Smith Yewell, Co-founder and CEO of Welocalize, joins SlatorPod to talk about how the team built the company from the first order, translating a single word 25 years ago, into what is now a global Super Agency.
Smith talks about co-founding Welocalize with wife Julia Yewell in 1997, and how they did not have a common language when they first met. He recalls how he navigated the many complex end-client industries and enterprise functions, especially with the maturation of technology.
The CEO shares the impact private equity investors have had on Welocalize and on the language industry — where roughly a third of the leading 100 LSPs are now PE-backed. He also reflects on the M&A space and shares the strategy behind expanding into digital marketing services as well as tech and patent translation.
Smith concludes with Welocalize’s plans for 2022: strengthening relationships and growth opportunities. He believes innovation around natural language processing (NLP) will change the way we communicate; from chatbots and digital assistance to automobiles and appliances.
First up, Florian and Anna discuss the language industry news of the week, with Google Translate adding 24 languages using zero-shot machine translation. At the Google I/O conference, the search giant also unveiled a smart-glasses prototype that can transcribe, translate, and display what the user is saying on the lenses in real time.
Having previously released a statement in April, saying it was in the “preliminary stages of considering a possible offer,” Baring confirmed they will not proceed with a takeover bid for RWS. Meanwhile, in Australia, the Supreme Court of Victoria has voided the sale of ezispeak to Technologie Fluenti, citing anti-phoenixing laws.
Florian: Tell us more about your professional background and how you got into the language industry coming from the military? You were an artillery officer in the US Army for four years so, how did that happen?
Smith: There were two big things that got me in the industry. The first was living in Germany and the second was meeting my wife. I was stationed in Germany in the military when the Gulf War started. I thought because the Cold War was still happening, they will leave us in Germany and the troops that go to the Gulf War will come from the states, but in a very unusual move, my unit got pulled from Germany and got attached to the unit that led the main attack in the Gulf War, so I was there for six months. Fortunately, it was successful and the very first weekend I got back I met my wife, Julia. We met in an Irish pub in Frankfurt, Germany, and famously, we had no common language. Most Americans never travel outside the country, so I was very fortunate to be able to do that and got to live in Germany and it was only when I was on the outside looking back that it brings home the point that the whole world does not speak English. I think that a lot of Americans assume that most of the world speaks English, but we know that is not the case and so the idea started to germinate from that. When I met Julia, it took off because we were experiencing it firsthand. We could not talk. With the growth of the internet, we figured people are going to need to be able to communicate in all sorts of different foreign languages on the world wide web, and off we went.
Florian: Did you move back to the US and start the company from there or was there an original seed sown in Germany?
Smith: We moved back to the US which was another interesting milestone because Julia did not want to leave Europe. She is very much a European and fortunately she did. We got married just one year later after we met and then we started to work on the idea of helping companies go global. We translate billions of words today, but it all began with our very first paying customer who famously only wanted to translate one word. You cannot make this up. That is exactly the way it happened. The first paying customer called and said, we are selling our product globally and we want to translate the name of the product. We were confused, but we said, okay, and the name of the product was Pathfinder. It felt like the skies had parted and here it was, we are trying to find our way. The first customer wanted to translate the word Pathfinder and we did and our path from then on went into this direction of localization and now that has morphed, so it all began with one word.
Florian: Let us fast forward. From Pathfinder to billions of words. You have a global marketing arm now that would probably be happy to do the Pathfinder project for a good consulting fee but take us forward 25 years, what are the key services now? What is the geographic footprint and leadership team, so people understand the scale and size of Welocalize today?
Smith: We have been super fortunate to keep on growing from the very beginning, from two of us to well over 2000 employees today all over the world in Asia, Europe, and America. Last year we earned about 300 million in revenue. We are on our third private equity investor which is also pretty rare but that has worked out well for us. Our leadership team is all over the world. We were distributed even before the pandemic, so we are used to that and we are incredibly proud of the team and where we are today. It has come so far from that first word.
Florian: You have seen a lot of waves of technology of major clients from translating, localizing Windows to now everything on SaaS to all these regulators becoming much more important. What have been some of the key milestones and challenges for you over the past quarter-century that you felt were transformative for you in the business?
Smith: There is definitely a wave that coincided with the dot-com wave. I remember eTranslate back in the day and eTranslate raised a bunch of money and the whole idea was that you could do it all automatically on the internet, get anything translated and it was all just going to happen magically. That did not quite work out. Translation management systems also took off. Compared to now where those technologies have matured and there are lots of great companies in that space. Curiously, I do not see any of them in the ranking. I do not know if these technology companies do not want to report their revenue or what is going on, but it is interesting to see that 25 years later, I have never seen a tech-only company in any top 10 ranking, so that is curious.
