The Business

Basically Votum helps ordinary polish people sue institutions, mainly banks and insurers for money they were unfairly charged. A normal person goes to Votum with a contract that a court somewhere has already decided that those type of contracts are illegal and Votum’s lawyers fight the case for them. If and when the client wins the case, then Votum keeps a percentage of the money the client makes from the lawsuit.

The company reports five revenue segments but one segment represented ~85% of all revenue in 2025.
That segment is the cash cow that has brought the company here and the segment I believe the market is discounting as already dead.

For you to really understand this segment, a little backstory...

In the 2000’s to 2008 when the Polish economy was booming, real estate prices were risings and like in many other places in Europe and around the world, owning a home was the way a person could say “I made it” to their family and their friends.

However, there was a problem with trying to own a house in Poland at that time. In the year 2000, the reference interest rate for banks hit a high of 19% (yes.. 19%…) in August. It is also true that the reference rate dropped in the following years to an average of about 6% leading up to the 2008 crisis. Those kind of rates seem like an impossibility in today’s world, but at that time they weren’t. If you were a bank manager at that time, even though rates are 6%, the fact they were 19% just a few years ago makes you very scared of the possibility they could go up again. That means the in the loans you’re going to give out, you will only accept a higher rate than 6% to make a nice margin of safety. In those years, the rates on house mortgage loans averaged around 9%.

Since the rates were considered high by the consumers, to prop up demand for house mortgages, banks began marketing the house loans either indexed or denominated in Swiss Francs (CHF).
They started doing it because interest rates in Switzerland were much lower at that time, hovering about 3% for that time period. Because of this, the monthly payments on the loans seemed to be around half of what they would otherwise be if they were contracted on Zlotys.

What Was Legally Wrong?

Banks stated that the loan amount would be converted from PLN to CHF on the day of disbursement using the bank’s own buying rate. Also, the monthly installments had to be paid in PLN but were calculated using a foreign exchange spread table determined unilaterally by the bank itself.

This created a big imbalance where consumers had zero bargaining power or transparency on those rates. The banks effectively bought and sold the currency at whatever arbitrary rate it pleased, embedding a hidden, recurring profit margin into the loan. This transactions meant that the banks transferred 100% of the long-term currency risk to the consumer without clear and transparent warnings about the catastrophic downside of a depreciating Zloty.

Well, what ended up happening? The zloty slow and steadily devalued against the swiss franc. See below:

What this ended up meaning is that the people who entered into this house mortgages, even though they had been paying every month for 10 or 15 years, they would see that the principal they owed on the house had actually become close to twice as large as when they entered into the loan.

And we all know the EU has strong consumer defensive laws and views. In 2019 the Court of Justice of the European Union (CJEU) issued a ruling on this, saying the bank’s actions were abusive and illegal. This meant that the banks had to return every single Zloty the consumer paid in interest and installments over the last 15+ years.

When Votum really began making money

Before this, Votum made personal-injury claims from car accidents, but this CHF story unlocked a whole different world for the company. You can see it below in the revenues.

You can see the growth was basically zero before 2019, then it starts growing gradually until it explodes in 2022. And it has a reason that has to do with how they record revenue. They make money with three separate things. First they charge a small upfront fee for taking the case, then they charge court attendance fees and finally they charge the success fee if they win the case for the client. The first two ones don’t contribute much, the last one is the real driver of cashflows.

Even so, in 2019 you can see that revenues jumped 34% only by the upfront fees of signing a lot of these cases. (This is also one thing that separates Votum from the rest of the other companies, and I will explain it further down. It has to do with the way they gather clients).

However, these type of cases typically follow a two-step process. The 1st Instance Judgment takes about 1 to 2 years. Even if Votum wins here, banks almost always appeal the decision to buy time. The 2nd Instance Judgment (final and legally binding) takes another 1 to 2 years.

Only when the 2nd Instance court rules in favor of the consumer is the mortgage contract legally terminated and the bank is forced to pay. From initial signature with Votum to final judgment, the process takes anywhere from 2.5 to 4 years and this is where the big discrepancy is created.

Under the IFRS 15 accounting standard, Votum records the bulk of its massive success fee revenue the exact moment they win the 1st Instance court verdict, which yes, is a bit aggressive.

Here’s how it works: Votum wins the first round in court. The bank will obviously appeal (which takes another 1–2 years), but Votum looks at their historical data and sees: “Hey, we win 98% of these appeals anyway”. Because the odds are so heavily stacked in their favor (~98% success rate), accounting rules allow Votum to say “This money is virtually ours.” So they estimate how much they will collect in 2 years, discount it slightly for the time it will take to arrive and book it immediately as revenue and profit today. Since the bank hasn’t actually paid yet (and is still appealing), that money doesn’t enter Votum’s bank account right away. Instead, it sits on the balance sheet on a line called “contracts assets”. That line, in the end of Q1 2026 had a value of 381 million zloty, basically the entire EV.

What this means is that the company you buy today for 382 million (considering you collect the cash on hand right away), has 381 million (which are already discounted for time value at 4,25%) of contract revenue that they will most likely receive, since it is already only 98% of what would be possible to receive if they won every case. One thing is for sure, on a nominal basis, the money that will enter the company will very likely be superior to what it costs to buy the whole business today.

The Illusion of the Drained CHF Pool and the Path Forward

As I said in the beginning, I believe that the consensus of the market is that this CHF pool has now been drained and the company will die without it. I usually say “in my opinion”, but this time I will say that the above is just factually wrong. Let’s look at the numbers.

The CHF mortgage book peaked at around 700,000 active loans originated in 2008-2010. Most have since been repaid, settled or won in court. But at the end of 2025, six of the largest banks still reported more than 231,000 CHF mortgages outside both settlement and litigation. In just Q1 2026 alone, they signed 1,733 new CHF contracts with new clients and they expect to sign close to 8,000 in 2026 in total. This was increased from an earlier estimate they had of 5,000-6,000 in 2026. How you may ask? The answer has two sides, how they acquire clients and the reality of the CHF market today.

The company’s secret sauce to acquire clients and a potential corporate governance red flag

Now, this part I’m going to write about, could either be a moat the company has or a potential red flag.
Votum is much better than the other law firms at acquiring clients, which in my opinion is the big reason why they are the best in their space in Poland. Instead of doing TV commercials or Google ads, they get clients in a very specific way.

They use an exclusive partnership with a network called DSA Financial Group.
Instead of waiting for an angry bank borrower to look for a lawyer, Votum utilizes thousands of independent sales agents who literally go out into the field, knock on doors and attend community meetings.

When the Swiss Franc crisis escalated, these agents were deployed across every corner of Poland. They approached borrowers, analyzed their mortgage contracts on the spot for free and handed them a pre-packaged lawsuit solution. This massive and aggressive distribution channel allowed Votum to get market share at a speed that traditional law firms couldn’t match. It created a massive customer acquisition moat.

