Memsource is a provider that is positioned at the core of the language industry, connecting language service providers (LSPs), enterprise clients, and linguists. To understand the context, strategic rationale, and future plans behind the deal, Slator spoke to Memsource Founder and CEO David Canek and Fernando Chueca, Managing Director at Carlyle Europe Technology Partners, and Carlyle’s lead on the transaction. The transaction establishes an important benchmark in the pricing of language technology deals going forward, so valuation was top of mind for many when the news broke. While both Canek and Chueca declined to comment on the deal’s financials, they confirmed Memsource’s revenues came in at around USD 8m+ in 2019 ARRs (Annual Recurring Revenue, a popular SaaS metric) and is on track to exceed USD 11m in 2020. According to Chueca, the business had grown around 40% YoY in 2019 and Canek said it was “highly profitable.” Based on this, Slator estimates that the transaction valued Memsource at around USD 45m. Canek said discussions with Carlyle began in August 2019. “We were not completely new to the M&A world,” he said, pointing out that once a business reaches a certain size, “you get contacted by different funds all the time.” Having been bootstrapped since its founding nearly 10 years ago, Memsource had considered the possibility of raising money on a few occasions — but “we didn’t really need it,” Canek said. Thanks to a few smart commercial deals, “we were really fortunate that we were able to grow the team to the 110 we are today without any outside funding,” he added. Memsource began conversations with a few funds between December 2019 and January 2020, and then “started talking more and more to Carlyle,” Canek said. They were about a month away from closing the deal when Covid-19 hit. At one point, early on in the pandemic, Canek began to think “oops, all this work [on a potential deal] was for nothing.” But, he said, “it’s great that we’ve always had this option to just go ahead with Memsource. We didn’t have to find an investor at any cost.” As it transpired, Memsource managed to “withstand the Covid outbreak for four straight months with very steady growth in our ARR,” Canek told Slator. Eventually, “we were able to close when it became clear that it wasn’t the end of the world,” he said. From an investor’s perspective, Chueca told Slator that they first met Memsource in December 2019 and that they had been able to meet in person both in London and in Prague, where Memsource is headquartered. “We had a lot of interaction before the [travel] restrictions kicked in. If we hadn’t had those face-to-face interactions, it would have been much harder. The fact that we were able to build that personal relationship was essential,” Chueca said. The second half of the transaction was done over video, he added. In middle to late March 2020, which Chueca described as “the time of maximum uncertainty” in terms of the pandemic, Carlyle’s approach was to step back and use the opportunity to understand how Memsource would cope. They wanted to know, Chueca explained, “How is the business is delivering its service? Are you able to deliver the service? Is the platform up and running? Are people using it? Are customers still using it? Is there any change in the ecosystem that will have a material impact on the business?” During that time, the two parties spoke at least once a week and went over a number of metrics together including cash collections and volume of words processed. “The good news is that the business was very, very resilient,” Chueca said, pointing out that Memsource “kept on adding customers during Covid, which gave us extra comfort that it was a very good business.” Asked why he felt the time was right for Memsource to take on an investor, Canek said “the main factor was that, after 10 years, we wanted to shake things up a little.” One of Memsource’s six shareholders had also decided not to continue, which was another contributing factor, he said. Canek’s desire was to “transform the company a little bit. Not revolutionize, but transform — and maybe find a partner for the business and get some additional cash on the balance sheet.” Having had advisory firm ACG Partners referred to them, Memsource began weighing up the options and felt that financial investors were an interesting route because “there’s quite a lot of them so you can choose,” Canek said. While he was quick to rule out a sale to an LSP because “it just wouldn’t work,” Canek believed a financial investor would bring a level of neutrality: “It’s not an LSP, a linguist or an enterprise customer, and it doesn’t compete with our customers — that’s the main thing.” For Chueca, the decision to invest was based on having “confidence that there is an underlying market demand.” They arranged calls with Memsource’s customers and concluded that “customers are very happy with Memsource. We felt that this was the right [business] to back.” Chueca gave his take on the landscape of translation management and productivity software: “We see that this market is thoroughly fragmented. We see that there is a good level of demand from customers having a tool that does everything that Memsource does.” The Memsource deal is a notable one for the TMS and translation productivity part of the language industry, particularly, as Chueca said because “this is a sector where there’s not been too much M&A. Or if there’s been M&A, it has been the big LSPs buying the small providers.” However, Chueca told Slator, “we’ve taken the view that it’s a very interesting space and that technology is going to keep on disrupting; we believe that there is a role for an independent provider of technology.” Memsource has now become one of the portfolio companies of Carlyle Europe Technology Partners IV (CETP), a EUR 1.4bn fund that Carlyle plans to invest across more than a dozen businesses. CETP has until 2025 to deploy capital and another five years to exit, which leaves “plenty of time to generate money,” Chueca pointed out. The plan is for CETP and