WEBVTT 1 00:00:00.000 --> 00:00:04.000 Eemeli Ervasti: All right, welcome to the Neuvottelija channel. 2 00:00:04.000 --> 00:00:08.000 Today our guest is Sami Miettinen, the mean investment banker. 3 00:00:08.000 --> 00:00:35.000 Great, it's a bit scary being on this side of the camera. Luckily we get to edit this episode together, so we can crop this prosperity belly of mine out of frame. Yes, today we have a so-called special episode. I've prepared a few questions for Sami, and the idea came from me thinking I need Translink's consulting services but can't afford to pay for them. So then I suggested, 4 00:00:35.000 --> 00:00:42.000 with a bit of an ulterior motive, that I'd interview you for the podcast and get to learn something myself at the same time. So yeah — tell us again who you are, 5 00:00:42.000 --> 00:00:56.000 in a few elevator-pitch words. Sure — I'm Eemeli Ervasti, a business administration student at Haaga-Helia University of Applied Sciences, and an entrepreneur. I'm active here too, and a silver medalist in a business competition, something like that. 6 00:00:56.000 --> 00:01:00.000 Yes, exactly. And that's why I need your services. 7 00:01:00.000 --> 00:01:14.000 So, it was Europe's largest youth entrepreneurship competition, and there, with my business partner and classmate, we won silver. That's the background — just this summer. All right. 8 00:01:14.000 --> 00:01:34.000 Yeah, it's funny — when I came to your studio for the first time, you said, 'well, I'm just wearing a sweater,' so of course, as a good host, you offered me a blazer too, but I didn't take you up on it this time. Luckily it's not a sweater today, though. 9 00:01:34.000 --> 00:01:51.000 Yes. So, I've got three topics lined up, specifically related to your expertise and work — I know you co-wrote the book Neuvotteluvalta (Negotiation Power) with Juhana Torkki, and you also work at Translink. 10 00:01:51.000 --> 00:02:07.000 Right, so three topics. The first is how M&A works behind the scenes, plus some things from your clients' perspective. 11 00:02:07.000 --> 00:02:36.000 Want to comment at this point? Sure, let's just dive in — maybe we don't need to explain the whole agenda up front, let's go piece by piece instead. So, Translink Corporate Finance is an international group that provides corporate restructuring services in over 30 countries. Often, for example, the company gets sold, or it's listed on the stock exchange, or it goes through an equity round or some other recapitalization, 12 00:02:36.000 --> 00:03:05.000 and often clients come to us with one of these needs — or really, it could be any of them, because the client doesn't necessarily know exactly what they want, they just want some kind of change in ownership. I'll start philosophically here: ownership is a scarce commodity in Finland, and in a way we provide solutions to ownership-related problems. One of those can be the need to find a new owner and 13 00:03:05.000 --> 00:03:12.000 change the current form of ownership, or even convert its value into cash. 14 00:03:12.000 --> 00:03:23.000 Meaning making a so-called exit. Right. And just before we started filming, you said there's usually a common misconception about company sales. 15 00:03:23.000 --> 00:03:26.000 Mm-hmm. So do you want to elaborate on that here as well? 16 00:03:26.000 --> 00:03:45.000 Yeah, well, it's difficult, because what's being sold and transferred is ownership, and the company's current owner might actually be the best owner for it. They might also have a very close relationship with it — like, they work there themselves, or... 17 00:03:45.000 --> 00:04:11.000 an active owner who's constantly creating value — take them out, and it actually reduces the company's added value. And if that's the case, then that person is a big part of the company's value, and in a way we get into the question of whether this owner should then go on to work for the new owner and possibly reinvest in the company. This often happens, for instance, when the buyer is a private equity firm. 18 00:04:11.000 --> 00:04:37.000 But there's really so much to it — people might think it's just like selling a used car, that you're selling 'used shares,' but it's really not like that. Ownership itself is difficult, and so the transfer of that ownership is difficult too, precisely because of that. And then, well, potential owners are a scarce commodity, at least in Finland. Abroad there's plenty of capital, 19 00:04:37.000 --> 00:05:06.000 so if the company is good enough, owning it can be easier to arrange abroad. Which brings us to Translink — we're also able to find potential foreign owners for a company. So what different owner profiles are there? Right, so a financial owner is, in a way, private equity — because private equity specializes in ownership. It buys 20 00:05:06.000 --> 00:05:17.000 these ownership stakes in companies, and then typically ties in the management team — for example with share incentives — to build the company together with them. 21 00:05:17.000 --> 00:05:46.000 into an international giant in its field, growing it and continuing to grow it — but then again, they too have the goal of selling that ownership eventually, say after seven years. This is the most professional type of buyer, and often in Translink deals, a good exit partner as well. So that's one type. The second type of buyer, for example, is a family-owned company, a family 22 00:05:46.000 --> 00:06:15.000 office. For example, Translink arranged a deal for a SaaS service in HR and accounting management, where Hartwall Capital — a Finnish family investment office — bought the ownership and is now growing the company. So that's a private equity investor, and then there's this kind of family office. Then there's also the stock exchange — so all of us could soon invest in, say, Posti, the Finnish state postal company, instead of the state owning it entirely on its own. 23 00:06:15.000 --> 00:06:42.000 So in that case the pool of potential owners is very diverse. That's the third type. Then there can also be wealthy private individuals, or professionals in the field, who could be owners. And then, as maybe the last category, another company in the same field — usually a larger company that wants to buy a smaller one. So those are the kinds of categories out there. What other opportunities and threats 24 00:06:42.000 --> 00:07:12.000 are there in a potential Posti listing? Well, the state tends to be a poor owner in general, and the postal sector as an industry — these logistics services are quite challenging, and letter mail, which sits alongside those logistics services, is basically a dying product. So the business has been disrupted. And how well they've managed the transition from traditional mail to logistics services, 25 00:07:12.000 --> 00:07:38.000 whether the state, as an owner, has botched that value or not, remains to be seen. It looks like there's now profitability and scale there. But this is also a question of price — anything can be sold if it's cheap enough, and once the price is low enough, there'll be demand. So the pricing risk is definitely interesting — whether the right value can be found for it, at least to the point where 26 00:07:38.000 --> 00:08:06.000 I'd actually bother investing. I'm a bit cynical myself about IPOs, so I don't invest in them very often, but you also run IPOs. Yes. And that's exactly the point — you have to think about the fact that the world