Now, the problem is that DSA Financial Group was founded by the same person who founded and owns the largest stake in Votum, which is Andrzej Dadełło. He owns 57% of the company and has been chair of the supervisory board since 2005.

Now, the fact that Andrzej Dadełło is such a controlling figure in the company could be a red flag, but I will explain further down why in this case I don’t think it is. The potential problem here is that he owns DSA Financial Group, and if things turn bad for Votum in the future, he could decide to increase the fees paid to DSA and extract capital from the main company at a loss for minority shareholders. Of course that would also hurt him since large part of his wealth is tied up to Votum, but a paper loss on Votum’s stock wouldn’t hurt so much if a physical gain was being made in his other company at the expense of Votum.

He is not the CEO, that part we’ll see in the next section, so he can’t really decide on these matters, but I’m sure he could probably heavily “influence” the decision.

The reality of the already paid off CHF loans

You see, in April 2026, Votum’s management explicitly stated that the market is only at roughly 50% penetration of the total addressable loans. The market ignores that beyond the active mortgages, there is a massive list of potential clients of approximately 300,000 borrowers who have already fully repaid their CHF loans. These people are still legally entitled to sue for retroactive compensation and why wouldn’t they? If you saw your neighbors who had the same type of contracts sue the banks and receive a lot of money back, you would probably want to do the same, even if you already paid off your house loan. Votum has recently shifted its aggressive sales network to specifically target this specific type of people who have been dormant until now, proving the pure CHF well is far from dry because they keep signing more and more contracts. As I said before, they expect 8,000 for 2026

On top of this, when the banks were issuing the CHF loans, they were also issuing EUR indexed loans with the same kind of terms as the CHF ones. While these are not nearly as many as the CHF ones, according to the finance authorities of Poland like the Komisja Nadzoru Finansowego and the Biuro Informacji Kredytowej (don’t ask me to pronounce these), there are approximately 60-65,000 active EUR loans still being paid off. Contrary to the CHF loans, I couldn’t find any information on the total of EUR loans issued, but if you use the CHF ones as a proxy, since 30 to 40% of them have been paid off (but are still able to being sued by Votum), take the middle of it at 35% and apply that to the EUR ones and we get approximately 92,000 of these loans that were issued and are possible to sue.

At the end of 2025, the company reported that the EUR contracts already represented 25% of the new contracts signed, which means they are redirecting part of the customer acquisition machine to this segment. Given that they signed 7,800 contracts in 2025, that means close to 2,000 of them were EUR contracts. In Q1 2026, it was already 32% of new cases. As I said before, Votum is the clear leader in Poland for these kinds of things now, at the peak of the CHF contracts, they controlled close to 30% of the market share. If you apply that to the 92,000 possible contracts, that means they can theoretically catch ~27,000 contracts in total. Even if you apply a margin of safety and say they only get 20,000 of them, there are still a lot of EUR contracts to be done. When the CHF contracts start to cool down in 2027, the EUR ones will replace them to keep a level close to 8,000 contracts signed.

Let’s not forget, this is all money on top of the 381 million already booked in the balance sheet! I explain in the valuation part how much I expect this to turn into FCF.

The infinite runway

So, what happens when these contracts eventually run out? Does the company die like the market is saying it will? No.
The company is already targeting a new market of contracts. These have much lower economics, but instead of a few hundred thousand like the CHF, there is an army of millions of these contracts in Poland.

The contracts i’m referring to are the Sankcja Kredytu Darmowego (SKD). What is it?
In Article 45 of the Polish Consumer Credit Act, SKD is a severe regulatory penalty designed to protect everyday consumers. Over the past decade, Polish financial institutions aggressively issued millions of standard consumer loans, cash credits and car loans. In their rush to scale this up, banks systematically committed technical and mathematical errors in their standard-form contracts, ranging from miscalculating the true Annual Percentage Rate (APR) to illegally charging interest on non-loan expenses (such as upfront insurance premiums or administrative commissions).

The legal consequence of these systemic errors is absolute (although keep this in mind to see the risks section, where I explain why this isn’t as certain as it seems): if a contract for a consumer loan up to 255,550 PLN contains an infraction, the borrower has the legal right to invoke SKD. This strips the contract of all profitability for the lender. The borrower is only required to repay the exact principal borrowed and the bank is legally obligated to fully refund every zloty of interest, insurance premiums and processing fees ever collected.

Unlike highly complex mortgage litigation, which requires a lot of legal work, SKD is processed by Votum like an automated assembly line.
They put their customer acquisition machine to work and the DSA Financial group network collects historical and active standard consumer credit agreements from millions of potential retail clients. Then, contracts are funneled through an AI-driven legal matrix that scans for known banking errors based on the specific bank’s contract templates. Because individual claim amounts are smaller (typically thousands of PLN rather than hundreds of thousands, which is why the economics are worse than a mortgage loan of hundreds of thousands), banks face poor risk-reward dynamics by fighting in court. Because of this, the rate of out-of-court settlements is substantially higher and court cases are settled far quicker than mortgage disputes.

Unit Economics of SKD, Run-Rate and Perpetual Cash Generation

The operational profile of the SKD segment offers two structural advantages. There is near-zero capex requirements (as it fully utilizes the legal back-office infrastructure already bought and paid for by CHF profits) and a drastically shorter cash cycle of just 1 to 1.5 years (compared to 3.5 years for mortgages).

The CEO has set an operational baseline to exceed 1,000 to 1,250 signed SKD contracts per month. In a normal, cruise speed environment, this translates to a steady-state run-rate of 12,000 to 15,000 contracts per year.
Because these are smaller consumer loans (on their Q1 report, they mention the average contract they signed was for 80,817 PLN), Votum’s average success fee per each case they win is significantly lower than in the mortgage segment, averaging 10,000 PLN. I will also expand further on this in the valuation section.

Management and Capital Allocation

CEO Bartłomiej Krupa joined in 2007 as a junior clerk and worked his way up the legal department. He’s been president of the management board since 2022. He’s a lawyer, not a finance MBA and he communicates with shareholders more openly than almost any Polish micro-cap I’ve seen so far. He gives quarterly investor chats on StockWatch.pl and writes honest annual letters. The forecasts he gave on the late-2024 chat I read have largely been met. That counts for a lot with me.

On capital allocation, the recent record is overwhelmingly dividends. They grew from 4.9m PLN in 2022 to 30.1m in 2023 to 34.8m in 2024. For 2025 in total, they distributed 100 million PLN, by paying in October 2025 1,67 PLN per share, then again 1,67 PLN in December 2025 and a final payment in May of 2026 of 5 PLN per share still attributable to 2025 profits.

As I said, the company is Net Cash, of about 154 million. They have 182million in cash, ~16,9million in debt and ~11,2million in leases.
For 2026 and 2027, the company has stated they will distribute at least 60 million PLN each year, which is about 5 PLN per share. If the annual profit is above 60 million, they will distribute more than that just like they did for 2025.

