In this week’s SlatorPod, Semantix Chief Business Officer Britta Aagaard joins us to talk about their recent major deal win — a EUR 34m translation contract for the EU GROW project. Britta discusses what’s involved in bidding for, winning, and setting up to go live on a contract of this size, as well as the role of Semantix’s technology partner ESTeam.
Britta also talks about her path to joining Semantix, which resulted from the 2017 acquisition of Textminded, the translation company that she headed as CEO before its sale to public sector interpreting specialist Semantix. Britta explains why she felt it was the right time to sell her company back then and why Semantix was a good buyer fit.
Four years later, Britta shares how learnings from the public sector carry over to servicing private sector customers, and how the acquisition of Textminded has seen Semantix expand its offering to become a bonafide full-service LSP.
First up, Florian and Esther discuss the news of the week, which saw dating website operator Match acquire a multilingual video chat platform built on Google Translation API for a cool USD 1.73bn, while Microsoft rolled out ‘Document Translation’ for LSPs and Enterprises.
The two unpack the latest financial results from Australia-listed companies Ai-Media — a captioning and subtitling provider that delivered upwards of seven million minutes in the six months to December 31, 2020 thanks to increased demand for live captioning in Zoom — and Appen, whose shares tumbled despite 12% growth in 2020.
Subscribe to SlatorPod on YouTube, Apple Podcasts, Spotify, Google Podcasts.
Stream Slator webinars, workshops, and conferences on the Slator Video-on-Demand channel.
Transcription
Florian: Britta is the Chief Business Officer at Semantix. Semantix is a Swedish headquartered LSP and you joined Semantix with the sale of your own company TextMinded in 2017. Tell us a bit more about your background with TextMinded and how did you get started in the language industry?
Britta: Coming back to where I was brought up in the border region between Denmark and Germany and as in every other, at least, European country, there are minorities on both sides of the border. In the northern part of Germany, there is a very active Danish minded minority and the culture is very much rooted in the whole educational system, so I have Danish ancestors because of historic reasons, but my family as such is German. My mother decided to send me to a Danish school, so I grew up bilingual, at home it was German and at school it was Danish.
After graduation, I moved to Copenhagen, Denmark to study and I was never supposed to be a translator or in the language field. My major was philosophy and I spent the first fifteen years of my life after school in academics. However, my field of specialization was the philosophy of language, which is still very important and also an important part of what I am doing, even if I have left academics. Without getting too nerdy it is language from a more epistemology point of view, language is actually the way we understand the world. If we do not have a concept, then we cannot understand whatever reality we think we see. There is this close correlation between language and knowledge, which still follows me a lot, so I would say even if I have left everything behind, it is still a big part of my vision today and especially when we talk about AI and data, I still see that essentially at its core is language. It is a representation of the world, so I would say I brought my academic background with me, even if I am on a completely different planet today.Â
Florian: What made you leave academia after such a long time and start your company? That is a big step.
Britta: It is and in many ways, the story was by accident. Being a research fellow was not so highly paid and because I was bilingual, I worked as a freelance translator at that point in time and worked for an agency here in Aarhus where I live and they had a huge project done in Trados at that point in time. They said to come into our office because that needs to be coordinated. I said, I can do that, so that is what I did and three weeks later, the founder and the CEO asked me to stay and I said, yes, I would love to stay, but I do not want to translate. It was definitely not my temper or my passion. However, what I like to do is use my organizational skills and I could see instantly there are a lot of things that we could do better or different.
Fast forward, I was asked to become a partner of that company and invest in the company, it was still quite a small four million euro company at that point. Then suddenly and very tragically, the founder and the CEO, who has been my mentor, suffered from brain cancer and within 12 months, 18 months, he died. He asked me during that time he wanted me to take over and again, I said, I can do that, so that is what I did. Already at that time, I could see that I wanted to scale that, I really wanted to do something more with it.