Memsource to interact on a weekly basis and to organize two board meetings a year, Chueca said. The board will consist of Chueca and Petr Rieger from CETP as well as two representatives from Memsource management and a chairman that will be brought in from outside. Chueca also told Slator that the 20-strong team at CETP will be “relatively hands-on,” but that day-to-day business will be run by Memsource management. Canek confirmed that he does not intend to step back and “the plan is really for the management to stay on and actually remain shareholders in the business.” Canek acknowledged that his role is likely to evolve into that of a more traditional CEO as he hands over ownership of marketing, sales, and product to new senior hires. Having already brought in a Chief Product Officer and a Chief Marketing Officer, Canek said, “our main focus will be to take our sales to the next level,” and more senior hires will follow. When it comes to Memsource’s involvement with LSPs, Canek was of the opinion that, although one “could view the relationship between LSPs and translation tech providers as one of conflict,” language tech providers and LSPs “need each other.” Memsource has “a great partnership with many LSPs already and we want to double down on our partnerships and figure out how to partner even better,” he said. According to Carlyle’s Chueca, LSPs “drive enterprise business” and are “an absolutely core part of our strategy.” He also acknowledged that Memsource’s customer mix may ultimately skew more toward enterprise clients given that there are more of them than there are LSPs: “Some enterprises will become customers through the LSP relationship, [while] some others may want to work with five different LSPs.” By contrast, “linguists are not an area where we seek to monetize heavily,” Chueca said. Rather, through freelance translators, they hope to build “a community of users that like the product, use the product, and can also [provide] a bit of critical mass.” Canek also pointed out that Memsource continues to invest in their translation tool and provides native versions for web, mobile, and desktop with support for Windows, Mac, and Linux. Given all this, Canek believes Memsource has “probably the best support for translators.” While both Canek and Chueca see a strong business case for partnering with LSPs, neither has any desire to become one. The two feel strongly about avoiding the services component entirely and ruled out the possibility of operating as a tech-enabled LSP (such as Lilt or Smartling) or a translation crowdsourcing platform. For Canek it is a matter of scalability: “If you focus on human translation, you’ll hit a limit at some point as there are a finite number of translators in the world and so this approach doesn’t scale.” It is also a question of laser-focus, and Canek said they intend to focus fully on automation and “machine translation (MT) enablement,” which aims to maximize automation across workflow management and translation. The eventual goal is to “make machine translation seamless, [although] we’re very far from it,” he said. Greater MT enablement will come by integrating new features, Canek said. For example, MT quality estimation (MTQE), which allows users to understand when MT output needs a human eye, and engine selection, where users are guided on which engine to select based on the language pair and domain. “This is where our AI research is going and this is our vision for what we want to develop,” he added. Enabling MT is very much a work in progress, and “we will need to have human translators for many years to come — forever,” Canek said. Meanwhile, Chueca articulated a holistic vision for the company where Memsource serves as a pure-play technology provider for LSPs, linguists, and enterprise clients. “The vision that we have is for [our] systems to effectively be at the heart of it — coordinating, automating the workflow across those three different stakeholders, and also increasingly integrating machine translation,” Chueca said. For Chueca, MT progress is very much an opportunity, and something that he believes will drive the uptake of tools such as Memsource. Although the technology is not perfect, it is still improving and “nobody in the industry should fight against machine translation,” he said. According to Chueca, finding smart and efficient ways to determine which content needs human processing will prove crucial. He believes that the trend of declining per-word rates will continue, but that overall volumes will explode — “I don’t think that the industry will shrink.” What will happen, Chueca said, is that users of translation will start to become more discerning: “Budgets will remain in place, but people will have to be smarter about what content, when, and how they get it reviewed.” And, he added, although MT will play an increasingly important role, some critical content will still need to be reviewed by experts, such as legal content and pharmaceutical documentation. Chueca told Slator that CETP is “seeking to deploy additional capital behind Memsource,” which is currently one of the smallest businesses in the fund’s portfolio. “We feel that we can accelerate growth, or complement the organic growth by providing additional capital for M&A,” he said. Speaking on what form future M&A might take, Chueca said it would be important for any business they acquire in the future not to compete in the same space, but “to the extent that there is a rationale for putting businesses together, we will.” Although MT enablement is at the core of Memsource’s vision, this does not mean that they plan to buy an MT provider or build their own MT engines. Chueca said they have no desire to compete with the “40-or-so machine translation engines that are on the market.” Rather, M&A could take the form of an acqui-hire, according to Chueca, and they would also consider companies that would expand Memsource’s geographic footprint. Finally, he identified niche providers that offer additional product functionality as a third potential target for M&A.