Below is the cash generated by the business since 2016:

Below is what they have paid in dividends:

While the payment of May 2026 doesn’t enter the cashflow statement of 2025, it still referred to that. So in 2025 200 million PLN entered the company and 80 million were paid as dividends. For a company selling for 382 millions, I would say that is pretty shareholder friendly, not just for the majority owner but for us minorities as well.

I do have one complaint though, share buybacks at these prices would be the perfect capital distribution method and more tax efficient, so that’s a negative. On top of that executive pay is tied to consolidated profit and segment results, not to free cash flow per share, which is not my favorite kind of incentive.

This kind of dividend policy also indicates that the corporate management red flag I highlighted before probably isn’t a big risk, but one to keep on mind for position sizing.

Risks of Permanent Capital Impairment

As usual, if you read my earlier posts, you know I rank the risks on a very subjective basis. I rank them by the probability I see and severity, which you might find to be different from your analysis. If you can find more than the ones below, please let me know.

1. The realization of the 381 million backlog. The entire margin of safety in this thesis is the contract-asset line, so it’s worth asking what happens if it doesn’t convert to cash as cleanly as the accounting implies. The number is already booked at a 97-98% win-rate and discounted at 4.25%, but “winning” the second instance and “collecting” the cash are not the same event. A bank that loses can still drag time on the payment, enforcement through the Polish courts can be slow and in a genuinely stressed scenario, say a wave of resolutions all at once or a bank in trouble, that 98% assumption could prove a touch optimistic, and a 5-8% reduction on that line would be material in zloty terms. I rank the probability of the mild version (timing slips, the cash just arrives later than I assume) as medium and the probability of the severe version (an actual impairment of the asset) as low. Severity of the mild version is low, it mostly pushes my IRR around. Severity of the severe version is moderate, because even a 10% writedown still leaves the backlog worth roughly 90% of what I’m paying for the entire company.

2. Polish wage inflation, specifically against the SKD model. In the CHF world this barely matters, the success fees per case are so large that paying lawyers more is a rounding error. In the SKD world it matters a lot, because there you are catching sardines, thousands of small cases that each need handling and the whole segment depends on operating leverage over a large, fixed and salaried back-office. If Polish legal salaries keep climbing at high single digits while the average SKD success fee stays small, the FCF margin on that leg compresses exactly the way I assumed in my conservative case in the valuation section and the perpetual annuity gets thinner. Probability: high, this is happening, not a maybe. Severity: low to moderate, because it's a slow grind on the least important leg of the thesis, not a shock to the part that's already contracted on the balance sheet.

3. The SKD thesis doesn’t materialize. This is a very real risk, and the one that could really affect an investment in Votum. In 23 of April 2026, the Court of Justice of the EU (CJEU) ruled in the case C-744/24 favorably towards consumers. While the CJEU decision is indeed binding regarding the interpretation of EU law (like I said before, specifically that banks cannot charge interest on credited non-interest costs like insurance or commissions), the domestic sanction for this violation is still governed by national law in Poland. Under Poland’s Consumer Credit Act (u.k.k., Art. 45), the drastic penalty of SKD (which strips the bank of all interest and fees) is only triggered by an exhaustive, specific list of statutory infractions (such as failing to disclose correct information in the contract).
   
   Bank lawyers are aggressively arguing that even if charging interest on credited costs is a civil contract violation, it does not automatically constitute a breach of the specific information duties that trigger the SKD under Polish law, especially if the APR was mathematically calculated correctly based on those terms. In short, they argue the punishment (SKD) is disproportionate to the breach.
   There have been cases already where this legal defense worked in Polish courts. Following the April 23 CJEU ruling, several regional courts including the Warsaw District Court in late April 2026 issued rulings unfavorable to consumers. Although, then on the Warsaw Court of Appeals, many of those were turned in favor to the consumers.
   
   Instead of an immediate wave of easy consumer victories, we are highly likely to see a prolonged “tug-of-war” in the courts. The next 12 to 18 months will see highly non-uniform rulings. First-instance courts will be divided until either the Polish Supreme Court or subsequent CJEU clarifications (such as the upcoming opinion on case C-831/24) explicitly rule on whether the SKD is a proportional sanction for charging interest on credited costs (which comes out tomorrow, 11 June 2026 btw).
   
   What I see happening is that the contracts that had clear mathematical mistakes in the APR and other easy to rule cases, will be won easily and the other ones will have a much lower win-rate than the 98% of the CHF cases. I would say the win-rate would be around 80-90% based on the fact that they have a high IT infrastructure and will optimize to only accept the best cases out of the millions of contracts out there.
   
   I consider is a medium probability of being an adverse scenario to the company and a high severity, because it would cut the second part of the investment thesis.

4. Controlling-shareholder action. I covered this above. A low-ball take-private or simply increasing the DSA marketing fees would deteriorate the investment thesis as well. I would rank the probability as low on the terrible version of a take private version, but the milder fee increase version is genuinely medium because the channel already exists. The severity ranges from moderate to existential for minorities.

These are the main ones that I can see really hurting the investment thesis.

Intrinsic Value Framework

I don’t use spreadsheets as you know. Let’s build the earnings power piece by piece, because the whole disagreement with the market is about timing and decay and both are very knowable thanks to that IFRS-15 lag and the contract volumes we already saw.

The CHF run-off is not running off yet. In Q1 2026 they signed 1,733 contracts and 46% of them, almost half, were of people who had already paid off the loan.

The unit economics of a whale. The average CHF case in dispute is around ~350,000 PLN. Votum’s total take (success fee plus the other fees) is roughly 12%, so ~42,000 PLN of gross revenue per case. With net FCF margin around 30% each CHF contract signed today is worth roughly 12,600 PLN of clean free cash flow.

They expect to close around 8,000 contracts in 2026 and I said before I expect them to close another 8,000 in 2027 because of the EUR cases. From what I saw, analysts are projecting less than me, around 5,500 cases, so let’s assume they are correct. Assuming a win-rate of 97/98% on these cases, we get 7,760 won cases from 2026 and 5,335 won cases from 2027.

Applying the lag, because of the 2-3 year wait to a final judgment the cash from a contract signed today doesn’t land until 2028-2029, we get the below:

7,760 sign-ups in 2026 × 12,600 PLN ≈ ~98m PLN of FCF arriving around 2029.

5,335 sign-ups in 2027 × 12,600 PLN ≈ ~67m PLN of FCF arriving around 2030.

This is only from the mortgage loans. The run-off everyone is pricing as imminent has effectively been pushed to the end of the decade and the cash that funds the dividend is already locked in by signings that have happened or will happen in the next 24 months.

Now, the unit economics of sardines. After the CHF whales stop existing, the plan is SKD as a perpetual catch of very small sardines. I will build a very conservative scenario and my base scenario for each case I will present below.

In Q1 2026, the average SKD contract was 80,817 PLN with a success fee of 12,38%, as I said before. I would say this is the base case since they are doing this in the midst of the regulatory turmoil, when things get defined in stone maybe it could even improve, so let’s say the conservative case is an average contract of 50,000 PLN, with a success fee of 12%. That’s revenue of 6,000 PLN per case.