Then again, maybe accidentally at a gala conference in 2011, I met Robert Etches, who had a similar-sized company in Denmark as well. Basically, after half an hour of talking, we decided to merge, not because we needed it, but because we really wanted to. With all the legal aspects and what have you, nine months later, we merged our companies and created TextMinded. In 2012, we then did a few smaller acquisitions in Denmark and Sweden and were extremely successful, we doubled our revenue, we tripled our profits. Then, of course, suddenly we were interested in the bigger space.
Florian: Then you were approached by Semantix?
Britta: Yes, January 2017, so the whole acquisition was very fast. In a couple of months, we did it. We got a lot of interest. We were approached by many companies, also larger American players in 2016 already but we do not really need that. Then at some point, we had a board meeting where we said, no, we think we should do it on our own. Then one of our board members said there is always a number and what is your number and if that number is hit, then you also go into these negotiations and at some point, it just becomes one of the drivers as well. When you first start embarking on that idea of leaving something behind and embarking on something new, then there is no way back.
Esther: Tell us a little bit about Semantix, just a general introduction for those of our listeners who maybe are not so familiar. Tell us about your areas of focus or some current initiatives.
Britta: Many who know Semantix, still know it as a very public sector heavy interpreting company. In 2015, the current private equity owners of Semantix actually bought an interpreting agency and had huge success with supporting immigrant languages for the public sector. Whereas translation was a byproduct, that was not a focus area. Then very suddenly, also with the immigrant crisis, the whole market crashed basically, and volumes were dropping dramatically. At that time, I did not know that buying TextMinded was the beginning of the turnaround of Semantix. To be completely honest, it was a little bit of culture shock also for me, because Danish Swedish business culture is a bit different but I realized when I joined Semantix just a few months later that I was embarking on something completely different than I had imagined.
One month after the closing of the deal, the CEO left and they installed a new CEO, my CEO today, Patrik Attemark, and we were now tasked with the turnaround of Semantix. The turnaround in terms of moving away from the public sector and also moving towards translation or what we call language solutions today. What we did was we changed that organization in not being country based in both business areas, interpreting and everything else. I am heading language solutions today.
Florian: How do you collaborate, if at all, with interpreting, is there an overlap or is it mostly back-office functions, finance, stuff like that?
Britta: No, there is very close collaboration so all the group functions are supporting both business areas, so it is HR and IT and finance. It is a group level combined. We also have a very close collaboration on technology development and also own resources, especially in IT and technology.Â
Esther: This big deal win with the European Union was announced in December time, but can you tell us a bit about what is involved, services, volumes, languages, et cetera? A bit of background?
Britta: Absolutely. This is one of the largest EU contracts for translation services or language services and it is called the Single Market Transparency Directive, which they say is a tool for providing information to companies, to basically citizens, as well as how the trade barriers are prevented and so on. It is a gigantic system of technical regulations for any products, services, you might imagine, that need to be checked, whether it also applies to EU legislation. All of this in the spirit of Europe always needs to be available in all European languages. All data always needs to be democratically available in all languages. It is translation and revision of quite complex, both technical but also legal, sometimes financial matters. It is classified and shows the diversity of the content in 13 different domains of all kinds of goods and services. Again, not only the complexity of the content but also the process that these regulations undergo, that is a complete life cycle where it goes from being a draft regulation, then there are all kinds of amendments and comments from different member states until it reaches a final stage of a final regulation that is adopted then by the EU for the European Commission.
Florian: How long does it take to respond to one of these European RFPs and who is involved? Who do you work with for that?
Britta: I did this mainly together with our partner ESTeam. I would say the heavy lifting especially for this contract is on the vendor management and vendor recruitment part because we have needed to actually submit names and CVs of around 1800 translators with specific qualifications. Just the gathering of that information and providing that was a huge undertaking, so I would say from a workload perspective that is the biggest part and then of course writing the actual tender. There are just a few people involved but that is definitely very heavy both thinking and writing.
Esther: I know you participated in the tender as part of a consortium so how did you get together with the other two companies and how are you sharing responsibilities, et cetera?