is full of cynical people, and you have to build the process, the company description, and the valuation so it's a fair deal for everyone. For example, in our latest listing, this excellent Tampere-based company called Tamtron makes these kinds of 27 00:08:06.000 --> 00:08:34.000 weighing solutions, and they're really strong in that space. Since the IPO they've grown quite a lot through acquisitions, creating shareholder value, and along the way they've had, for instance, Mika Heikkilä of Proprius Partners as a kind of co-investor, alongside Ensto. Sounds interesting. I believe 28 00:08:34.000 --> 00:09:04.000 this is going to be a long episode, as long as the content stays interesting. I've got a whole battery of questions here, so let's run through the topics — what really happens behind the scenes. Beyond that, I'm interested in your view of the future; as a 22-year-old, that's especially interesting to me. And then at the end, advice for owners — meaning buyers and sellers in this context — 29 00:09:04.000 --> 00:09:09.000 on building a business and being an owner, tips for that as well. 30 00:09:09.000 --> 00:09:37.000 Yes. So let's start with what happens behind the scenes — the 'mean investment banker.' When a client reaches out to you, how does the process start moving? Often we get new clients through recommendations from former clients, which is the most natural way, because they talk among themselves, without us, about how the Translink team handled things, and whether 31 00:09:37.000 --> 00:10:05.000 we kept our promises — and they sort of pass on the trust from the previous successful deal to the new client. That's quite good, because then they already know what they're getting into, and they've heard that we're reliable, sensible, and a good negotiating partner. Sometimes people come along who don't know anything about us and just reach out. 'Ah, these guys probably 32 00:10:05.000 --> 00:10:32.000 handle corporate restructurings' — and the worst kind is when a group sends a letter to, say, ten advisors, saying 'you're hereby cordially invited to make a very cheap bid on our process.' We often don't bother with those if that's the category they fall into. Yeah, this is interesting — Juhana Torkki, in Tarinan valtakirja (The Power of Story), makes the claim that a company has two 33 00:10:32.000 --> 00:11:00.000 ways to create its own story. And I've been thinking about this from a marketing angle — the fact that customers tell the story themselves, in other words recommend you onward, versus the company doing its own marketing and saying 'this is who we are, these are the services we offer.' So, a question related to that: have you noticed a difference between clients who come through a recommendation 34 00:11:00.000 --> 00:11:02.000 versus through your own marketing? 35 00:11:02.000 --> 00:11:30.000 So is there a clear difference between those clients? There is, yes. The one who comes through a recommendation is immediately warmer, because they're usually fine with us talking to the person who recommended us. And the recommender will often say — 'you went through the process with us, so just share your own story, tell them what it was like' — subject to the non-disclosure agreement and the limits of what can be disclosed, share your experience and maybe some tips, 36 00:11:30.000 --> 00:11:59.000 so trust plays a big role. There's this 'trust account' concept in the book, where you draw on the trust built up in a previous client's successful deal, so the new relationship starts with, say, a few dozen units already in the account. Whereas if someone comes in completely cold, say through inbound marketing, that trust might be based on nothing more than nice pictures and convincing references. Though we do have written references too, like 37 00:11:59.000 --> 00:12:26.000 from all those deals, where clients praise our professionalism and highlight certain things — but it's still a colder relationship. In general, building that kind of confidential space matters in almost every transaction, and it's especially important in M&A, and it includes meeting each other as human beings. Even though during COVID deals did get made where 38 00:12:26.000 --> 00:12:31.000 nobody ever met in person, that's still pretty rare. 39 00:12:31.000 --> 00:12:37.000 Have you made a conscious decision to focus specifically on customer satisfaction, and treat that as your marketing? 40 00:12:37.000 --> 00:12:46.000 Yes — this kind of reference marketing. I remember going through this mandatory 'guru phase' once, where we went through these... 41 00:12:46.000 --> 00:12:51.000 Brian Tracy books and others, and there was this one... 42 00:12:51.000 --> 00:12:56.000 referral marketing guru, and I somehow became convinced that... 43 00:12:56.000 --> 00:13:00.000 it's by far the best. A... 44 00:13:00.000 --> 00:13:10.000 recommendation from a previous customer in the same industry, doing the same kind of deal, is completely superior — at least in this case, since... 45 00:13:10.000 --> 00:13:15.000 investment banking work is such a significant, or in a way such an intimate, thing that it requires... 46 00:13:15.000 --> 00:13:25.000 that. Whereas with something more generic — say if you're selling phone plans — you don't necessarily need to collect that many references from previous... 47 00:13:25.000 --> 00:13:27.000 Elisa customers. 48 00:13:27.000 --> 00:13:45.000 So there's a spectrum here, but in this kind of demanding, expensive, small-team, high-stress process — investment banking or similar fields like strategy consulting — reference marketing simply wins out. 49 00:13:45.000 --> 00:13:50.000 Do you have any specific criteria for not taking on certain types of clients? 50 00:13:50.000 --> 00:14:17.000 Yes. Right now — for the past couple of years — we've focused on recurring-revenue B2B software services, or B2B SaaS as the industry calls it, typically with an enterprise value somewhere between 20 and 50 million euros. So it's quite a specific strategy, and those deals make up by far the majority of our transactions. 51 00:14:17.000 --> 00:14:44.000 So it's a bit like — if it's outside that segment, there needs to be almost a special reason for us to take it on. So maybe adjacent areas like software, AI, e-commerce, or media — if it's close enough to our focus, that's fine, but we really are very, very focused. And then if it's something like 52 00:14:44.000 --> 00:15:07.000 a deal completely outside that focus, then there'd have to be, say, a years-long personal relationship — which does happen — where we want to do it simply because we're such good friends that we'll take on the deal anyway. Some people really do take on just about anything that comes their way, but I think that's a bad strategy, because you inevitably end up doing bad deals. 