If instead of the 30% FCF margin, they only get 15% FCF margin because of wage inflation, the FCF per case is 900 PLN. That’s for the conservative case.
For the base case, I would assume the 80,000 PLN average contract, 12% success fee and 20% FCF margin, we get 1,920 FCF per case.

Now, let’s estimate the average volume of SKD cases per year.
The base case management is expecting is 12,000 per year, so let’s say the conservative case is half of that, at 6,000 cases per year. Let’s also assume a win-rate of 85% for the base case and 75% for the conservative case.

In the very conservative case we get: 4,500 won cases x 900 PLN = ~4 million PLN
In the base case: 10,200 won cases x 1,920 FCF = ~19.5 million PLN

Now.. That’s a very big difference between the conservative case and the base case but keep them in mind.

For the rest of the segments the company operates, in Q1 2026, the personal injuries segment made 6.9 million PLN of revenue and the rehabilitation segment made 8.1million of revenue (with consistent YoY growth of 5%). If we annualize that we get 27.6 million for the injuries segment and 32.4 million of revenue for the rehabilitation. In a very base case for both, for simplicity, if you assume 10% FCF margin on that, which is usually higher than that, you get 60 million total revenue x 0.1= 6 million PLN of FCF on a steady state basis, without counting for revenue growth.

Now, if you put it all together, the sardines and the other segments, you get:

On a conservative basis: 10 million of steady state FCF
On a base case: 25.5 million FCF

So now, putting it all together… The company has 12 million shares outstanding. At a share price as of today of 43,10 PLN, you get a market cap of 517.2 million PLN.
With net cash of 154 million, what you are really paying for the company is 363.3 million PLN.

So, what are you buying for 363 million? You are buying a company which on it’s books has already contractually 381 million PLN to receive in the next 2-3 years, which are also discounted for the 98% win-rate and time value at 4,25%.

On top of this, we have to add the FCF it will receive in 2029 and 2030 from the cases it is signing now.
To make sense of how this cash actually lands in Votum’s bank account, we have to map out a timeline. We have three main cash flow engines:

The Backlog Collection: The 381 million PLN of contract assets currently on the books. Since the operational costs for these cases were already spent in 2023–2025, they convert into almost 100% pure cash inflow as they win their 2nd instance appeals. We will assume this is collected over the next 3 years (2026–2028).

The New Cases (with a 3-year lag): The ~98 million PLN from 2026 signings landing in 2029 and the ~67 million PLN from 2027 signings landing in 2030.

Other Segments + SKD (Steady-State): This is the stable, non-CHF cash flow. We will assume the other segments generate a steady 6 million PLN per year. As the SKD “sardines” business model scales up, it joins this pool to reach our conservative steady-state target of 10 million PLN per year by 2028.

In 2026:

• Cash from Backlog: ~130 million PLN (this is an estimate on what Votum collects in the first major wave of 2nd instance wins from the 381M backlog. If it collects less this year, it collects more next year, but in the Q1 report they had 142 million in the short term assets, so it will likely be more than the 130million).

• Other Segments: +6 million PLN.

• Total 2026 FCF: ~136 million PLN.

In 2027:

• Cash from Backlog: ~140 million PLN (The remaining bulk of the 381M backlog resolves in 2nd instance).

• Other Segments + Initial SKD: +8 million PLN.

• Total 2027 FCF: ~148 million PLN.

In 2028:

• Cash from Backlog: ~111 million PLN (Clearing the last of the 381M backlog).

• Conservative Steady-State Non-CHF: +10 million PLN.

• Total 2028 FCF: ~121 million PLN

In 2029:

• Cash from 2026 contracts: ~98 million PLN (The 8,000 contracts signed in 2026 finally reach final 2nd instance judgments).

• Conservative Steady-State Non-CHF: +10 million PLN.

• Total 2029 FCF: ~108 million PLN.

In 2030:

• Cash from 2027 contracts: ~67 million PLN (The 5,335 contracts signed in 2027 reach final judgments).

• Conservative Steady-State Non-CHF: +10 million PLN.

• Total 2030 FCF: ~77 million PLN.

As I write this, I noticed that I've treated the 381 million backlog as almost pure cash in 2026-2028, which is right, because the costs of those cases were already spent back in 2023-2025. But in those same years the company is also spending money working the new 2026 and 2027 cases, the ones whose cash doesn't show up until 2029-2030. So I'm collecting the backlog "clean" while assuming the collection of the money for free inside the same buckets, which means my 2026-2028 numbers are a little flattered and my 2029-2030 numbers a little understated (by then those costs are behind us).

The sum across the period barely moves, I still land around 590 million either way, but the cash is less than my rundown makes it look. If you calculate an IRR for the numbers above you get close to 21%. The real number is slightly lower than that, maybe by 2 percentage points. It changes nothing about the thesis, because the floor under all of it is still that the backlog alone is worth roughly what I'm paying for the whole business.It also changes if you assume a terminal value for the 10 million FCF steady state and it jumps to the mid twenties.

But you don’t need to do an IRR calculation. If you sum all of that cash, you get 590 million.
For 363 million today, you receive in total 590 million in FCF until the end of the decade, you don’t need a calculator to know that’s a good return, especially when you consider it’s based on the very conservative numbers and the margin of safety given by the already contracted cashflows.

Conclusion

At the end of the day, Votum is the classic case of a stock where the market is so focused on the imminent “death” of the business that it has completely forgotten to look at the cash on the table.

If you bought a house today for 363k PLN, and the seller handed you the keys along with a legally-backed contract guaranteeing you’d receive 381k PLN in cash over the next three years, you wouldn’t spend your time worrying about whether the pool you’re planning only gets built in 2031. You’d buy the house immediately.

That is exactly what is happening here. At today’s prices, you are buying the net cash and the secured and already-won backlog of cases for less than they are worth. Even if the SKD “sardines” thesis takes longer to materialize due to court friction and even if we apply an ultra-conservative lens to the other segments, the sheer margin of safety provided by the current balance sheet and the immediate cash inflows is massive.

To me, this is a classic “heads I win, tails I don’t lose” scenario. The market will eventually realize that Votum isn’t a dying cigar butt when cash keeps flowing.

Disclaimer: I am long Votum and please, for the love of god, do your own due diligence. I don’t speak Polish, I trusted the machine to translate the filings for me and I’m still a new-to-the-game investor learning a lot every day. None of this is financial advice, just my opinion.

Thank you very much if you made it all the way down here, this was a long one. Every critique, correction or just a hello is welcome in the comments and if you spot a grammar mistake, please point it out, English still isn’t my first language. (Or just forgive me and move on.)

Best regards,
Tiago

I’m writing to you again but this time because of a Polish company. While doing my A-Z there I found a fish company that I thought was interesting and cheap enough to be worthy of a write-up here.

