9 months ago
November 20, 2020
SlatorPod Episode #48: Summa Linguae CEO Kris Zdanowski
Joining the pod is Kris Zdanowski, Founder and CEO of Summa Linguae Technologies. Kris talks about his entrepreneurial journey, from paying for receivables collections with shares, doing a merger with an LSP in India, to rolling out an M&A strategy and launching into the data annotation industry.
Florian shares a few choice insights from Slator’s Inside India’s High-Growth Language Industry webinar, which this week attracted more than 130 participants and featured four expert panelists.
The two discuss startup Otherside AI, which raised USD 2.6m in funding just six months after launch. Florian unpacks the tech, which transforms bullet points into complete sentences (including in other languages), and its potential benefit for translators. They also venture down the translation test rabbit hole — a divisive issue, as highlighted by Slator’s recent reader poll.
Esther talks about the language industry’s latest association: Entertainment Globalization Association (EGA), which launched this week with 60 members from within the media localization niche, and is led by ex-Netflix exec Chris Fetner.
SlatorPod Episode #48 Transcript
Florian: Tell us what Summa Linguae does? How long have you been with the business and tell us a bit more about the background?
Kris: I started the business in 2011, but the story goes back a bit. I come from a bilingual family, a Polish-Canadian family. I moved to Kraków for my studies and just had to support myself. I had two skills, I spoke English and I’m good at being friends with people, I have interpersonal skills. I was a little too lazy for translations and so I started consecutive interpreting. Then I was, modestly enough, quite good at interpreting, so I just got more and more jobs. I was a little too lazy to attend to all of them, so I started outsourcing to a group of friends.
Then essentially this one larger LSP hired me as a freelancer. I’ve never worked for anyone per se, as an employee, and the contract was with ArcelorMittal, a steel plant here in Kraków. In 2008, this whole steel industry crisis hit the world and I got fired from that contract or that LSP that hired me got fired from the contract. I lost my main source of income, so I decided to start a one-man shop, as a freelancer, LSP freelancer and just grew a little bigger, got more jobs, and more clients.
Then I met this guy literally in an elevator and just started talking. I think he was back from golfing, he was carrying a golf bag, which is not something you see very often here in Poland. It’s not a very popular sport and I just asked him what do you do? And, I told him I have this problem with collections, I probably had the size of a business of 50,000 euros annually or something. I told him I had this issue with collections and that I didn’t know how to solve it. Then he says, okay, I’ll help you solve them. He runs some sort of a business as well, and then I ended up giving 50% of the company to the guy.
Florian: That’s good. That’s generous of you. Did he help collect at least?
Kris: I don’t think so actually, and we soon got into some issues and some problems, but then I met this other guy, and very soon I realized I have to get rid of this guy whom I’ve given 50% of the business to.This other guy also had an LSP, and then in 2011 we knew this guy in Poland as well who runs a smaller LSP. We decided to join forces, so I talked to my partner to join forces with the other guy and we joined these two LSPs. During that process, we bought the initial guy out, so I got rid of him and my business also doubled in size. This is how in 2011 Summa Linguae was born, out of a forced merger between two small Polish LSPs.
Florian: How was the return for him? Was he happy? Was he a happy person to exit?
Kris: He didn’t pay anything for his shares. He got them for a false promise of helping me to collect my debts. It was just a crazy story. The guy, he had some issues, he went into financial problems, he had a family to support. I think the bank had forced him to sell his house, he couldn’t pay the mortgage. He just tried to just squeeze us for whatever we made, a small LSP, we’re talking a 100,000 euros maybe. We made some money and I didn’t pay myself any salary whatsoever. Whatever money came in, he just wanted to take that money for some reason. He was always like “I’ve given you this advice or that advice or whatever”. Started threatening us. We just had to get rid of him, so we did.
Florian: That’s a long road from what you just described to being a publicly listed company so give us the version of what happened in the meantime.
Kris: In 2011/2012, as Summa Linguae merged with these two smaller LSPs, we thought it’s maybe a good idea to do some sort of a rollout strategy in Poland. Consolidate the Polish market, the Polish LSP market. This was the initial idea, we didn’t have money, so we had to go and secure some funding, mainly debt funding, some early equity angel funding in the beginning as well. To secure some funding to pay in cash, but mostly we tried to pay in stock and shares.
