David Čaněk, Founder and CEO of Memsource, joins the pod to talk about the recent acquisition of localization SaaS provider Phrase.
David walks us through Memsource’s milestones, from their launch in 2010 as a cloud-based TMS to growing as a bootstrapped company until July 2020, when investment firm giant Carlyle took a majority stake in the company.
He also shares insights into the acquisition of Germany-based Phrase, discusses why their developer-targeted product is complementary to Memsource’s offering, and outlines the direction of the company.
The two also discuss TransPerfect’s 2020 financial results which saw their revenues grow 11.5%, adding USD 88m to its topline, and how they won’t be following in Lionbridge’s footsteps in selling off their data-for-AI division.
Stream Slator webinars, workshops, and conferences on the Slator Video-on-Demand channel.
Florian: David, you are the founder and CEO of Memsource, the translation productivity and translation management technology firm. Glad you could discuss this transformative acquisition of Phrase, a Hamburg based software localization technology firm. First, tell us a bit more about the background, history, some of the key people, and key moments of Memsource.
David: Sure, I founded Memsource in 2010. We had our first paying customer in 2012, I think that customer stayed with us for many years, so the software already worked but it was a lot to do. We had a team of 10 to 15 people in 2012. Every year we would add 10 more people, sometimes 20, then at some point 40. In 2016 there was a major milestone where we got some extra cash, not through investors or funding, but actually through some commercial deals. We decided to expand the company to just below 100 people.
Around 2017 and 18, gradually, we started the AI team, the mobile team, the integration team, a lot of teams. All the way we were bootstrapped, we were always just above breakeven, trying to run a company in a very elite way and in July 2020 that was really the big milestone, when Carlyle invested in Memsource. There was also the plan about the acquisition so that leads to the news that we shared about Memsource and Phrase joining forces.
Florian: Just to go back a little bit, what is your background? How did you end up in this industry? Are you an engineer? Are you a language person? What got you started in this?
David: I am a language person by training, I studied translation and I worked as a translator during my studies, but then I thought that it was a little too solitary for me. I went to work for different companies, a tech company and also an internet startup, developing some interesting language technologies. Then I ended up working in the translation industry, in a smaller translation company in Prague. I saw the kind of technology that was being used and I saw that opportunity. I really liked technology, I worked in some product roles in tech companies before and I was wondering if maybe this is the opportunity to do something on my own which I enjoyed. Combining my language background with technology, this was my passion. Here I am, I have been with Memsource for over 10 years now.
Florian: How do you deal with the technology component? How did you get up to speed on some of the technical aspects of this?
David: I am not an engineer, even though some people think that. I probably give that impression a little bit, although not intentionally. I would say that is the only thing I am not doing it in the company, other than that I am doing everything. I think we had great engineers in Memsource and great AI researchers, great tech people overall. That is not something I need to worry about. I am pretty comfortable not coding at Memsource, however, it is very important to understand the product and everything around products related to customer sales marketing.
That is what I enjoy, I am pretty much a focused product branded CEO. This was also an important role that I had in Memsource from the beginning. Obviously, we did not have a product team, I was the product. I was also the tech support, I was also the marketing and sales at the beginning, but then luckily other people joined and helped out. Now all of a sudden we have this great team based in Hamburg.
Esther: Can you tell us a bit about Phrase as a company and a product?
David: Obviously, I knew Phrase and I would meet and chat with Frederick, one of the three founders, roughly every year at a conference. I noticed that we have some similarities and they are also bootstrapped and they also run the company in a wider efficient way without external capital. One of the reasons why we decided on partnering with an investor, in the end, it was Carlyle because we found that was a great match. Apart from organic growth, an acquisition was always on the table, that was part of the plan right from the beginning when we partnered with Carlyle.
We looked at a lot of companies in the space and in the end, we saw that Phrase is most complimentary with Memsource. They are a software organization platform and so they really focus on software developers and the needs that they have around localization. Our focus is much more general, we cover the full range of use cases for enterprises, so Phrase has this great in-depth focus on this fast-growing vertical. They have a very similar company culture, like Memsource, and they are doing a great job in software localization and they are excellent, so it kind of clicked.
Florian: It has been six months since Carlyle, is that when you started looking at a long list of targets or did you have this in mind earlier? You knew Frederick from before, but how did you get started on the actual M&A, or the acquisition talk and then how does this work doing a deal in Covid-19 times?