53 00:15:07.000 --> 00:15:33.000 Are these usually buyers, or sellers who come to you? So sometimes we have acquisitions specifically with our international partners in the group, and there can be really massive chains of acquisitions. Like the Swedish company Miss Group, which is owned by the private equity firm Perwyn — they've made dozens of acquisitions with us globally, around 30 as I recall, and 54 00:15:33.000 --> 00:16:00.000 some of them here in Finland too. So there are those exceptions, but then there's also Smartvatten, which does this kind of water management — alarm and early-warning services for properties — and we've done buy-side acquisitions for them as well. One last question on the 55 00:16:00.000 --> 00:16:28.000 transaction side — how are your services priced, and why exactly like that? Yes, well, I can't give you an exact figure, but the price is in the hundreds of thousands of euros, so there's really no point hiring our team unless you're prepared to pay hundreds of thousands for it. And it's paid out of the purchase price — if the company sells for, say, 25 million 56 00:16:28.000 --> 00:16:56.000 and someone pays 25 million for it, the fee is paid out of that money. If no new money is actually coming in, it's much harder to pay, because then it would have to come out of individuals' own pockets, or the company's — and those kinds of sums usually aren't sitting in individuals' pockets — or out of the current owners' pockets. Which is why it requires 57 00:16:56.000 --> 00:17:23.000 an actual transaction where money changes hands, out of which the fee can be paid. So is it purely a commission, or is there also a fixed component? There is a fixed component — we have what you could call a materials fee. We don't do the retainer model. Many competitors do — a rolling monthly fee paid continuously for ongoing work. We instead have these milestone fees 58 00:17:23.000 --> 00:17:51.000 for preparing the materials — once we've got all the info memos (information memorandums) and the preliminary data room and so on set up, a certain fee is paid for that. Then for the signing of the LOI (letter of intent), there's another milestone fee, you could say. So those are fixed fees, not commissions, but they're still tied to a clear milestone. But then the hundreds of thousands comes from the 59 00:17:51.000 --> 00:18:01.000 success fee — that's only paid if the deal actually closes, and it's fair in the sense that if there's no transaction, nothing gets paid. 60 00:18:01.000 --> 00:18:14.000 So in terms of opportunity cost, it can end up being a really expensive project for us — we might have a team working on it for over half a year and end up empty-handed. 61 00:18:14.000 --> 00:18:23.000 Yes, those sound like large sums, and I'm also interested in the team side — what kind of people succeed in this industry. 62 00:18:23.000 --> 00:18:47.000 Yes. For example, we've got an 11-person team here at Translink's Finnish office. It includes partners like me, and also Tero Nummenpää and Jari Lauriala, for instance. At this level, we're usually the ones holding client responsibility, and in a way acting as the lead negotiator on the deal. 63 00:18:47.000 --> 00:18:51.000 Then there are the younger academic interns. 64 00:18:51.000 --> 00:19:19.000 There might be another partner involved as well, doing the project work and the day-to-day client work, and then there's the analyst level, and then we often have what we call 'analyst interims.' They're students near the end of their studies who, in a way, sacrifice half a year of their studies to not do any 65 00:19:19.000 --> 00:19:25.000 studying, and instead work a fixed-term, half-year stint with us, 66 00:19:25.000 --> 00:19:52.000 and get a proper monthly salary for it. We often have people like that on these projects, and for them it's a total jackpot — they get to learn about real deals, real things, so it's really meaningful, and the positions are highly competitive. But our permanent team is 11 people, and with that we close about five to ten transactions a year, 67 00:19:52.000 --> 00:20:17.000 in total. You said it's a jackpot, this kind of internship — so when the applications come in, what exactly do you look for? Yes, we have a very structured process. There are a lot of well-thought-out processes in investment banking work too, along with ready-made templates and standard operating procedures. 68 00:20:17.000 --> 00:20:45.000 But for this recruitment, we have a system where first we announce the position through student associations. There's a deadline a few weeks out, applicants send in their CV and transcript of records, we go through them, and then we invite about ten of them — sometimes more — to do a case study, 69 00:20:45.000 --> 00:21:06.000 where we give them a real company and say, 'value this.' We also have them describe the company's industry and competition, and then they do the work and present it. Based on that, we're already pretty close to a decision — like, if... 70 00:21:06.000 --> 00:21:09.000 it's been done well, then there might be very good chances... 71 00:21:09.000 --> 00:21:13.000 to get in based on that, but then there's still another... 72 00:21:13.000 --> 00:21:17.000 discussion, where all the possible partners, for example, get... 73 00:21:17.000 --> 00:21:22.000 to join that discussion, and then based on that, usually... 74 00:21:22.000 --> 00:21:28.000 about four people per year are chosen for these... 75 00:21:28.000 --> 00:21:32.000 interim roles. Yes. Now we'll move on to a slightly... 76 00:21:32.000 --> 00:21:37.000 future-oriented topic. So, how has AI changed your work so far? 77 00:21:37.000 --> 00:21:41.000 Yes, it has. This week I founded the Translink Corporate Finance vibe coding... 78 00:21:41.000 --> 00:21:47.000 society. So, this isn't a joke — we now have this, 79 00:21:47.000 --> 00:22:15.000 for example, GitHub, which is like a repository where you save code, and this week, while filming this, we've actually been vibe-coding real programs in Python. I mean, this is partly just playing around, but we also use, for example, the paid version of ChatGPT for everyone at the firm, with the security settings configured so that customer data isn't used as training material. Everyone knows how to use it — it just got this code block feature that lets you 80 00:22:15.000 --> 00:22:43.000 run Python code directly like that. So you don't need local environments or anything like that if you want to vibe code. That's what I mean by the vibe coding part. Then, I've also used it myself, for example in video production — I use Adobe's Firefly image AI. So when I edit images, I use AI quite a lot. Then I've got all these LLM 81 00:22:43.000 --> 00:22:59.000 models, like Gemini and Grok. Then there's Perplexity, and I've built quite a bit on that. Three weeks ago I switched to Comet as my main browser — the Perplexity thing that's got Jeff Bezos's money behind it. 82 00:22:59.000 --> 00:23:17.000 And it has this feature where every window you open in the browser is its own AI instance, which is absolutely wonderful — you can basically delegate tasks left and right, and it's like a small miracle when it actually gets things done. 83 00:23:17.000 --> 00:23:43.000 when you look at it. We also have various experimental systems in use — for example HubSpot has its own AI assistant now. Pretty much every CRM or customer data system these days is getting some kind of AI built in. We also have Microsoft Copilot, but it's so bad that almost nobody uses it. So we're quite picky about quality. 