Being publicly listed helped us pay in stock so this was why we wanted to go public. This was the main reason and we went public in 2015. By then Summa was a million and a half in revenue. We bought a few smaller LSPs in Poland, and really then we were an unexciting small, typical, usual, LSP translation agency. Nothing exciting about it, just one of the tens of thousands of such companies, working 50/50 for public clients in Poland and some end clients. We never worked for other LSPs though, but it was nothing exciting.
In 2015 we went public and then we realized, whomever we could buy out in Poland, we have either merged or bought out. Then it’s also fragmented, there’s no local leaders as such. There’s ourselves and one more company. We’re Polish but we don’t really work in Poland per se. Out of the Polish LSPs, probably the largest is around 3 million euros. We couldn’t really grow and it’s not an interesting or big market per se, for acquiring revenue. One partner, the one I merged with in 2015/2016, he also exited from his other business because he’s never run that LSP he founded. He had people run it for him, he was doing something else, and he exited that business. We sat down and said we’ve done whatever we could have done in Poland, how about we just start travelling to these industry conferences?
We would have gone to the Slator Conference, but in 2015 the conference didn’t exist, so we chose others. I attended the first SlatorCon in London, I remember that very well. We had to go to other conferences. We just attended the galas and ATCs of the world and we formed an idea that this business model is not really sustainable, running a LSP like any other is probably not the best idea. We wanted to specialize in a certain vertical or a few verticals. We wanted to get a little more techy, it was zero technology at Summa, just the usual, a couple of CAT tools and that’s it by then.
Then at one of these conferences, I met Madhuri. She was the founder and owner of Mayflower, an Indian company. We sat down and we’re like, so how about we merge? She wanted to open herself up to the Western Europe a bit and the US, and we wanted some technology and some industry knowledge that we never had in Poland, and some production capabilities. She brought that to the picture, she was fairly well-structured for the size and geography she operated in. In 2016 we merged and we rolled the new strategy out. We adjusted our values and decided to build something serious, something real.
I always say Summa started in 2011, but we really started in 2016. In 2016, we were a million and a half in revenue. The same revenue since we went public. We’ve plateaued, we didn’t really grow that much. Global strategy was vertical focus, more technology again, more M&As and we thought it’s probably a good time to think about acquiring a serious financial partner. We laid a strategy out, at some point there was a three-year plan, to acquire a private equity fund. Grow the business from $1, $2 million to $10 million or so. We wanted it to have a global footprint, we wanted to work within clients. We wanted to have our own TMS, we changed that idea a little later, but we wanted our own technology. To focus on capabilities or production and lower cost geographies revenues in higher-margin geographies, like the Nordics or the US and we did succeed with that plan fairly well. We acquired a company in the Nordics, we finally found a private equity partner in 2011. We grew the business to over 20 million dollars in revenue in 2019. I think it was a successful plan that we’ve rolled out and now we have a new one for the next three years.
Florian: You gave us the picture there with the revenue, so it’s quite a sizable business now, done via M&A. I want to get your take on it. There’s so much M&A happening in this industry and how do you see it, does it add value? Do most fail? What are your thoughts on the big M&As, such as, RWS, Amplexor, Acolad? As somebody from the M&A scene who’s done a fair amount of deals in this industry, what’s your take on what’s going on at the moment?
Kris: In the first place, I think it’s a highly transactional industry for a number of reasons. There are many reasons why the industry is so transactional. First of all, clients are fairly sticky, the sales cycle is quite long. I would say it’s a big and growing industry, quite economy dependent. Obviously, this year is very difficult. We’re actually dependent on the economy of our clients, but it’s fairly spread out across all economies or industries. It’s a must-have service rather than a nice to have service. The volumes are fairly unpredictable, it’s a little difficult to estimate. You have a SAS business where you just know how many licenses you have. It’s a recurring business, fairly difficult to estimate, but still a recurring and must-have service.
The industry is very fragmented, so there is a lot of room for M&A. Naturally, the PE partners like the price or valuation arbitrage, where there’s a lot of smaller companies, 1, 2, 3 million to 10 million. For a lot of them, you can probably buy them a little cheaper than you can value yourselves at the exit when you put them all together. I think it’s a honeypot for all these PE firms now, so they’re jumping on it. M&A will continue. I think we clearly have the very large players, the super agencies, and they’ll need to change hands. The private equity-owned will need to be sold. H.I.G will probably think about selling Lionbridge at some point, whatever is left of it. We localizers on the LSP side will also need to be sold at some point. The ‘I buy everything’ Acolad will keep buying everything.