This post is going to be about another cheap stock one buys for a basket of statistically cheap companies and waits for the math to math it out. I warn you, this isn’t the kind of dirt-cheap stock that triples overnight since it doesn’t have a gigantic gap to close (book value is around 40% higher than today’s price). But there’s a nice margin of safety in the balance sheet, an 11% FCF yield if the cash on the balance sheet isn’t distributed and closer to 17% if it is.

It’s basically a small profitable business that you buy cheap, with a gap to book value that may or may not close, an ok dividend of about 7% that pays you while you wait and a controlling family whose entire compensation, in any real sense, is the share price going up. Even if the gap never closes or the cash doesn’t leave the company, between the 7% dividend and the profit it keeps and reinvests every year, you earn more than the Polish government bond. The rest comes for free.

The company is SEKO S.A. (GPW: SEK), based in Chojnice, Poland. As I’m writing this it trades around PLN 11.15 per share.

Business Description

SEKO buys fish, mostly herring with some mackerel, pickles it, fries it, puts it in jars and cans and plastic tubs with sauces and salads and sells it to the big Polish supermarket chains. Poles eat a lot of pickled herring around Christmas and Easter, so the company makes most of its money in the fourth quarter and the run-up to Easter and the least in the warm summer months when nobody wants to think about oily fish. There is also a single petrol station they run under a Circle K franchise which management openly calls a sideline it will never expand (I appreciate the honesty, it would have been very easy to dress that up into a “diversified retail platform” in the investor deck).

In 2025 total sales were PLN 243.4 million. Of that, products and services were PLN 228.1 million, about 94% of the total and up 4.2% over the prior year. The petrol station and other resold goods were the remaining 6% and fuel volumes actually fell 13%, so nothing exciting there.

Here is the first important number. Fish products by weight were 16,295 tonnes in 2025, up 4.2% from 15,631 tonnes in 2024. So almost the entire revenue increase was volume, not price. Revenue per tonne actually slipped slightly year over year even as volume rose. Hold that thought, because it tells you most of what you need to know about this business: there is no pricing power here and I don’t think there ever will be.

Where does it sell? Domestic sales were 91% of the total and exports 9% (mostly Czech Republic and Slovakia, then Germany, the US, Lithuania and France). Domestically, the retail chains take about 60% of sales and they are what you would expect: Biedronka, Lidl, Kaufland, Aldi, Carrefour, Auchan, Dino, Netto, Żabka and Eurocash.
Biedronka represented 11.6% of 2025 sales and no other customer was above 10% which has to do with another factor, because the chains deliberately cap any single supplier at 30-40% of their purchases and that will become relevant later in this analysis.

The cost structure is what you would expect of a thin-margin processing business. In 2025 materials and energy were 64% of total costs which were the single dominant line. Wages were about 20%, external services (mostly transport) about 11% and depreciation under 4%. Materials are largely variable because you cannot make a jar of herring without buying herring and herring is a euro-priced global commodity that SEKO imports from Norway, Iceland and Denmark. The real dependency in this business is not a customer, it’s the euro price of North Atlantic herring.

One genuinely good structural change is that the materials-and-energy share of costs has been falling for years (it used to be over 70%) and a big part of that is a new cogeneration plant the company built (I had to go and learn what cogeneration even was, it’s basically generating your own electricity and heat on site instead of buying it off the grid). In its first full year it cut energy purchase costs by about 13.5%. For roughly PLN 4-5 million of capex that produces close to PLN 1 million of annual savings which is around a 20% pre-tax return and is hard for a small rival to copy.

Gross margin sits stubbornly around 18% and net margin around 4%. Management itself says the chains “exert strong pressure to keep prices low.” Basically SEKO is sitting between a global raw material it does not control and a handful of discount chains that won’t let it raise prices.

TIKR (and probably others) will show you a 2025 gross margin of around 4.5%, down from 17% the year before, which makes it look like the business just fell off a cliff. It didn’t. This is purely an accounting artifact and it’s worth understanding, because it’s exactly the kind of thing that scares people off a stock nobody is watching.

Here’s what happened. A change in Polish accounting law forced the company to reclassify where certain costs sit in the income statement, moving a big chunk of cost into the “cost of goods” line and leaving the selling and admin line nearly empty. Now, the important thing to understand is that moving a cost from one line to another doesn’t change anything real. The operating profit is identical either way because you subtracted the same cost in the end, you just put it on a different shelf on the way down. But it does crush the gross margin on paper, because gross profit is calculated before that lower line. The data provider then mapped this reclassification badly, dumping everything into the cost line and leaving the other blank, which is what produces that scary 4.5% number on the screen.

The annual report tells you the real story which is that gross profit on sales was actually up about 6%, and operating income was stable. The true margin is right where it’s always been, around 18% gross and 4% net. The fact that this kind of thing scares off the screen-filtering crowd is part of why a debt-free, net-cash, dividend-paying business is sitting here this cheap in the first place.

Industry Structure

The Polish fish-processing market is worth roughly PLN 16 billion. Poles eat about 12-13 kg of fish per person per year against an EU average closer to 22 kg, so there is theoretical room to grow, though I wouldn’t build a thesis on Poles suddenly deciding to eat more herring. On PLN 243 million of revenue SEKO is around 1.5% of the whole market, which sounds tiny, but that overstates the relevant competition, because SEKO plays in one specific slice: pickled and marinated herring and fish salads and pastes. In that specific niche, the trade analysis I read from the Polish Marine Stewardship Council says Lisner is around 35-40% and SEKO close to 15-18%. SEKO is small against the whole thing but a clear number two in its own category.

Now the cyclicality, which is the part you really have to pay attention. Revenue is fairly steady, but margins are not. If you walk the net income line over the last decade, it bounces around the mid-single-digit millions and then in 2021 it essentially vanished with net income of PLN 0.22 million. The cause was the 2021-22 raw material and energy spike that the company could not pass through to the chains fast enough. I went to check this to see if management were overstating it but it’s true. According to the Polish Institute of Agricultural and Food Economics the import purchase price for wild-caught Atlantic fish skyrocketed by more than 25% in 2021. The reverse image was 2023 which was a peak year of PLN 14.94 million helped by falling input costs and some one-offs. So this is a low-margin business whose worst visible drawdown was a near-total wipeout of one year’s profit, followed by a full recovery. Cyclical, survivable, but you have to respect that 2021 number.

The big recent event in the industry is consolidation. In November 2024, Lisner (the leader, owned by Germany’s Theo Müller group, a roughly €7 billion parent) acquired the processing business of Graal, the other large player. So the giant above SEKO just got bigger. Your first reaction, and mine, is that this is bad for the little guy. But the new CEO has a contrarian read that I find genuinely persuasive.

The retail chains deliberately cap any single supplier at 30-40% of their purchases in a category, because no buyer wants to be hostage to one producer. Now that Lisner and Graal are one company, that combined entity bumps up against the ceiling much faster. The chain can’t just keep buying more from the leader, its own purchasing policy won’t allow it. So it is structurally forced to give shelf space and volume to a credible number two to keep its supply diversified. And the only credible number two in this niche is SEKO. The CEO also points out that after Lisner and Graal, only much smaller firms are left, so there isn’t an obvious next deal to worry about.