84 00:23:43.000 --> 00:24:09.000 Then various service providers pitch us things too. We had one AI tool meant for building infrastructure, but it was so bad that I half-joked we should just code a better one ourselves. So yes, AI is used very extensively, and in a sense we're AI natives — our people actually know how to direct the 85 00:24:09.000 --> 00:24:38.000 AI properly. It's not just 'here's a PDF, summarize it' — we do these small custom tasks. I myself, for example, coded a better OCR tool last year when I had a magazine article with bad print quality that Adobe couldn't read properly, so I ran it through an OCR library in Python instead of typing it out by hand. We're not exactly professional coders, but we're smart people, so 86 00:24:38.000 --> 00:24:44.000 we make these kinds of small fixes using AI. 87 00:24:44.000 --> 00:25:12.000 Does it come into play in the deals themselves — anything concrete, like processes getting automated, or something similar? Well, you do have to be a bit careful with data security. There are cases, for example, once the data room comes in — meaning the company's entire set of data — where you could run it all through these AIs. But you have to be careful that the company's confidential data doesn't leak into the LLM, so you have to, well, 88 00:25:12.000 --> 00:25:39.000 we have strict bans on that. But document summarization and translation, that we do — if there's a Finnish-language source document, for instance, that's done with AI nowadays. So those are the kinds of things. Then there's tracking down buyer candidates — yes, that's done too, like saying 'these six have been identified as likely buyers, so suggest similar ones' 89 00:25:39.000 --> 00:26:06.000 in these particular countries — that gets done. And then, well, some people do format their emails using AI. Personally I don't really bother with that sort of email automation, because I don't want to put much effort into email anymore, it just feels so old-fashioned, email. And I think in terms of interface, we've had Slack, for example, as a common discussion platform for five years, 90 00:26:06.000 --> 00:26:36.000 both within Translink and with clients, so these old-world tools started to feel really stiff by comparison. So what would a dream world look like, where, when you're running a deal, what else could you use AI for? Well, plenty could certainly be automated, and some of it already is. This manual work won't disappear though, because this transfer of ownership is so extremely confidential and personal that 91 00:26:36.000 --> 00:26:39.000 you don't want it to leak under any circumstances. 92 00:26:39.000 --> 00:27:04.000 These are processes that take place under non-disclosure agreements, and it's not a desired situation for them to be discussed openly. But if there was a slightly more cynical owner who didn't care that everyone would know they're looking to exit, you could pretty much just dump this into every database out there, saying it's available and people can get in touch, and 93 00:27:04.000 --> 00:27:28.000 maybe put out media, the kind of content that would let people know this asset is for sale. So that kind of thing — boosting the probability of a sale event, so to speak — could be automated by sharing information about the asset quite openly. 94 00:27:28.000 --> 00:27:42.000 Then, of course, you could actually try to automate the process itself, so you could see a future where AIs negotiate a sale and purchase agreement among themselves, without an investment bank or lawyers being needed. 95 00:27:42.000 --> 00:28:07.000 Do you really see that as the future? Well, in a sense you can model anything — you could take, say, 100 real M&A deals and their negotiation processes step by step, and model that. Then the bots just talk to each other while you watch from the sidelines. But then things can go wrong, and that's where the delegation problem comes in — and this is actually interesting, 96 00:28:07.000 --> 00:28:27.000 how much you want or dare to delegate to AI, and whether you still want a human check on top of it. An M&A deal is just — pretty much every single thing has to have the right decision-maker specifically sign off that this is how we're proceeding. 97 00:28:27.000 --> 00:28:53.000 So that kind of 'close your eyes and hope for the best' while process agents talk to each other is still quite far off. But there can be something like bringing everyone to the same level of information. In an acquisition, the current owner usually has very good information about the target being sold, and for the buyer it's like 98 00:28:53.000 --> 00:29:22.000 peeling an onion. First you give them teasers and info memos and management presentations, then comes the data room, and finally — in principle — you give all the information the buyer wants to know. And once the buyer and seller reach nearly the same level of information, trust is in a sense maximized, and when the deal is made, the current owner's entire 99 00:29:22.000 --> 00:29:47.000 know-how is transferred to the new owner, so you don't have to pay a kind of risk premium for the possibility that something was hidden. So this is generally the philosophy behind a company sale — that you really try to share as much as possible about the business to reach that same level, so there's no need to discount the price for the potential risk that 100 00:29:47.000 --> 00:29:50.000 the seller pulls a scam or withholds information. Mm-hmm. 101 00:29:50.000 --> 00:30:13.000 The same principle applies to, for example, real estate deals or car sales — in principle, a good car salesman or real estate agent lets you take test drives or actively tells you about the flaws, so there won't be a dispute later. You mentioned that you specifically sell SaaS companies, so how do you see the future of SaaS software developers yourself? 102 00:30:13.000 --> 00:30:33.000 Yeah, well, that's probably Finland's strongest area of expertise right now — B2B SaaS, meaning software for a specific need, like specialized payroll processing software, for example. 103 00:30:33.000 --> 00:31:00.000 as a software module of its own — or it could be something like Profinder, run by Ilkka Oja, which provides this kind of business information — you can look up a company's revenue, decision-makers, profitability, and other data, and compare across industries. 104 00:31:00.000 --> 00:31:28.000 They collect data and then, through some kind of user interface, it's used for specific functional purposes. Some of these use cases are, in a sense, easier to disrupt than others, and we produce this kind of quarterly report, the Translink Index, for SaaS companies, and there's a sort of two-way split — a horizontal 105 00:31:28.000 --> 00:31:58.000 SaaS and a vertical SaaS. Horizontal means you have a specific function, like payroll or ERP or accounting, or some other sales database like HubSpot. Vertical means it's for a specific industry — for example Admicom, a listed company that focuses on software services for the construction industry, and in a way 106 00:31:58.000 --> 00:32:24.000 it sits in that construction vertical. And these vertical SaaS companies — their valuation multiple is currently about five times ARR (annual recurring revenue), while the valuations of these horizontal SaaS companies have actually dropped, to just over three times recurring revenue. So already from that you can 107 00:32:24.000 --> 00:32:54.000 see — and these used to be roughly the same figure for a long time — so from that you can already see that it's better to be in a focused vertical than in a generic horizontal SaaS. That's one clear trend this year. Then there's this AI disruption — you can take pretty much any piece of software nowadays and reverse-engineer it with AI. So if I, an old investment banker uncle, can whip up code 108 00:32:54.000 --> 00:33:21.000 reasonably well based on my thirty-year-old programming skills, then imagine what a real