Florian: We’ll ask them at SlatorCon
Kris: Then you have the midsize companies like ourselves, and there’s only maybe a hundred of them in the industry. Again, very interesting and transactional, but there’s only a handful of them, few compared to the size of the industry. The clear case for a private equity investment is actually the M&A. Unless you’re a tech business, or memsource or CAT tool or XTM, or one of these, you’re a product company. In my opinion, if you’re a services company, it’s not really that interesting of an industry for private equities. I think it’s interesting because you can execute M&A, so the growth is backed by M&A. It becomes tricky at the same point, because there’s only a handful of companies who’ve actually executed a rollout strategy successfully. There’s quite a few of them who failed. There’s some who’ve done one or two, and who knows if it’s a success or not? How do you measure the success of an M&A? It’s a little difficult as well. These PEs either focus on these huge super agencies or they look at the 50 hundred companies that they can invest in, who’ve executed a successful MNA strategy, ourselves modestly being one of them. But to onboard capital, to fundraise for organic growth, I don’t think that’s very interesting for a private equity firm to invest in, in an LSP, in a services company.
You’re asking me about the future, I think valuations at the end of the day are a little crazy. Obviously these product companies, tech product companies are a different story, but talking about LSPs, forget if you’re a small LSP, let’s say sub $10 million, and you’re asking anything more than six times your profits, I think that’s crazy. There are some transactions like that, but if it’s just the services company with nothing extra, I think that’s insane, because any other services industry will never pay any such valuation. Benchmarking against other industries, it just doesn’t make sense. Even if you’re a larger, in the midsize LSP, you’re asking valuations upwards of 8, 9, 10. I think it’s high and we’ll see these valuations drop. We’ll see them drop because there’s more benchmarks, there’s more transactions, and if you’re really, just a services company, you’re going to have a hard time selling for eight or nine times your profits. So I think that’s going to change.
Florian: You’re doing some of the negotiation here. Now you’re in the market for buying, I’d be setting expectations now.
Kris: Yeah, at some point the owner will want to exit, so it works. Honestly, if you think about it, you’re a services company, why would anyone pay? You’re growing at the industry’s pace, 5%, 4%, 6%. Why would anyone buy your business for times 10 your profits? In my world, it doesn’t make sense.
Florian: So let me just ask you about the Lionbridge transaction then, because that’s sold for 20 times and frankly, I have no particular insight into that business, but it is a service business. It’s called AI, but at the end of the day, it’s a lot of human students, a lot of human micro tests. So why would that go for 20 times?
Kris: It’s different, again, it’s not an LSP industry. I think you guys don’t rank Appen in your LSP largest, and I think you’re very right. Although they do some translation services, very little, you’re actually getting it right. It’s a very different industry altogether.
Florian: Yeah, it’s apples and oranges.
Kris: I think it’s because this industry is just booming. Clearly Lionbridge at $200 million in revenue was a market leader alongside Appen. It’s a relatively new industry, so if you have a chance to grow, if you have a chance to triple that business in three years, I think it’s worth paying because your effective price is obviously much less. But if you’re growing at 5% rate, nobody, nobody will ever pay you anything close to that valuation. I think the growth of that industry is the reason.
Esther: Let’s talk a bit about what happens after the deal was done and about post-merger tech integration, how do you go about that? How do you deal with that, when it gets so complex after multiple M&As and mergers?
Kris: Yeah, good question, Esther, because I think it’s fairly easy to buy. A lot of companies claim success when they buy a company. I think it’s just 10% of a success. It’s really difficult to integrate them and in our case, it starts with a C-level position. We have a Chief Strategy Officer, but she’s actually in charge of integration, I don’t do that, nor does my CFO. We’ve made these mistakes in the past, so we have a dedicated C-level role to just oversee integration. That’s very important, and I think integration is a clearly well-defined process. Production, technology, sales, brand, values, HR, everything is structured into a process. It has a risk lock against every item, clear deadlines and clear KPIs.