David: We did the transaction with Carlyle also in Covid-19 times, so we already had this experience. For the most part, not in such an extreme way as today during the pandemic, we already were in a Covid-19 mode before Covid-19 because we have customers all over the world. We already had the whole set up for this, and with the transaction, that is a whole different level.
It helped that both with Carlyle and with Phrase, there was some initial personal contact. Although I have to say that with Phrase it was very limited because I knew Frederick, other than that, I did not know anyone. I do not think no one from Memsource or Carlyle knew anyone from Phrase in-person. I had the only kind of relationship with Frederick and I only met the CEO of the company, also one of the founders, Wolfram, once in person and that was in September. It was just as everything was closing down and we planned a meeting for October. I called Wolfram and said we are coming tomorrow because I think we will not be able to do it later. That was the case, it was very difficult. That was the only time we visited Phrase in person, but then everything else was on zoom.
Esther: Do you know what Phrase’s founders and management team are planning to do now? Are they staying on with the business?
David: Yeah, we discussed all this in great detail and the fact that we did not meet a lot in person does not mean that we did not have a lot of interaction. I do not feel like we missed out on some important information and I think our relationships evolved greatly. We just learnt how to deal with this, and it would be great to see everyone in person, but I think we were able to find ways to learn about each other. There are three founders in Phrase, Wolfram, Frederick, and Tobias and they are staying not just with the business but are staying as minority shareholders in the combined company. They already have key roles within the Memsource group and they continue to be driving all these important initiatives. This is a very important part of this transaction.
Florian: In terms of branding, are you going to phase out the brand Phrase, are you going to build on it? Is it going to be Phrase brought to you by Memsource? What is the plan?
David: Both of the companies are very product-driven, so it is the Phrase product and it is the Memsource product, I knew these are the two main things. The plan for the two companies is we want to have one company, we want to be one and we have already made some important steps toward that. I think it is going to be relatively fast and because we are so complimentary and we have many shared values, we will have one company. We will definitely want to have two products, the Memsource product with its brand and the Phrase product with its brand, so the Phrase product is not going anywhere and the Memsource product is also staying.
I think there will be an important evolution in terms of the additional resources that we have, all the surges that we have, and the bigger team, almost 200 people that we have. We will be able to focus even better on the needs of software developers by the Phrase product, either startups or our software development teams in big global companies and with the Memsource product. Memsource is the enterprise TMS, so we serve enterprise customers, all the full range of enterprise localization use cases, but also we serve translation companies and translators with the Memsource products. I think this will be the direction and there will be some interesting evolution and integrations and synergies but they will take some time, that will not happen right away.
Florian: In terms of the strategic fit, I do see that it fits very well, the software piece, localization piece and the general one. You mentioned it before, but this was one of the central aspects why you did that deal. How does it fit into the broader vision? Long term, maybe there are going to be some other pieces that are coming to it, but just tell us a bit more about the strategic fit of this business in the context of your strategy going forward.
David: There is certainly not one size fits all. It is important to recognize that the Memsource TMS, with all the complex needs that large enterprise customers have and large translation numbers, there is some complexity. Also, the need to cover the full range of localization use cases is really not a simple thing and it adds some complexity. This approach by focusing and making sure that startups have some localization needs and their localization maturities still need to evolve, they will be looking for something that is maybe not an overkill, that is a little more focused on them. It is part of the strategy, working with customers when they are starting with localization, maybe that is going to be more the domain of Phrase. Then also continuing to meet their more complex localization needs, not just in software development, but also in other areas with the Memsource TMS as these startups evolve into enterprises and covering better the full customer’s localization journey through these products that we now have.
Since you mentioned LSPs, we had these questions already with Carlyle Investment since Carlyle has such a big fund, it is one of the biggest investment funds. One question was, are you now going to also expand into providing human translation services and trying to expand into the LSP space? We were always very clear on this, no, that is not our plan and on the contrary, we want to continue to partner with LSPs. We hope we can be more interesting to LSPs. This continues to be the strategy, we want to focus on technology, we want to focus on automation. We have never been trying to add LSP services to our offering and would always route any leads or requests to our LSP customers.