software developer could do with AI — take a program that's making good money and think, 'let's sell this same thing, but usage-based instead of a monthly fee, so you pay for what you actually use.' And as an extreme example, Elon Musk announced that he's founded a company called Macrohard, whose job is to disrupt Microsoft. So, well, 109 00:33:21.000 --> 00:33:48.000 Microsoft is the world's most valuable SaaS company — it gives each of us email, PowerPoint, Excel, and so on. Well, Elon Musk's rather wicked vision is that these kinds of services could be bought on a usage basis from this AI-built Macrohard instead. Is this some kind of new trend? It's this kind of raw disruption — the idea of the SaaS model is 110 00:33:48.000 --> 00:34:15.000 that you're charged a monthly fee for a service, plus usage on top. So if you worked for me, I'd have to buy you Gmail, and then Microsoft and OneDrive so we could share files — and Microsoft's bill might, say, go up by 1.5 times once you start working there. Whereas in this Macrohard model, we'd just buy you usage-based service — so if you only use it maybe once a month, 111 00:34:15.000 --> 00:34:42.000 to send that one file, then it would only cost something like 15 cents that month instead of a full tenner. So this is a trend that's bound to break through at some point. But doesn't this sound really risky — a big risk? Sure, but there's opportunity in it too, coming out of the software side. Well, but then it comes down to focus, and that kind of specialization is always sort of like 112 00:34:42.000 --> 00:35:09.000 the kind of thing where you don't have the energy for every single industry or language market — even a giant's payroll management, say, since nobody really wants to deal with Finland's ridiculous labor regulations, cooked up by a bunch of old-timers, to the point where even AI starts smoking at the ears trying to handle it, so they just leave it to the humans. So there are these 'moated' niches, like 113 00:35:09.000 --> 00:35:36.000 culture, language, special legislation, and that whole 'we're doing this exactly according to working hours and collective agreement terms' approach — the collective agreement appendices have to be in order, and people can't be bothered to deal with those. There can be industries that are quite unique in that way, so they won't really get disrupted. But this kind of disruption, or flipping 114 00:35:36.000 --> 00:36:03.000 an industry's prevailing business model — you can do that with a quick, vibe-coded minimum viable product, test whether it starts selling, and only then put real resources behind it. It's actually a great time for this, because these dinosaurs can fall — suddenly a giant like Microsoft or SAP or whoever is no longer 115 00:36:03.000 --> 00:36:08.000 protected. When did these things first start to emerge? 116 00:36:08.000 --> 00:36:12.000 Well, it's this AI that's been around the whole time. 117 00:36:12.000 --> 00:36:40.000 So, as I said, even though Translink has also been using AI — ChatGPT 3.5 — for over two years now, it has to be done smartly, because a lot of people wasted a ton of money last year building their own fine-tuned language models, which now looks completely pointless. We never went down that money-wasting road, that 118 00:36:40.000 --> 00:37:07.000 it needs to serve a real, practical purpose, a bit like an app. Those are being built all the time, and Finns have been a bit slow on this front. But, for example, I had Mikko Alasaarela here as a guest — he founded Agentics Finland, which pulled in about a thousand people in a single day, a group of people in Finland who are also taking this AI revolution seriously. Can you join it? Yeah, I'll send you the link again, so 119 00:37:07.000 --> 00:37:30.000 well, yeah, it's a WhatsApp group, and we've got a meetup coming up there soon. There are these hardcore coder types there too, people who actually know how to build things — specifically build practical applications, not just some nonsense. Sure, all of us know how to code some philosophy bot on an LLM, but you don't really make any money with that. 120 00:37:30.000 --> 00:37:34.000 Now I'm shifting the topic a bit, away from AI and technology. 121 00:37:34.000 --> 00:38:02.000 Yeah. I'm personally interested, as a 22-year-old, in this — if you were my age now, with, say, 10 years ahead of you and Finland as your market, what would be an industry where you personally see potential in the future — and I mean not AI or tech. Yeah. Well, physical, real-world services will always exist. In other words, we all need 122 00:38:02.000 --> 00:38:29.000 some kind of warm home, and in that sense there will always be services tied to that kind of living. Then, well, you can think of Maslow's hierarchy of needs — you need a home, food, a relationship, sex, that sort of thing, then you need a job in general, and then 123 00:38:29.000 --> 00:38:57.000 the highest needs, self-actualization, start to emerge — where we start thinking about philosophy, visiting art exhibitions, educating ourselves through literature or art — and across all of these, the human element will remain. Of course they can develop alongside AI and in a way bring the transaction cost down, but it's going to stay there. I'd 124 00:38:57.000 --> 00:39:25.000 say that if we rule out AI and data, then robotics is really interesting. I mean, Elon Musk has been developing this Optimus humanoid robot for a couple of years now, and they're already frighteningly good. So we're starting to get human substitutes into physical spaces — in the coming years those humanoid robots will arrive, and if nobody in Finland has really started leading in that space, it could 125 00:39:25.000 --> 00:39:41.000 be quite interesting to get in on that with these physical robots. Right now those Wolt delivery robots are rolling around Helsinki, but in a way the next step from that is bringing humanoid robots in 126 00:39:41.000 --> 00:39:58.000 here into physical space. They'll of course have AI features built in too, but that's something I don't see as much of a business in itself, so I think if someone became, say, the Optimus reseller for Finland, that could be a pretty big thing. 127 00:39:58.000 --> 00:40:15.000 Well, what if we say we don't tie this to any particular industry. But that I'm starting a company now, and my goal is that within 10 years I'll build the business and the brand, make it profitable, and then I want to exit. 128 00:40:15.000 --> 00:40:24.000 So if that's the goal, what would be the important things to keep in mind right from the start? 