What does it actually mean that a M&A has failed or has succeeded? It’s a difficult question to answer, unless you set yourselves with these expectations, these KPIs, you will not know. We have them set before any acquisition. There about the upsell and retention of customers, but also about the retention of employees that we want to retain. It’s about running surveys among employees that we have acquired through an M&A, and comparing the results to the employees that we have hired initially prior to the M&A. Looking at how these two groups of employees answer questions like ‘do you know what a three-year strategy of the company is?’ And you would be surprised.
These are success metrics that you need to define and obviously technology and the TMS or ERP that governs the company is a backbone. It’s very important and crucial. The problem is it’s integrated with so many tools, it’s usually very difficult and if you ask a lot of companies, a lot of LSPs who’ve merged or bought companies, it’s probably the largest pinpoint, and so it has been our case as well. We have bought a company in the Nordics that had their own TMS. We wanted to have our own TMS and we’ve learned the lesson and abandoned the idea of developing our own TMS. Over 50% of our business has nothing to do with localization, 55% of our business is in data annotation.
Florian: How is that to integrate? There’s probably very little in-house linguistic capabilities?
Kris: My principle is that we’re not integrating this. It’s a very separate business. Actually, there’s some cross-sell potential, but I think we sacrifice that potential for a greater good. I think that greater good is not mixing the two, they’re completely different businesses. Talking about the LSP side of business, our idea is not to have our own TMS. Our idea is we’re a services company, we don’t want to develop technology. We have some connectors, we have the API, and the APIs that connect to different content management systems. It’s a structure that is a puzzle of a data warehouse and a lot of components plugged into the data warehouse with Plunet being the production environment. As little as we can develop ourselves, this is what we want to do.
Florian: That’s the Canadian company you acquired right about nine months ago?
Kris: Yeah, in December. We experimented with data, like some LSPs, because it’s an industry where LSP has some sort of segway into the industry because the clients that I’ve worked for have asked them to perform some of these tasks. We started experimenting with data in 2017. We’ve realized it’s a very interesting and lucrative business, valuation wise. Growth potential is humongous but we also realized that it’s not a sustainable business to run as part of an LSP. It has to be a very separate business because it’s a combination of crowd, human and technology. It’s still a services business, it has some correlation and similarities to the language and LSP industry, but it’s still very different.
Some LSPs, you had one or two on your podcast who play with data as well, they tend to put them together and mix them together and hire linguists to perform these tasks. I don’t think that’s the way to go, do in our case, we separate that business completely, and it has a separate management, a separate strategy, a separate technology. All the technology that we use, around 90% of the technology used is in-house developed. The annotation platform, the mobile app that collects, and processes image and voice, and different kinds of data. All of that technology is our own and I think that’s the right strategy.
Florian: Data annotation market is not as distributed as the language industry? You don’t have this cascade of players, you have these giants and then niche players, or how large would you say it is? There’s not thousands of data annotation service providers, correct?
Kris: You have some startups mainly in the States who have developed technology, and you have some companies in Asia who have not developed technology, but they have a lot of crowds or they have this cheap labor. Then you have these giants like Appen and Lionbridge, and then a handful of midsize companies, three, five, six companies, ourselves included, who actually perform these services in a professional way, with technology and the crowd.
Then you have a lot of LSPs who’ve been asked by their customers to perform some data annotation, data collection, tagging services, mainly on the client’s platforms. If you look at the LSPs claiming to be data annotators on their websites, I think they staff people to perform data annotation tasks on the client’s platform. Very few actually own a lot of IP and actually run that business as a separate business unit. In our case, it’s 55% of our revenue, so we give it a lot of attention and our strategy is around growing that business. It’s not that saturated, it has more potential. The large accounts are much easier to acquire, there is some consolidation going on, the industry is trying to professionalize a bit. The AI annotation industry is trying to professionalize a bit, but you talk to very different teams of people. You talk to the LSPs, the LSPs talk to the language departments. The data annotators talk to a very different set of people, even at least the same customers. It’s just a very different business, and to ask translators or linguists to perform data annotation tasks, it’s not the right way to go. I think that’s the mistake that LSPs make.
Esther: I’d like to talk a bit about India. It’s very fresh in our minds because we ran an India webinar yesterday. It’s a fast-growing market, still very much emerging, and you’re one of the few LSP entrepreneurs who’s actually acquired an Indian LSP and you do business there. So can you tell us a bit more about that deal in particular? And how is it to operate in India?