Florian: You are working in all these super complex industries: life sciences, MedTech, software and tech, and global marketing. This is very complex to manage and for you to wrap your head around all of these different client segments, how do you do that? How do you understand all these different pain points from the clients? Tell me a bit more about that.
Smith: The first phase as I was describing was very experimental. Right out of the gate, we brought our first investor in the summer of 2000 and the whole premise was we were going to ramp up the company and build out a global sales team. We went out and spent a bunch of money that we raised from our first private equity investor and then everything changed overnight. At the beginning of 2001, the internet bubble burst, our biggest client went bankrupt and we had to cut two-thirds of the company, so right out of the gate unfortunately we had our first big failure and we had to recover from that. From there, we fortunately did. We made some tough decisions early. Those were the toughest things I have ever had to do because at the beginning of a startup you hire a lot of your friends and we had to make some of those tough choices. The second phase was seeing that experimentation mature around automation because my first big client after the Pathfinder client was Cisco. The only reason we won it was because we were using Trados and we met a little company in Germany that had created all these innovative macros to process files. This was back when you would get the source visual basic files and get a whole lot of the source code back so that maturation phase was around technologies with translation memory leading the way. Those were the early days of the adoption of that technology. Not many translators were even using it, so that helped us differentiate at the beginning with Cisco and they are still one of our top clients 25 years later. Other technologies started to come into play. The challenge was there was no intra-operatively, so we were one of the first companies, with a number of others, that formed this initiative called Interoperability Now. We worked closely to try to get exchange between different systems because Welocalize took the view which was novel at the time, we were not going to create a walled garden around technology. We wanted to open it up. We wanted customers to use any technology they wanted and we would build around multiple products, so that is what that second phase was all about. We saw the technologies mature in their capability and then learned how to collaborate, create connectors, and use standards to create connectors. We led a number of those initiatives.
Florian: You have quite a strong internal piece of ERP that you are building around and you are adding these additional systems to it. Can you tell us a bit more about that?
Smith: We had a fun debate many years ago. It was at a TAUS conference and myself and the Chief Technology Officer of SDL at the time got up on stage and debated the walled garden approach because SDL at the time was very much into proprietary products and even getting the APIs was difficult. We took a different view and have stuck to that view ever since. We very much want to create technologies, from prospecting with a customer, in our case that is Salesforce, all the way through to invoicing and collecting from a customer and making it fully digitized but the only way to do that is through an open environment where you are connecting best of breed products. That whole initiative in the beginning which we helped lead called Interoperability Now was about connectors. Our system Pantheon has over 200 connectors to client systems, to our competitor systems, to third-party systems such as Microsoft Products or Workday and that is how that ecosystem has evolved. That has been the approach we took.
Florian: When you think about CAT tools and connecting them, is it not hard to maintain a dozen different technologies from a workflow perspective, a billing perspective, a quoting perspective, and then maintaining them among a large distributed linguist space that you have?
Smith: It is super hard. You are correct. A whole lot of work and investment has to go into those connectors. Is there a way around it? I used to think we could probably get there if we could get the industry as a whole to adopt certain specific standards to make that much easier. That was harder than we thought. Getting a whole lot of competitive companies to agree just was not entirely possible, so it did not happen. What does that leave us today? Only about half of the tasks we do today at Welocalize are translation. It is interesting how that has evolved. Everyone wants their own work surface. Everyone is comfortable with where they are most productive, so we have got to try to enable that experience so it is easy for them to do the work. That is the approach we have taken.
Florian: You mentioned that only half of the tasks now are translation tasks, so what is the other half? Is it around your AI training data services, for example? Tell us a bit more about that and some of the other lines of business that you are in?
Smith: The interesting thing is we started in localization. That is about a little bit more than a third of our business, so that means 10 years ago, we were not even two-thirds of our revenue. That is how things have evolved. It is incredible and as I said, half our tasks are translation-oriented today, so what are they? We are trying to create the ability for a customer on their global journey to have a guide, a trusted partner, almost a travel agent, for lack of a better word. You are going to a new country, you do not speak the language, you do not know where to go, so for us that begins with legal, protecting your product. From there, marketing your product, so we bought two digital marketing agencies over the years. We have got core localization across all the different content types to support our customer’s customers and then we rounded out with the data and that is the most exciting thing for our industry today. Effective communication is at the heart of all human interaction, whether it is personal, whether it is business. If that is not understood, business relationships break up, so where does our industry sit? At the core of that is the data because it is all digitized now. In the beginning, we even took faxes. It was all analog. Now it is all digital which means with all of that natural language processing data, our industry is sitting on a treasure trove of data. How we can use that to make communication better around the world is the ability to change the world, so it is an exciting place for our industry to be right now.