129 00:40:24.000 --> 00:40:51.000 Well, ownership is a scarce commodity, and only the kind of ownership that's internationally interesting is truly liquid — meaning it's realistic for a large pool of buyers. In terms of negotiating power, there's a real tension here: if you don't have alternatives, all you have is terms dictated to you. So when it comes to your exit, 130 00:40:51.000 --> 00:41:16.000 if it depends on a single card — like one company in the same field buying you — that's not good. You should have many possible paths to that exit, and it should be possible to make the company internationally interesting. For example, on the SaaS side, that means preferably over 5 million in annual recurring revenue. Then it attracts interest abroad too, and it's large enough that 131 00:41:16.000 --> 00:41:42.000 there start to be dozens of people working there, so that ownership becomes transferable. So, in a way, raising the scale to a level where it starts to be internationally interesting brings those exit options. Then there's your own role — in a way you need to reduce it, if you want to turn it into cash so you're no longer working at the company or having to reinvest in it. Well, in a certain way, you have to make yourself a bit 132 00:41:42.000 --> 00:42:12.000 unnecessary. Meaning you might stay on as an active owner, but perhaps no longer as CEO or the single most important person, because if you're the CEO and the key person, the new owner will often say, 'we can own this together, but you're not going anywhere.' So — that can be fine, maybe that's what you want and you stay on. But if the goal is to be a serial entrepreneur, to grow a business large and 133 00:42:12.000 --> 00:42:29.000 sell it and move on to the next one, then that won't work. So, in a sense, it's about shaping that ownership structure right at the start. In the beginning, in companies, ownership and management are always the same thing — the owners are hands-on. 134 00:42:29.000 --> 00:42:58.000 I'd say that especially if you're not doing it alone, if you have business partners, then the shareholder agreement is extremely important — it needs to be well thought out and also cover things like: if it doesn't work out, the way it sometimes doesn't in marriages, and you need to part ways, on what terms and how does that happen, and are there mechanisms that allow for it. You should never assume the situation will stay exactly the same in, say, 135 00:42:58.000 --> 00:43:02.000 five years' time within the owner group. 136 00:43:02.000 --> 00:43:31.000 These kinds of things. Of course the business model should ideally be able to generate its own cash flow and fund itself, or else from the start there needs to be some external funding channel bringing in the cash flow for that growth. And that has to be thought through, because if the business naturally consumes cash flow for growth and you end up running it in a mode where it has to earn its own money for that 137 00:43:31.000 --> 00:43:44.000 growth, that inevitably leads to slower growth. And then, again, that first goal — becoming a large, internationally viable company — gets pushed back. So those five years might stretch to, say, eight, 138 00:43:44.000 --> 00:44:08.000 or so. But do you mean there aren't good opportunities right within the Finnish market itself? Well, there are, yes — for example Hartwall Capital, which is a purely Finnish family company, so yes, they do exist, and you shouldn't be pessimistic about it. But in a certain way, you shouldn't start a company just to exit it — you should start a company because it's a good 139 00:44:08.000 --> 00:44:37.000 company that generates shareholder value for you too. So the idea that shareholder value can only be created through an exit — that's the wrong way to think about entrepreneurship or starting a company. In principle you should run the business as if it could be yours, your ownership, perhaps even forever, so you don't end up thinking, 'I'll just run this for a five-year stretch to an exit and then I'll throw 140 00:44:37.000 --> 00:44:57.000 in the towel.' That mindset doesn't necessarily result in a good company. And on another note — is it visible, or have you heard, that now that the population is aging, there's been an especially high number of ownership changes, or that they've increased because of this? 141 00:44:57.000 --> 00:45:24.000 We have this really unfortunate problem, where this right-wing government backed down at the last minute in a cowardly way; it was very close to us doing what other civilized Western countries — Sweden, Norway, and others — have done. We didn't change the inheritance tax into a capital gains tax, we chickened out, and it's a terrible mistake. And if we now get a blue-red coalition or an SDP (Social Democratic Party)-led government, then this won't get fixed for, like, 142 00:45:24.000 --> 00:45:52.000 six years, this problem. So we're stuck with this awful period where this sort of 'death tax' remains. I mean, heirs would have to pay, on a potentially completely illiquid company, something like 7% — or more, it rises above 20% if you're not a close relative — tax in advance to the state. And then once you get the shares, you'd have to figure out whether the company can pay 143 00:45:52.000 --> 00:46:21.000 you dividends, say, or something else to cover those taxes, and/or whether you could sell it so you at least get those taxes covered. So this is a kind of fear — as people age, and since building a genuinely good company has typically required a 10-to-20-year growth runway, the owners are often approaching what you might call the danger zone of death, meaning over fifty years old. So they have to think about 144 00:46:21.000 --> 00:46:50.000 how the heirs will be able to own the company and pay this inheritance tax, and that logically creates pressure to exit. This kind of pressure doesn't exist in, say, Sweden or Norway, because you can die anytime — from a heart attack, say — and it won't trigger any inheritance tax. So you can die without worry there. But in Finland, because of that... and this applies to everyone, from the 145 00:46:50.000 --> 00:47:01.000 Vaisalas down to even large listed companies — that seven percent is a lot. Especially if it's a growth company that doesn't pay dividends, it's a really big incentive to exit. 146 00:47:01.000 --> 00:47:12.000 And then other people who'd otherwise have the money, like those over fifty, it's not worth it for them to buy either, because they know the company even less well. They'll die with the same probability as you, so they won't buy. 147 00:47:12.000 --> 00:47:26.000 But then again, a Swedish person of the same age can buy it, because once again it doesn't matter — dying doesn't trigger the inheritance tax there, only selling does. 148 00:47:26.000 --> 00:47:34.000 Could there be a market niche here for an active private equity investor? Think of a company where the owner is currently also the manager. 149 00:47:34.000 --> 00:47:43.000 Yeah. And the owner wants an exit, so a private equity investor buys it and brings in a new manager, for example as CEO. 150 00:47:43.000 --> 00:48:08.000 Yeah, you're absolutely right, and that's exactly what these capital owners, these investors, do — they provide a solution for this. We also had, for example, this great Tampere-based company Aurora Yhtiöt, which does this kind of serial acquisition, buying up smaller companies and possibly combining them under certain industry umbrellas. These 'compounders,' and it's good that they've emerged. 