Britta: We partnered with ESTeam, which is a quite small Swedish technology company, but they have been in NLP and AI for almost two decades. Again, more or less by accident, I met Jochen Hummel who is known as the original founder of Trados, and when he sold that to SDL, he eventually joined ESTeam and became the CEO. In 2016, when we were TextMinded, we were discussing a big contract for the European Intellectual Property Office, the EIPO. Jochen has been working with the EIPO as a provider of technology as well and they have been working with the EU on technology for a long time. Now, this tender comes out where it is about technology and services and ESTeam is not doing the services part at all so he said, we should combine that. Being a big believer in the combination of technology and services, that is where real scalability is possible.
As you also covered in 2016, 2017 we won that contract and it was a contract, not again on translation, but on revision and editing translations already and with the use of AI and technology. Ever since we have been collaborating and also a few months after winning that EIPO deal, this specific tender that we have won now came out and we said we could do this together. We applied and we did not win it at that point. Today I am quite happy because the magnitude of the operational side is huge. Not sure if we could have done it at TextMinded, but that was the first time that we worked very closely together on writing such a tender.Â
Esther: What is the technical part or the technology part of the contract and how is it all going to be integrated into the services?
Britta: We have suggested a solution where we are not moving away from the traditional translation part. However, we do preprocessing and postprocessing, where ESTeam is doing mainly the preprocessing part we use all the linguistic assets available. We use their NLP tools to provide an output that is, I would say, much better than you would normally see MT output done for the translators to enhance. That is the whole core of our solution, to deploy technology and AI, to provide an output that is not only in terms of fluency and quality of the language but also has annotations towards these domains. As I said, because they are so different, you need to be extremely careful of also assigning competencies. We use data not only for the actual translation process but we also use data for the matching of the competence so this is how we have designed a whole workflow based on the data that always travels with the assignments.
Florian: With Semantix it is a big organization so you can run this at a much bigger scale because people tend to, looking from the outside, think that it is all done with a great custom engine but it is not. It is managing all of these different text types, different areas and then allocating it in payment, etc. is a huge task.
Britta: Exactly. I would say that MT is just a small part of this, and we would also use eTranslation (a free automated translation tool for European SMEs and startups that can translate to and from 27 different languages), but that is a very small part of it. As you said, the real magic that happens is when you deploy that into an intelligent workflow where everything as much as possible is automated, but always with human intervention. Given that the nature of the content is very complex, it is technical, but it can also have a real-life impact on people’s lives because it is about legislation and there are quite some timelines to respect. If, for example, a draft regulation is not treated properly by the EU Commission, then an EU member state can basically adopt legislation that is not in line with the EU so it has real-life implications.Â
Esther: As I understand it, the contract is currently being serviced by another provider. I am sure you started prep work but in terms of it being live, how will it be handed over on day one?
Britta: We actually have started. Normally you would have two, three months of preparation and hand over time. However, due to all kinds of stuff, Covid-19, the people of the unit at the commission, were not in the office. The actual time from signing the contract until they started working with us was five weeks but we knew what to do. We then made project governance and approved program management of this huge implementation project, but not until a few weeks ago we got access to the linguistic assets, everything around processes and access to the database and so on.
This has basically happened the last two weeks and as you can imagine, many of our processes are heavily dependent on the data and still the quality, the amount, the structure of it we do not have a full grasp of yet. From a process workflow perspective, we know how it should work and now we are assessing the quality of the data. The commission is extremely helpful and very collaborative so basically, three, four times a week, we have work meetings with them.
Esther: This is a pretty major example of a public sector contract, but I know that Semantix also works with private sector clients, so what is your experience in terms of how those two differ in terms of requirements and solutions and expectations?
Britta: In the beginning, just three years ago, we had interpreting as 75 percent of our business. Today, language solutions are 80 percent of our business so we made that complete turnaround already. However, I would say then the move from public to private has been our main focus but still having that experience from decades of working with the public sector is very valuable. In general, as we also said, writing tenders for public clients, you need to keep formalities, that is core and there are requirements that are just yes and no. Whereas private customers, you would have a different approach. You might be able to influence them in discovery meetings, really understand the needs and so on. From that perspective, going into tender is very different, I would say, and in general, working then with the public versus private there are also in many ways still differences.