Florian: We are seeing all these large language models that are open source by companies like Facebook, Meta, Google, et cetera, and then there is this ecosystem of startups that are building all kinds of additional solutions on top of these powerful models. How do you see this fitting into the more traditional language industry? Do you think we understand where this is going? Do you think there is another wave coming that we need to catch? Generally, where do you see this powerful language AI going, and how it is affecting the language industry, but also Welocalize?
Smith: We are in the wave already, I would argue, and we are trying to ride that wave but things are changing so quickly. The wave could crash and you could be caught up in it. We have had those experiences because we try to stay on the edge of the wave, so NLP is a big part of that. You were talking about if we could change an engine industry acronym, I will throw out another idea, let us change LSP, language service provider, and call it NLP, natural language provider because that is what we have become. In that context, the ability to use linguistic data. Let us take, for example, the convergence between product localization and product market. When I first got into the industry, all our customers were product people. We never got a meeting with marketing because those are separate areas in the companies. They were not even communicating, but now those have converged, so when we see product localization and product marketing converging, then you can take this language data, which our industry is fortunate to be sitting on and you can use it to help identify what is more effective in SEO, as an example. Is it short-form languages? Is it long-form? Is it formal? Is it informal? What reading level is it? You can use your performance data, your Google Analytics, or whatever products you might be using, and you can link it to experiments through algorithms with the linguistic data itself. Strategically what we always wished for at the very beginning 25 years ago is can we be a deep partner for a customer, not just a project provider? Now we are there.
Florian: You have been in the industry for 25 years, you have seen these hype cycles, how do you stay balanced between making sure you are not missing out, but also not just going into the completely wrong direction because there is just so much happening and people have to prioritize.
Smith: I never believed the hype that MT was going to change the world and put us out of business. We are on our third private equity investor and each time we met a potential new investor that was the first question. Is MT going to put you out of business? It is just another NLP technology and it is very effective and we have got significant usage of it ourselves but is it going to radically change the industry? There are other things that will change it even more profoundly like NLP across the board rather than just that single NLP technology. NLP is using the data to not just train the machine translation engine, but all kinds of predictive capabilities, so you can understand, for example, where is the risk in quality? Quality historically has been very wasteful because it is just random sampling. The algorithms should be able to tell us where exactly the risk is and you can use the data and that can help point to where the risk might be in quality? Is that hype? Definitely not because that is mission-critical to a company’s business.
Florian: Why do you think the role of private equity is so strong in the language industry? One-third or half of the top 100 providers are owned by private equity. Why is that? What do they like about this industry and why are they so active?
Smith: You are right. I saw on your report that about a third of the top 100 ranked by revenue are private equity-backed in our industry now. It is fun for me to see how far that has come. We were one of the first private equity-backed companies in this industry back in 2000. We had a board meeting last week and I asked, the other industries where you invest, do they have so much private equity like this one? They said, no. Why is it? I think it has got a few of the basics that all private equity companies like to see. First of all, is it a growing industry with a bright future? Yes, and it just keeps on growing, so great. We have checked that block. What happens when there is a problem? In other words, how resilient is the industry to downturns? In the 2008 financial crisis when everything collapsed around the world, how did we do? We were flat, so we did not decline. In Covid, everything collapsed around the world again. How did we do? We were flat again, so through both of the biggest challenges in the last 20 years, our company and the industry as a whole, weathered the storm. You are growing. Great. If things go bad you are in good shape. The next area is technology. Can we automate? Can we invest to help improve not just growth but profitability as well? Absolutely, and then finally, is it an industry where we can consolidate? Private equity companies are very attracted to buy and build and Welocalize has acquired 19 companies over the years in that context. That opportunity is still as big as it ever was. There has been a lot of consolidation, but there is still an enormous amount of fragmentation and you will see that consolidation continue and we want to be one of those consolidation leaders. We could not be more fortunate to be in such a great industry and as an industry, we love what we are doing. We love the people we are working with, so it is just a fantastic spot to be in.
Florian: What is the playbook? How do you scout for deals? When do you pounce? When do you buy something? What do you do after? Is there an integration playbook? What are the typical challenges of having gone through this 19 times?