I’d treat this as a plausible and structurally logical tailwind rather than money in the bank, because it is still the read of a CEO who has every reason to sell it and the chains could just as easily decide they want cheaper private label instead. But if it plays out, it lifts SEKO’s volume and maybe its margin. I’ll come back to why that matters for the return at the end.

Management and Capital Allocation

The family holding company, Złota Rybka (which translates to “Golden Fish” which I found to be pretty cool), owns 62.49% of SEKO. Two different pension funds collectively hold about 25.74% and the free float is roughly only 12%. Inside the family holding, the founder Kazimierz Kustra, who ran the company for over thirty years, holds the majority with his son Tomasz and the daughters-in-law holding the rest.

The central recent event is the succession. On 30 September 2025 the founder announced he was stepping down from the management board at year-end and his son Tomasz Kustra became CEO from 1 January 2026. This is not an outsider parachuted in. Tomasz has been at SEKO since 2001, climbing through settlements, sales, CFO and Vice-President, so 25 years in the building, a finance-and-operations man who knows where everything is. The founder stays on the supervisory board. A clean, telegraphed, inside succession is about the best kind there is.

Now the compensation is the best part. Total 2025 pay for the entire three-person management board was about PLN 666 thousand. Not each. Combined. And every single złoty of it is fixed salary. Zero variable, zero options, zero equity grants, no severance and no pension top-ups. The policy even caps any board member’s monthly pay at ten times the average company wage. Set that against a business earning roughly PLN 10 million of net income and doing PLN 243 million of sales. Three executives running a company this size for two-thirds of a million złoty.

If the management here wants to get richer, they have exactly one lever to pull and it isn’t their salary. It’s the value of the 62% of the company their family owns, which is to say the share price. Their pay packet is, in any meaningful sense, the stock going up. This is the alignment we’re always claiming to look for and almost never actually find, the kind where you genuinely don’t have to worry that the people running the business are quietly enriching themselves at your expense through some gameable bonus metric. They eat what you eat. The only thing a purist could complain about is that there’s no per-share value-creation target written into the pay which is redundant because the family’s entire fortune already is the per-share value.

The capital allocation record:

Acquisitions: none, ever. The company has grown entirely organically. The new CEO has signalled he’d look at an opportunistic deal, which is the one thing to watch, because a first-time acquirer is always a small risk, but the family’s thirty-year track record is one of caution which is good.

Buybacks: none. With the stock at about 0.70x book, buybacks would be expected from great capital allocators, so this is not great.

Dividends: disciplined, roughly 50% payout. They’ve paid steadily for years, and the one detail I love is that they skipped the dividend entirely in 2022, right after the 2021 profit collapse, which is exactly when a prudent owner conserves cash. The proposed dividend for 2025 is PLN 0.74, which at the current price as I’m writing this is a yield of roughly 6.6%.

Surplus cash goes into bonds. And here we get to the thing that makes me want to write about it.

The Bond Portfolio Hiding Inside the Fish Company

Rather than let excess cash sit idle, SEKO parks it in PKO BP corporate bonds and Polish treasury bonds, held to maturity with maturities running from 2029 to 2033. At the end of Q1 2026 this bond book was PLN 24.85 million, earning the company over PLN 1.3 million of interest income a year.

Add to that the regular cash on the balance sheet of about PLN 7.9 million, subtract total debt of PLN 13.29 million and you get a net cash position of roughly PLN 19.5 million once you count the securities.

So to calculate enterprise value, at PLN 11.15 a share and 6.65 million shares, the market cap is about PLN 74 million. If you credit the full bond book and net it off, the enterprise value of the actual operating fish business drops to around PLN 54.5 million. Even if you’re a pessimist and treat the bonds as untouchable money you’ll never see (a fair worry under family control, more on that in the risks), and count only the regular cash, the EV is around PLN 80 million.

That gap, the difference between what you pay and what you’re actually buying once the bonds come out, is the whole opportunity.

Intrinsic Value Framework

Because the net income jumps around too much to trust any single year, you have to try to build a normalized owner earnings number.

Mid-cycle net income looks like roughly PLN 9.5 million (the 2024 and 2025 run-rate). The 2023 figure of PLN 15 million was a one-off-aided peak, 2021 was the rough part of the cycle and the trailing twelve months look flattered because an early Easter this year pulled sales forward into Q1 2026, which will partly reverse, so I wouldn’t annualize that. To net income number I add back depreciation of about PLN 8 million, then subtract maintenance capex. This last one is a judgement call, because reported capex of PLN 8-10 million in recent years carried a lot of growth spend (the cogeneration plant and two new packaging lines) and in years with little growth spending capex was only PLN 2-2.5 million. So true maintenance capex sits below depreciation and I’ll call it about PLN 7 million.

That gives me owner earnings of roughly 9.5 + 8 − 7 = PLN 10.5 million on the income approach versus a cleaner multi-year free cash flow average closer to PLN 8 million. I’ll split the difference and deliberately lean toward the more conservative cash figure and call normalized owner earnings about PLN 9 million.

Now the test I always come back to, free cash flow yield against the local risk-free rate. The relevant benchmark here is the Polish 10-year government bond, which has been bouncing around 5.5-5.9% this year and sits near 5.7% as I write.

• On the bond-adjusted EV of about PLN 55 million: PLN 9m / 54.5m ≈ 16.5%

• On the conservative cash-only EV of about PLN 80 million: PLN 9m / 80m ≈ 11%

So even on the most pessimistic enterprise value this offers around 11% against a 5.7% risk-free rate. SEKO yields roughly two to three times the Polish 10-year depending on whether you count the bonds or not. That is a meaningful premium.

Over the past decade equity grew by about PLN 45 million in retained earnings while annual earnings rose by maybe PLN 3-4 million, which is an incremental return on reinvested capital in the low double digits at best. The cogeneration project earns its ~20% on a small base, but the bulk of retained capital is sitting in bonds yielding ~5%. So reinvestment here is roughly value-neutral. I’m considering this at no growth and treating any volume gains as cherry on top.

Risks That Could Permanently Impair Capital

As always, I list the risks I see and rank them on probability and severity, in the simplified and subjective way I always do. You’ll likely weigh them differently, and you might spot one I missed.

1. A herring supply and cost shock, plus the euro. The mechanism is the 2021 scenario repeating itself: tightening fishing quotas and a weak złoty raise the euro cost of imported herring faster than the chains allow price increases, and the 4% net margin compresses to zero. Probability: medium. Severity: severe but existential only in the worst-case scenario where herring stocks are structurally depleted and the cost base rises for good. This is the main thing to watch (gross margin per tonne, EUR/PLN, North Atlantic quotas).

2. Retail buyer power and private label. A few discount chains control shelf access and a sustained shift to chain-owned private label or losing a major chain, would hurt volume and the brand. Probability: medium. Severity: moderate-to-severe.