In general, I think it is fair to say that public clients are more cautious to adopt technology as well so it is still very much about the service as such. I think even in Scandinavia we at least perceive ourselves being very far with digital transformation but still, we can see the public sector is a little bit behind. On the other hand, once a project like this, the EU contract or other governmental bodies are moving, then obviously the volumes are huge. From a contractual and maintenance perspective, it is not heavy lifting because in some ways the expectations and requirements are much more clear than if you work with an enterprise, with numerous stakeholders involved. Once you have the operational workflow and best practices, it is maintaining a good delivery in quality but there is not so much business development or sales involved.Â
Florian: You are one of the somewhat rare examples of a founder or business owner who stayed on with the company that it sold to. Can you just tell us a bit more about the transition and the thinking process and why you stayed on? I want to know a bit more about the background to that decision.
Britta: It went very fast and from my perspective, there is always a number but also when you go into this you need to also be willing to let go because you do not know what hits you. However, for me going into that process, it was extremely important that I could still continue my vision for this industry and also for the company. Last but not least also for my people. I could see that I had reached a level where I was not able to grasp everything. We were 80 people or something at that point and even if you think you are great and doing great stuff at some time, that is limiting for scale. We could see all the wonderful ideas that we had and we were always driven by having maybe a bit too crazy ideas but still executing on them. We wanted that to live further and the scalability comes with growth and comes with power. For me, TextMinded also was still my baby, and I would say it was obviously a step to let go of that freedom of being in charge. When you actually see that you essentially can become limiting for your own growth and also your people’s growth, then I think it is time to do that. I still think that companies in the range of four or five million euros want to scale and grow M&A is probably the right way.
That was also one of my choices to sell to, let us say, a Nordic company, because we share some not only values but also business as such, the culture as such, we are very flat organized and so on. At the beginning of the year, we got a new CEO immediately, not from the industry and I am blessed with a CEO that trusted me as well. Today I can say I have the freedom of being in charge and at the same time very supported. That was for me, one of the best experiences in joining Semantix also on the C-level. Suddenly I have a great team of professionals around me who are smarter than I am and I think that it is learning you appreciate and you do not necessarily know that when you are your own CEO.Â
Esther: What general advice do you have for LSP owners that might be considering an exit or a sale? Are there any dos and don’ts?
Britta: I think when you are an owner of a smaller LSP or you are an entrepreneur, you need to make up your mind if there is a vision you want to carry on and how important is it that your people can also be in that new company. The integrity of TextMinded was very important to me, and I could see that not only live further in Semantix but develop further. In that sense, it is a very personal decision but at the same time, no matter how emotional that is for you or personal, you need to be able to let go and be willing to let go. Also at some point say, yes, they might be taking some decisions that I would not take. Even Semantix will also be sold to someone else at some point and something that I will not be in charge of, but still keeping the vision in mind.
Florian: It seems like the language industry has gotten through Covid-19 not unscathed, but quite well compared to many, many other industries. How is the business going right now? What is your outlook for the remainder of the year generally?
Britta: We at Semantix and also my business area of language solutions, actually had an extraordinary 2020. We had quite an ambitious budget going into 2020 and we almost reached those budgets but the best thing about it is that we actually significantly improved our profits. At the same time in this transformation that we were still in, also executing all the strategic initiatives. That is almost the biggest achievement for people who have been working hard day and night and that we at the same time have not lost focus on what we need to achieve in terms of the way that Semantix is not only consolidating but also integrating. We want to integrate completely into systems and culture and so on. We succeeded with that as well and now heading into 2021 we have this contract, which was a nice Christmas present, obviously, but already trading on our budgeted level. I think it goes for the majority of our peers in the market that we have a bright future.