Smith: If I could sum up the challenge in one word it is trust. Building trust is the challenge. If you can do that, you can get through the ups and the downs of which there are many, because anytime you are integrating it is the same as being in a personal relationship. You have got to compromise and those compromises are tough. Especially, if you have been doing it the same way for so many years and all of a sudden that way has to change and have the humility to be willing to do that which comes because you trust each other, you trust in the relationship. That is what we are trying to do. That is the brand I want Welocalize to stand for. The most trusted brand in the industry whether it be through with clients, with vendors, with staff, or with acquisition targets. What is the playbook? We come in and we try to build that and what I tell everybody when I first meet them after we acquire a company, is you do not know me. I am not expecting you to trust me today or tomorrow, but if you are patient, if you are willing to give us a chance, then I pledge we will earn that. We will do our best and we are going to earn it and more than half of our leadership team worldwide all came through acquisitions, including half our C team, so we proved it is possible.
Florian: Do you typically rebrand or do you keep a couple of brands on their own?
Smith: We have kept a couple on their own. We have got a brand for legal, Park IP, and we have a brand for digital, Adapt Worldwide. We do not want to confuse customers and when we bought Park IP it was a very strong brand. Park and RWS were two brand leaders in the IP space at the time and we did not want to confuse anybody, so we kept it, and off we went.
Florian: This is one of the more challenging parts of the language industry to understand the whole IP space. How was that for you buying that company, acquiring these capabilities, and then using the resources of a larger company to get ahead in that very complicated space?
Smith: I did not understand it at all. RWS was public and you could go learn whatever you wanted. I had heard about RWS, I had heard about patent translation and it was just interesting to see this hundred million company doing well as a public company. I started to get curious about that business because they were the only game in town in that space at the time. Our customers were asking about it and that has guided us from the very beginning. We started as a traditional localization company and evolved into much more than that because customers asked us about it, so we investigated it, looked into it, and found a great company. We were, fortunately, able to acquire Park. I think RWS took notice of that one because they turned around not long after got into localization. We jumped into Park and they jumped into Moravia, so that was an interesting tit for tat, and kept on evolving since then.
Florian: What are some of the initiatives and projects you have for 2022 and next year? What are you working on and how does the current climate influence decision-making or not?
Smith: What is going on with the war in Ukraine and inflation around the world are changing everything. I have never had to encounter or work with inflation before. The last big inflation cycle was in the seventies, so before we ever got into the business, this was new territory. How do we navigate through it, especially, with the great resignation going on and incredible pressure on wages? We have got to continue to create these relationships with people, so how do we do that? We have two people goals and that applies to all people: staff, vendors, and clients. Those two goals are first, demonstrably prove we care and the second is to answer the question, where can I go from here? People can work anywhere and what reason can we give them to want to work here. In terms of where they can go, it is the best place to be. We have got three OKRs in the company, objectives, and key results. They are: be the greatest growth opportunity. Again, for everybody involved, investor staff, customers, and how customers grow. Be the easiest to work with. You mentioned the complexity of the different technologies, we are not there yet. Many of the things we are doing today are still not easy, so we have still got some ground to cover there and then finally be the greatest company in the industry, not necessarily by revenue, but by reputation. That is what we are shooting for and that is how we navigate through these challenges. People want a safe harbor when times are tough. I have been there. I was in the Gulf War and I can tell you when the bullets are flying and you are in a box hole, who do you want covering your back and who are you going to trust to do that? That stuck with me ever since and that is how we get through this challenge. You do not have it with everybody, but we work hard to try to get it with as many as possible.
Florian: Let us assume we get through that challenging period. Where do you see the language industry in two to three, or pick your timeframe, five years? How have we evolved? Will we be very similar to what we do now, or are we at the cusp of something big and new happening?
Smith: We are definitely riding that wave. Big, new, exciting things are happening. I do not think that is hype. In terms of the ability to automate and do exciting innovations, we are not the only ones. There is lots of interesting innovation happening. Especially, around NLP and that is going to change the way we communicate with chatbots, with digital assistance, with appliances, with our automobiles, with all of these things that we use in our daily lives. There are only two choices of an interface. You are either speaking to something or you are typing something into something. That is all you have got. In both of those languages at the heart and NLP technology, which is capturing whatever you say. Whatever you speak into the device is going to change the experience and companies appreciate that and the importance of culture and respecting culture through localization at different quality levels of language, depending on the need, which could go from raw MT to being exactly right. The average translation task for us is less than 200 words and for our biggest customer, there are 60,000 of those tasks that go through the pipeline every month, so this is the kind of change that is happening and our industry is at the center of it, so I find that to be very exciting.