3. Secular decline in herring eating. Polish fish consumption is already low and it comes from older people and traditional costumes. A slow generational drift away from pickled herring would erode volume and on a fixed-cost plant that means operating deleverage. Probability: medium. Severity: moderate because it is a slow grind rather than a cliff.

4. The consolidated Lisner/Müller competitor. The newly enlarged leader, backed by a €7 billion parent could fund a price war that SEKO can’t match. The counterweight is the CEO’s diversification-ceiling argument that buyers will protect a number two. Probability: low. Severity: severe if it happens.

5. Controlling-family and minority risk. With 62% family control and a ~12% float, minorities have little leverage and the permanent-impairment path for an outside holder is a cheap take-private before the value is ever realized. This is also the reason a careful person discounts that bond portfolio since the cash may never actually come out to minorities. The counter argument I see to this is the behaviour so far in the modest fixed pay, ~50% payout and no abnormal related-party dealings. Probability: low. Severity: moderate because it caps your upside rather than destroying the business.

6. Single-site, food-safety catastrophe. There is only one plant. A serious fire, a contamination or something could halt production and damage the brand and the chain relationships. The fact they have a clean operating history and a modern plant help. Probability: low. Severity: severe.

Plain cyclical earnings dips like 2021 do not make this list. They’re the price of admission to a fish business and the balance sheet is built to absorb them.

Conclusion

SEKO is a small and boringly competent Polish herring processor, the clear number two in its niche, run for thirty years by a founding family that owns nearly two-thirds of it, pays itself a pittance in pure fixed salary, carries no goodwill and a net cash position once you count its bond hoard. It also hands shareholders half the profit every year it can. It has no pricing power and it never will, because the discount chains own that part. Its margins are thin and 2021 proved they can briefly vanish. It will not compound your money at 15%.

The part that matters is that the 40% gap to book value is worth wildly different amounts depending on how fast it arrives. The same gap is a spectacular return if it closes in two years and a mediocre one if it closes in ten.

If the gap never closes and the cash stays trapped, you’re left with the ~6.6% dividend plus a tiny amount of growth. Call it an 8-9% IRR. Notice that even this floor beats the 5.7% Polish bond. That’s the worst case, and the worst case is still acceptable since it doesn’t lose you money (not counting opportunity cost).

Base case, the gap closes gradually over four to six years while dividends creep up: something like a 12-15% IRR.

The good case, a real catalyst arrives fast (a buyout at a premium, the family or the Warsaw exchange push harder on distributions, or the market simply noticing it): the 40% gets packed into a short window like 2 years and the IRR goes north of 20%.

One thing I deliberately left out of all three numbers is the consolidation tailwind I described earlier. I assumed essentially no earnings growth, because reinvestment here is roughly value-neutral. But if the Lisner-Graal merger really does push the chains to hand more volume to their credible number two, that grows SEKO’s tonnage and maybe its margin, which lifts the one leg of the return I set to nearly zero. I’m not putting a number on it because I don’t want to sell you a CEO’s thesis as if it were fact, but it sits there as a free bonus on top of the base case rather than something you’re paying for today.

So the honest shape of this investment is a high base case of around 8% and a stack of free options on top that don’t cost you anything, because you’re buying the whole thing at 0.70x book. You’re paying for the base case and getting the upside thrown in.

Will the gap actually close? I don’t know. Which is exactly why I mentioned in the beginning this would sit in a basket of statistically cheap stocks.

Disclaimer: I am long SEKO and this is not financial advice, just my opinion. I’m still a new-to-the-game investor learning every day and on top of that I don’t understand Polish, so I trusted a machine translating the filings. Please do your own due diligence.

As with every post, thank you very much for reading this far. If you have any critique, advice or just want to say hello, please do it in the comments. And if you spot any grammar mistakes, please point them out, English is still not my main language.

Best regards,
Tiago

The solution came from my girlfriend, who suggested I should build a recurring segment for exactly these weeks and I liked it immediately.

It’s called Small & Inevitable.

The idea is simple: take a company that was once a micro-cap nobody followed and ask what a careful analyst could have seen at the time. Which signals were already present in the filings? What was the market missing and why? What does the case teach about finding the next one?

The first one we are going to talk about is Constellation Software. Everybody knows it so I think that’s a great starting point for this segment.
Constellation Software listed in Toronto in 2006 at roughly C$345M (about US$275M which was ~1.7x sales). Today it is worth about ~C$57B… That’s why everyone knows it and why we will be analyzing it.

The business in 2007

Exactly like they do today, at that time it was the same. Constellation buys vertical market software companies and then, contrary to what is normal in private equity, just keeps them.

In essence, vertical market software means software written for one narrow industry and useless to everyone else. Imagine things like scheduling systems for public-transit agencies, billing software for water utilities, management systems for golf clubs and funeral homes. These are some examples of what they buy. Each product is very boring but very important to a few hundred or few thousand customers. It is also completely irrelevant to anybody outside that niche which is the whole point. The niches are too small for Microsoft or Oracle to bother with, so nobody shows up to compete the returns away.

The money comes in three forms: upfront licenses, recurring maintenance contracts and services. Maintenance is the prize because it renews every single year and the customer almost never leaves (more on that below).

By the May 2006 IPO they had bought 45 of these little businesses and were running about US$165M of revenue across 40 offices.
One detail that mattered enormously and that almost nobody noticed is that the IPO raised no new money for the company. It was purely a liquidity event for an early private holder. There was no fresh capital and no sponsor with any incentive to go around hyping up the stock. This means that constellation basically listed in silence.

At the same time, the timing of the listing was terrible. This was 2006–2008 which was the Web 2.0 era. Every dollar of capital wanted consumer internet or high-growth horizontal SaaS. A Canadian acquirer of boring single-industry legacy software and growing organically in the low single digits was the precise opposite of what the market wanted to own.

Which, of course, is exactly why it was available.

The signals that were already on the page

They had recurring and sticky revenue which was the centre of everything.

Maintenance is contractual and renews annually. At the same time, in niches where that software is vital, customers basically don’t switch. Why would a water utility risk ripping out the software that runs its billing to save a rounding error? Mark Leonard (the founder) said it plainly in his letters (which I recommend you go read and then come back to this article, his words are greater than mine): he treated net maintenance revenue as one of the best single indicators of the intrinsic value of a software business. And that value kept climbing, net maintenance revenue hit a record US$37.8M in Q4 2007, up 28% on the year before.

The whole business was also very capital-light. Software needs almost no physical capital. The real “capital” being deployed wasn’t in factories, it was money spent in acquisitions and it earned high returns. Leonard’s own definition of ROIC was Adjusted Net Income over average invested capital and in these early years the redeployment was already running at rates most businesses never see. The public-sector segment alone was earning roughly 25% on invested capital in late 2007. For someone who is reading my posts trying to read on some interesting stock ideas, I don’t have to tell you how great 25% ROIC is.