Kris: India for us was our first international deal, very important and Summa, as an international global company, was born out of this deal. I think my business partner Madhuri was the key here. I would have not done this deal if she was not on board and aligned with the strategy. We bought some minority shareholders out, but it was actually a share swap. She’s a very significant shareholder at Summa, so it’s actually a merger rather than an acquisition. A path paved full of challenges, you can probably imagine a small company buying or actually swapping shares with a small Indian company. If I had to repeat this I don’t think I would have, I’m happy about the result, but I don’t think I would have embarked on that road again. All the challenges you can probably associate with India, we faced them. It’s so great to have Madhuri on board because, without her, we would have failed before we even boarded that Lufthansa flight to Bangalore.
Florian: You’re generally a bit of a risk-taker? You’ve done the Exfluency project on the blockchain and then India. Now the data annotation, that wasn’t a sure bet either, but it seems that it’s really helping you in 2020. So where does that come from? Is that your pilot gene?
Kris: I’m an entrepreneur. It’s not a noun, it’s an adjective in my case. Yes, I am a pilot. Outside of my Summa life, I have done other VC-like investments and many such things I’ve failed a hundred times. I’ve bankrupted companies. I’ve invested in companies that have bankrupted. I’ve done many things. And I think that experience is born out of these mistakes. Being a pilot is just a hobby, but for business, I’m really an entrepreneur. All these things come from risks. There’s a huge difference between owning a business and being an entrepreneur. If I were just to own Summa Linguae it would not be so exciting for me, it would be a nice profitable LSP, but it was riskier, it was more difficult, it caused me sleepless nights and tears. Summa is what it is because I was brave enough. I took all that risk in a sense. Pivoted to India, out of all places, fired the two partners that I have joined forces with, before many very difficult decisions were taken. Pivoted into the data annotation business. Invested in things like Exfluency.
Exfluency is a product, blockchain is a part of it, but it’s essentially a product that is trying to redefine the workflow of how translations are made and substitute the linguists by subject matter experts in a monolingual editor. In one sentence, it’s trying to orchestrate a workflow where you take the MT output and you have a subject matter expert look at the monolingual MT output and enhance it. They claim this is a better way, so we’ve invested in Exfluency, I was part of the founding team. Summa Linguae has a life on its own, it has a separate management, it does not really overlap that much. We use Exfluency in our workflows. Exfluency has some first initial customers, some POC, some revenue of half a million dollars this year. So it’s a risk, and I’m a risk-taking guy.
Florian: I hope not in the pilot role though, don’t go towards the thunderstorm. Can I talk a bit more about Poland as well? I remember I was there for the first time, about four years ago, and I was so impressed by going to Warsaw, the skyscrapers and all the building activity. For people who are active in that region it’s not news that it’s a very robust economy. But when you look at the figures, this economy has been growing uninterrupted ever since the early nineties. What’s going on in the Polish economy? How important is it for you? Give us a bit more insight there.
Kris: Poland is only about 5% of our legacy revenue, so it’s not really that important as an economy. I don’t think it’s also very interesting for LSPs, but as an economy it is booming. It has a lot of IT focus so there’s a lot of IT engineers here. A lot of companies have opened their BPOs, their outsourcing activities in Poland. A lot of all these large corporations have opened offices in Poland. Kraków being the second or the third biggest outsourcing center in Europe.
Florian: When it’s booming and salaries are going up, isn’t that challenging the business model of outsourcing? I remember it from my days in China, when I started salaries were a third of when I left, so the whole outsourcing play at some point falls apart.
Kris: Fair point, and I think it is the case because the salaries are obviously much higher and these outsourcing centers have suffered, but they’re still fairly low compared to what we would have to pay in the States or in Western Europe, so there’s still a clear financial gain. The balance between India and China, and the lower wages there. With Poland you have to pay more in wages, but still, you’re closer. Since the onset of the pandemic that has become very important. Any business would rather have an office in Poland than in India or in China at the moment, even if you have to pay a little more. The command of English is fairly good, you can communicate in English very well, it’s part of the EU, so you have some stability of the laws and regulations.