151 00:48:08.000 --> 00:48:11.000 In Sweden there are dozens of these — Sweden has 152 00:48:11.000 --> 00:48:20.000 private equity investors, plus even wealthy individuals can buy, and you don't have to fear death there, and on top of that they still have dozens of these compounders, 153 00:48:20.000 --> 00:48:46.000 by the dozens. So the Swedes are like an incredibly lucky country in having this ownership culture in what's basically the best corner of Europe for it. And Finland is really only just now starting to even slightly appreciate ownership, because we have this kind of hatred toward ownership. A perfectly typical narrative in the press right now seems to be, 'oh, if only there were fewer rich people.' So, I mean, purely 154 00:48:46.000 --> 00:49:09.000 just because it's not understood that, in a market economy, ownership is a scarce resource, and if we don't let Finns own things, then the ownership of those smaller businesses can disappear — and with it, market-based employment disappears too. 155 00:49:09.000 --> 00:49:37.000 That's kind of what we're experiencing here — this 10% rise in unemployment, and the fact that, even though it's really hard for your generation to find these entry-level positions, as you said, part of that is because we haven't understood that ownership is a scarce resource. But think about this, Sami — for our generation it's hard to find entry-level positions now, like you said. I was at Haaga-Helia University of Applied Sciences last semester, and we had this course called the Year as an Entrepreneur program, where you set up a company 156 00:49:37.000 --> 00:49:41.000 that, in the eyes of the tax authorities, looks like a hobby activity. 157 00:49:41.000 --> 00:49:46.000 So we start this kind of company that operates with real money. 158 00:49:46.000 --> 00:50:08.000 And it included these 'Dare to Try' competitions, where the winner of the Finnish round went to Europe, to the EU finals, which you mentioned at the start. And guess how many — well, you know the answer, but the listeners can guess for a moment how many competitors there were at the university level. 159 00:50:08.000 --> 00:50:35.000 There was a big fat zero. Our company was the only one that took part in this program, so I was just thinking — if those starting slots don't even exist, then what does my generation do? It's a great opportunity, but maybe it just hasn't been marketed enough, or whatever the reason is. But yeah, this is, for example, Finland's — Sanna Marin's — and 160 00:50:35.000 --> 00:51:03.000 the 2023 YEL (self-employed persons' pension insurance) reform, made by the tripartite (government, employers, and trade unions), was a catastrophic error. I mean, the fact that an entrepreneur now gets pre-billed based on the pension insurance company's estimated earned income, at a rate of 24.8%, is such an incomprehensible brain fart that I don't know how 161 00:51:03.000 --> 00:51:17.000 something like that got through, other than these people being so out of touch with how business actually works that they figured they'd manage to scrape it together from somewhere. But anyway, this came at the worst possible time, with the stupidest possible system. 162 00:51:17.000 --> 00:51:38.000 Just when ownership had become a scarce commodity in Finland because of thin capital and anti-business, anti-ownership tax policies — this YEL business made running a company as, say, a one-person firm, unreasonably difficult, and a total cash flow risk. 163 00:51:38.000 --> 00:52:08.000 So now this pension contribution is basically a forced expense, and this really needs to be fixed as a matter of urgency, because it's, in a way, become the natural refuge for someone whose life is already a bit — for an alienated young person, they end up in the unemployment register, or as a perpetual student, or they don't start a family and so on, they somehow get stuck. Or then they take — this is 164 00:52:08.000 --> 00:52:35.000 the kind of thing where they don't dare take a risk. And it's really awful to hear how these opportunities either aren't communicated, or people don't even bother to look — so even this should be pushed through with AI, like, search out every possible competition like this, here's a site, here's what you can see, and go sign up for these, because it could simply be that nobody knew they existed. Yeah, well, that's very possible — and it did feel a bit bad being there in the finals of the Finnish round when there wasn't 165 00:52:35.000 --> 00:53:02.000 anyone else competing. We then got to Athens and took silver — there were about 23 other teams besides us. By the way, before you ask — back in the stone age, when Juha— well, Jyrki Katainen was Prime Minister, I got to meet him, and I pitched him the idea that everyone should have a Business ID (a Finnish company registration number) in addition to their social security number. In a way, to make it 166 00:53:02.000 --> 00:53:30.000 equal — because this system in Finland is a really off-putting one: if you have a Business ID, you become a kind of welfare-state outsider. So coming back from that, to becoming a recipient of social benefits and so on, has been made tricky. I don't — I don't know about this, can you tell me more? Yeah. So, if you own a company, then in the eyes of these social benefit schemes 167 00:53:30.000 --> 00:53:46.000 and so on, you're treated as an entrepreneur who should be able to earn a living through that entrepreneurship — i.e., through that limited company's income — and you're not really allowed to, say, shut it down just so you can get into the employee category. 168 00:53:46.000 --> 00:54:13.000 So then the safe option becomes never owning anything, because that way you at least get social security, unemployment benefits, and so on. So there's a huge incentive for people to just be employees and avoid entrepreneurship. And now that things like the YEL bomb have been loaded on top, it's been made even more off-putting. But if everyone had a Business ID — meaning everyone was basically an entrepreneur — then 169 00:54:13.000 --> 00:54:42.000 the system couldn't play this game of 'ah, you're an entrepreneur, go earn your own social security or unemployment benefits by just doing better at some gig work.' I think that would be a healthy thing. Similarly, the equity savings account should be mandatory for everyone, and the state should put in, say, 300 euros for every child born, so that everyone becomes a capitalist from birth. Then we'd have, in a way, these 170 00:54:42.000 --> 00:55:08.000 basic building blocks, or hygiene factors — a Business ID, an equity savings account for securities, and of course a social security number — and then we'd have all kinds of equal opportunities, and the system wouldn't be able to discriminate against you just because it has, from its perspective, such a convenient excuse to reject you, like having a Business ID. 171 00:55:08.000 --> 00:55:38.000 Well, I've had a Business ID for four years now, since high school, so this isn't a familiar problem for me. What have I missed here — is it that if you've run a business and the company's since been closed, and you then register as unemployed, the employment office says, 'you're an entrepreneur, not unemployed, so no, we won't pay you anything, maybe some support — but are you sure you couldn't invoice the company for something and earn your cash flow that way?' Okay, well, I've never 172 00:55:38.000 --> 00:56:06.000 actually been a wage earner, so if I were a wage earner and this happened, then what? Well, for a wage earner it works like this: you register as an unemployed jobseeker, you go through the employment office, and the system tries to push you into finding work. First you get earnings-related benefit, which is based on a multiplier of your previous salary, and then after a certain time 173 00:56:06.000 --> 00:56:25.000 that runs out and you drop down to basic unemployment allowance. But if you have a Business ID and are an entrepreneur, the system might say you don't deserve this — try invoicing something through that Business ID instead — so they won't even let you into the employment office system. 