Esther: In tech, you started to talk a little bit about some synergies that you might expect between you and Phrase, what would you expect in terms of synergies across the group? Would it be in product management, in marketing and sales? What do you envisage?
David: I think it is going to happen relatively fast and is already happening, it is amazing actually the speed of some of these things. For instance, we announced the acquisition on Monday and on Wednesday we already were on Slack. We are working very quickly to create one company so there is going to be a lot of synergies around the know-how that there is in Phrase, and the teams are incredible. We also have an incredible team in Memsource and some best practices have been implemented in Phrase and Memsource, and we will be able to combine them. This will be the same with leadership, we have some very complimentary situations there. One of the first teams that we combined is the marketing team, so we now have one marketing team, headed by Phrase VP of Marketing, Stephan, and so this will be very important in synergies that we have here. There will be very interesting products, we will be able to utilize some of the back end technologies and develop them for both products and free up some of the capacity for real innovation for software localization needs, software developers and for enterprises.
Esther: On those innovations that you mentioned, what kind of things can clients expect in terms of new functionalities and offerings in the short to midterm?
David: The first thing we want to reassure our customers is that there is no distraction for them or us due to combining companies. They will be able to continue to enjoy the great product, from both Phrase and Memsource, and the teams, the people that they are in touch with, there is no change there. As we move forward, we want to talk to our customers and understand what are the integration opportunities for which we could meet our customer’s needs even better. What are the changes that we should do? What are the improvements? What is the innovation? Prioritize in line with the market needs? These conversations are already taking place and obviously we have an idea, but I think the main message right now is this is great news, we will be one team very soon, almost 200 people and we will have more capacity, more know-how, more experience to meet our customer’s needs and innovate faster, and have a bigger impact also on the industry.
Florian: We have not mentioned AI and machine translation or anything like that yet. First, would you agree that maybe for Phrase, this is a smaller issue than for Memsource because you are in this big volume type of a business and Phrase is more on the smaller, more technical side of the software loc? What is your thinking for the next three to five years, where is this going, how are you going to integrate it?
David: Yeah, that is exactly where we are and that is exactly how we compliment each other. Memsource definitely is the one that has a lot of data and we have a lot of high-volume customers, so we started our in-house AI team in 2017. We were growing the team, making sure that our product and the entire architecture is optimized for this. The Phrase team was not focusing on this, so this is definitely an area where the AI team coming from Memsource will be able to cover both at some point for sure.
Florian: For software localization where do you see good first use cases or deployment areas for MT or post-editing, anything that is related to human translators?
David: Machine transition is one of the single biggest innovations that we have and that is already making an incredible impact on how our customer’s translate and localize. I think we published in December some stats from Memsource that show that 2020 was the first year when post-editing machine translation became the dominant translation method. If you want to translate professionalism content, MT post-editing is the default method now. This is where we are and I think it is just so obvious that the trend is clear to everyone.
Machine translation with all the investment that is going into the technology from all the players in the industry and outside the industry is going to get better and is getting better every year, and we can see it in our data. The post-editing effort to correct, to post-edit in machine translation is getting smaller and smaller every year. That is confirmed by our data, and so this is what will happen, MT will get better. Human translation will probably not get better, so the actual production of the translation will be heavily impacted by MT even more than today.
The workflow will be more and more automated by including AI. There are some workflows at Memsource that already are AI-powered. For instance, there is a decision-making mechanism for whether something needs to be post-edited or not, if the quality of the machine translation output is high enough. That is the AI feature machine translation quality estimation that can make this decision to route it to post-editing when it is low quality or to skip post-editing when it is high quality.
There is still a lot of work that needs to be done in the UI and the transaction tools, or with MT to fully utilize the power of MT and all the features, and this will no doubt happen. There are people that are thinking why have not the TMS providers done it, what are they waiting for? It is just a lot of things to do, and also sometimes the MT providers do not expose the necessary API calls that we would need to make this app. A few things that need to happen, they will happen, but there are also many other things that are the customer’s requests that are MT related or integrated with our content management systems. There are so many of them.
Florian: It is a good point because sometimes we do not interact with big clients every day and look at their needs. We tend to look at it in a somewhat simplified way, but when you have all these real-life enterprise challenges and feature requests and complex workflows, then you need to get those sorted out first and look at some of the other aspects.
Note: White & Case advised Memsource on the Phrase transaction.