I remember transferring money to India and making an international wire transfer out of India. The share swap, I think it took me seven visits to the RBI, the Reserve Bank of India, to actually explain to these guys what a share swap was, and the law or the regulations of share swaps in India were only voted in half a year before we executed that share swap. I’m fairly sure these guys heard about a share swap for the first time from me. Just imagine these challenges, and in Poland it doesn’t phase them. Obviously the wages in India are cheaper, but I think these corporations weigh the balance between the quality and the price, and just the comfort and ease in us of doing the business. I think Poland is doing well, but that’s just going to become more and more challenging, and companies will start turning to countries like Serbia, for example. It’s not going to last forever, but we have our economy on our own as well, and a lot of startups from Poland that are also doing fairly well.
Esther: I wanted to get your take on another ongoing discussion in the language industry, around pricing models and the future of where pricing models are going to go. Do you see pricing moving long term to a per-hour model or a high bid model? How do you see that evolving?
Kris: I think the pricing model is a result of the business model in a sense, and if the industry sticks with linguists and puts a linguist in the center, I think the variation of per word will be the going business model, the going pricing model. If the industry believes in something like Exluency, where a subject matter expert is put in the center of the equation and not so much a linguist, then I think it will just turn to any subject matter business pricing model. You don’t pay lawyers by the pages they look at, you pay them by hour. I think the same will apply to the language industry. Obviously, with the continuous use of MT, the per work prices will naturally be deteriorating, but probably productivity gain will be higher. I think that per hour model makes sense because I think productivity is what matters and not so much how many words you look at and you can actually process. Companies like Lilt have figured this out and most of their linguists are now paid by hour. I think that’s the way to go.
An interesting idea we have been experimenting with quite a lot is the discussion around quality. What is quality, and the balance between paying the linguist and getting the right quality. What we’ve done, especially in the e-commerce industry (one of the four verticals that we service), is the real voice of the market. We don’t care what a linguist says about the quality of the translation, we want to check what the target audience says about it. What we do is, we have a standard workflow for the translation of product descriptions. The so-called SKUs, where we take the MT output and we would distribute through our own mobile app, we would distribute a couple of passages from the MT output product descriptions. We would distribute them between a crowd of five and buyers for a given brand. And we would ask them questions like does this sound right to you? Do you spot any issues? Do you see anything that would go viral and could be a problem for the brand? We don’t ask them questions like is it linguistically correct? Do you see a grammatical mistake in that translation? These are not the questions we ask, and if a product description score is five out of five, it automatically goes back to the CMS and is published at a client’s website. So there is no linguist in that equation. There is MT and there is the target audience validating the quality of the translation. I think that’s the way to go. How do we pay them? We pay them by hour, we don’t pay them by words they look at. I would say these are not linguists, so I think that’s the way to go.
That’s where data comes together. We use our mobile to crowdsource and I would say, it’s successful. It’s not a product, it’s a means of providing our service, but our app is used by over 50,000 unique users every month and it’s growing quite fast. For two months we tracked that and it’s not about downloading and paying us for the app, it’s about performing tasks. We have over 50,000 crowd members performing tasks on our unique platform every month. So it’s a fairly large product.
Florian: Interesting. Let me pick your brain on the outlook for 2021 and a little bit beyond, what are your plans? Where do you see the business going? The data and the language industry? Give me the crystal ball gazing.
Kris: I think the traditional language industry will stay the same. We’ll keep growing at it’s natural, 3, 5 per cent rate. Judging by our revenue and the people I know and have spoken with and even looking at these public announcements, the translation revenue is back to what it was, obviously with some industries suffering a little more, but I think that’s just going to be normal. The data is going to boom, is booming and will be booming, so that’s where we see a lot of growth potential. Consolidation will continue, so you’ll see us buying companies. You’ll see Acolad buying companies and everyone else buying companies. And smaller LSPs will see the threat and will want to exit and valuations will keep dropping. I think that’s also important. Bitcoin will grow.
You and I, Florian, have also spoken about this. Early in March you’ve asked me about my take on which industries are going to suffer and what’s going to happen. I was very reluctant to say ‘yeah, the e-commerce industry will boom because of the pandemic’. Maybe the e-commerce industry has boomed because of the virus, but it does not mean the LSPs have boomed alongside it. If you think about it, the supply chain from China, for a lot of e-commerce companies, has been suffering. You had the e-commerce companies selling more products, but not necessarily introducing a lot of new products and this is where we make money. I’m not saying ‘the world will change so much, there’ll be maybe more remote work than before’, but I think we’re lazy as humankind, we forget easily, so we’ll be back to normal.