174 00:56:25.000 --> 00:56:29.000 Okay. Well, luckily I don't have to worry about any of that myself. 175 00:56:29.000 --> 00:56:40.000 Hey, I've still got a few more questions I'd like to hear your answer to. What are the typical mistakes people make when carrying out ownership changes, transactions? 176 00:56:40.000 --> 00:56:45.000 The kind that people don't usually think to watch out for. I mean, the most common mistakes. 177 00:56:45.000 --> 00:57:15.000 Yeah, every deal is of course unique, but you could think about the deals that didn't end up going through, and why. A poor shareholder agreement is probably reason number one. Basically, you have the current group of owners, and they've agreed in the shareholder agreement on the ground rules — like who gets to decide when the company is sold, and why — and if that's poorly thought out, then basically 178 00:57:15.000 --> 00:57:42.000 even a small minority owner, who might not even work at the company, could block the acquisition just because they're having a bad day, or something like that. That should be prevented, and it can be fixed before an exit, but it's definitely worth getting it right from the early stages — using good, carefully drafted contract templates. That's a 179 00:57:42.000 --> 00:58:07.000 pretty lethal issue, honestly. There can also be situations where, say, management salaries are below market rate, and then at the exit stage they want to negotiate better pay right then — but that conversation happens too late, and it increases the company's costs, reduces profit, and affects the purchase price. So 180 00:58:07.000 --> 00:58:16.000 that kind of thing is a major risk. Of course, having done a lot of these deals now, we never fall into traps like that. 181 00:58:16.000 --> 00:58:44.000 So we really do check management salaries against the shareholder agreement in good time, and discuss it directly — whether the salaries are currently at market rate and whether they need to change — and if so, they get adjusted right then as part of the process. So from that you already get two of them: terms of employment, and problems with the shareholder agreement. Then, of course, there's people's unrealistic expectations — there might have been, for instance, 182 00:58:44.000 --> 00:59:10.000 some potential buyer who tossed out an offhand figure that isn't based on anything, maybe a high one — 'we'd like to buy you for, say, 20 million' — and then, looking at it completely objectively, even if the company's actually worth 13 million, the seller thinks, 'well, they must know what they're talking about, so we definitely need to get 20.' So that kind of quite irrational expectation, which then becomes 183 00:59:10.000 --> 00:59:37.000 a millstone around the whole process — and it might be driven entirely by emotion or ego: 'I won't sell unless I get that 20 million,' even if the objective value is lower. It can also go the other way — selling too cheaply to the first person, just because they seemed nice. In those cases we usually wouldn't be involved in the process at all, because we'd say, 184 00:59:37.000 --> 01:00:04.000 'don't do it that way — let's run a real process and approach a lot of people,' because, as I said, if you have no alternatives, all you have is terms dictated to you. So there can also be a mistake in the other direction, where you sell far too cheaply to just the first nice guy who happened to make an offer. Hey, I heard some good news — that the Negotiation Power book is getting more attention again now, right? 185 01:00:04.000 --> 01:00:21.000 Yes, we did another print run of it this year, and it's now sold over 30,000 copies across various formats, which is nice. It really is the overwhelming market leader in Finnish negotiation literature — though that's a pretty low bar, since we don't have much competition in this genre. 186 01:00:21.000 --> 01:00:26.000 Hey, if I go buy this book and read it, what can I get out of it? 187 01:00:26.000 --> 01:00:55.000 Well, there's the philosophy, for one. It has these classic examples from Finnish negotiation history — like how Risto Ryti tricked Hitler into getting German help during the Continuation War, or how Nalle [Björn Wahlroos] sold — sorry, sold Sampo Bank to Danske Bank at an overprice, or how Risto Siilasmaa negotiated with Alcatel- 188 01:00:55.000 --> 01:01:25.000 Lucent and things like that, or Martti Ahtisaari's thoughts on peace negotiations, or Bengt Holmström's game-theoretical ideas. But then there's the theoretical framework Juhana and I developed ourselves, where there's this, in my opinion, quite robust, time-tested idea that a good negotiator uses four basic levers — namely power, analytical skill, sociability, and being principled. And for each of those there's 189 01:01:25.000 --> 01:01:38.000 an analysis of exactly why they matter. And for example, within that 'being principled' lever, building a 'trust account' is one of the main concepts. 190 01:01:38.000 --> 01:01:47.000 Hey, finally, to wrap up here — what do you say, can you tell the viewers, from your perspective, how I ended up here? 191 01:01:47.000 --> 01:02:15.000 Yes, well, through sheer persistence, of course. A bit like — you blasted through with a cold call. I usually screen calls, so if it's an unknown number I just say my first name and check whether it's some scam caller from India or someone who actually happens to be Finnish — but you got through that filter with some real energy, and then you already had your pitch ready. By the way, there always has to be 192 01:02:15.000 --> 01:02:36.000 a goal for every interaction, a clear next step — so you managed to sell that low-commitment ask, a Google Meet call. And then, with this 'yes ladder,' we moved on from there to a meeting — a classic cold call, then to Teams, then to an in-person meeting, 193 01:02:36.000 --> 01:03:03.000 and from there to the studio here. Yes, it's been really nice, it's been great getting to know you, and I don't have anything else to ask at this point. Great. And it's fun being on this other side of the table, by the way — so leave some comments about how we did here, and whether we should do more of these 'grillings.' And since you've watched this far, be sure to subscribe to the channel. Actually, one more thing 194 01:03:03.000 --> 01:03:27.000 came to mind — if anyone has questions, they can drop them in the comments, and if we do another episode like this, we'll pull questions from there. Yes, that works really well — nice closing move there at the end. And actually, if you want, you can also share your contact details in case people want to brainstorm with you on these things — so where can people reach you? 195 01:03:27.000 --> 01:03:30.000 Isn't that only for the inner circle, then? 196 01:03:30.000 --> 01:03:34.000 I don't know, it's not... well, you can find the contact info online. Yes. 197 01:03:34.000 --> 01:04:01.000 The name was Eemeli Ervasti. It's been a pleasure being here. Yes, and since you've made it all the way to the end, please subscribe to the channel, and we'll of course bring Eemeli Ervasti into the inner circle — and we'll actually chat a bit more about that transition into working life from the perspective of a 22-year-old, whether you go the entrepreneurial route or into an organization, and how to tackle this AI challenge, 198 01:04:01.000 --> 01:04:10.000 and how human working-life skills combine with AI. Would that work? That sounds very interesting.