The business was also run with an owner-operator with real skin in the game. Mark Leonard ran it and held a meaningful stake. (As an aside, if you want to know how the man was even like the answer is: famously and almost aggressively private, a tall 1.95m guy with a beard somewhere between a wizard and a hermit that gives no interviews. I find that oddly reassuring in a CEO.)

More important than his beard: managers were required to take a large chunk of their cash bonuses and buy CSI shares on the open market. This means they thought like owners because they literally were buying in alongside you. The company never raised equity again. The share count stayed around 21.2M and stock-based compensation was essentially zero. All the free cash went back into buying more software companies. There was also no dividend being paid. This was a team compounding capital, not a team extracting it.

On top of all of this, his discipline was already on the public record too. In a late-1990s newsletter (”Opportunity Above All”), Leonard reckoned a software business growing revenues in the high teens at low-20s margins was worth around 1.65x revenue before an illiquidity discount. He capped what he’d actually pay in cash well below that, around 1.25x revenue for the best ones. You could read, in his own words, that this was a man who knew exactly what he was buying and refused to overpay for it.

Growth was funded out of internal cash flow and a bit of borrowing, nothing crazy. After the IPO, the acquisitions were financed from earnings (around US$73M) plus a modest amount of borrowings and cash (around US$26M), with zero new shares issued. As late as early 2008, the revolving credit line was still only about US$105M. This company walked into the financial crisis from a position of strength on it’s balance sheet.

And the big sign that one had to pick up on was the FCF conversion that the income statement was actively hiding. This is the tell that made the whole thing cheap.

When you buy 45 software companies, accounting forces you to record a mountain of intangible assets and then amortize them. This creates a non-cash charge that gets subtracted from earnings every year even though no cash leaves the company. That charge crushes GAAP net income. So Constellation reported its own number, “Adjusted Net Income”, which adds that non-cash amortization back and told shareholders directly that ANI was generally a better measure of cash flow than GAAP net income. Leonard even pointed out that GAAP does a poor job with goodwill and intangibles since it implicitly assumes the asset’s economic life eventually runs out, when for his businesses that life was often closer to perpetual.

Translation for us: the headline GAAP P/E made the company look more expensive than it actually was, because the “E” was being artificially suppressed by a charge that wasn’t real cash. The cheapness lived precisely in that gap.

Want a feel for the size of it? Quarterly ANI went from US$4.1M in Q4 2005 to US$9.0M in Q4 2006 to US$12M by Q2 2008. Annualise the Q2 2008 figure and you’re at roughly US$48M of cash earnings against a company that had listed at about US$275M. It was a cash earnings yield somewhere in the mid-teens and that’s before the 2008 panic let you buy it cheaper still. (Quarters are lumpy, so treat that as a sketch, not a spreadsheet.) Against a Government of Canada long bond yielding around 4% at the time, a mid-teens cash yield on a business growing maintenance revenue 28% a year was a clear mispricing.

There was one more signal, and it was the most important of all. The runway of reinvestment was very long. The pool of tiny vertical market software companies was enormous, fragmented and had no natural buyer because private equity wanted bigger targets. And the funnel was visibly widening in real time: in 2007 the company signed roughly 50% more non-disclosure agreements than in 2006, with the value of letters of intent up something like 52–65%. This turned out to be a decades-long global runway.

What the market was missing

The story the market told itself went something like this: it was a boring acquirer of legacy software, with ~1–3% organic growth, run by a recluse who won’t talk to anyone, trading on a thin GAAP P/E that makes it look pricey, listed through a no-raise IPO that nobody had any reason to promote. And the word “acquirer” carried real scar tissue because the previous decade was littered with serial acquirers that piled on debt, played accounting games and then blew up.

The market was so unconvinced that at one point there was a reported short position of more than 30% of the shares. Leonard wrote that he was, at first, more amused than annoyed. A third of the float was betting against a business compounding maintenance revenue at nearly 30%.

Some of the concerns were perfectly fair. Organic growth genuinely was low. Key-man risk was real. The construction and housing-related markets were getting hammered in 2007–08 as the US property market rolled over. And the model did depend on cheap deal flow continuing.

But the central worries were checkable, and a careful reader could have checked them at the time:

The acquirers that blew up overpaid and over-leveraged for synergies they never got. Constellation paid no more than ~1.25x revenue in cash, used very little debt, didn’t gut the companies it bought, left them autonomous and measured each one on ROIC. The filings let you tell a disciplined acquirer apart from a financial-engineering one if you bothered to read them.

They also said that the low organic growth was a melting ice cube. It was the wrong lens entirely. The value was never going to come from the existing businesses growing, it was going to come from taking their cash and redeploying it into more businesses at 20–30% returns.

One of the biggest fears was the saying “it’ll run out of targets”. The accelerating NDA and letters of interest counts said the exact opposite. The funnel was getting wider, not narrower. Now, I must admit that this one comes more in hindsight than anything else. If I was looking at it at the time, I would probably also think the runway was way smaller than what it actually was. But that’s the reason we are analyzing these cases and learning. We should be conservative and skeptical but also look to make a deeper analysis before judging it.

And finally the “expensive on P/E.” As we saw it was only on GAAP. On cash earnings it was cheap, because the amortization crushing reported earnings was non-cash and attached to intangibles that weren’t actually eroding. The company told you this in writing.

The lesson for today

This is the part I actually care about, because the point of these look-backs isn’t to create nostalgia or regret of not seeing it then (even though I was only 7 years old at the time), it’s to build a template that me (and you) can use on the next one we come across.

1. When acquisition amortization is large, throw out the GAAP P/E and rebuild cash earnings yourself. A serial acquirer of intangible-heavy businesses will always look expensive on reported earnings and cheap on cash. The mispricing lived in that gap and it still occurs today. So when you’re hunting for ideas, for companies where amortization of acquired intangibles is large relative to net income, do the owner-cash-earnings math by hand. When management hands you an “Adjusted” number and explains why it’s closer to real cash, like Constellation did, that’s not automatically a red flag for accounting games. Sometimes it’s a gift and your job is to check which one it is.

2. The compounding came from reinvestment rate × return on reinvestment and not from organic growth. When looking at Constellation at the time, one shouldn’t have been thinking “how fast does the core grow?”. The correct approach was “what share of free cash flow can be redeployed, at what return, into how large a runway?”. High ROIC plus a long fragmented runway, plus a manager who reinvests instead of distributing is a very powerful engine. A low organic growth rate does not disqualify a business that can put nearly all of its cash back to work at 25%. If anything, the low organic growth is part of why it was cheap enough to buy.

3. A discredited category is where the edge actually lives. An acquirer like Constellation was basically a slur in 2007 so the market threw a disciplined and cash-funded compounder right next to the frauds and walked away. The edge here was doing the work and go see the cash versus leverage, autonomy versus forced synergies, ROIC discipline versus growth-at-any-price and an owner-operator with skin in the game versus an empire-builder.

As always, this is just my opinion, not financial advice and these look-backs are educational by design. I don’t own a time machine and I obviously wasn’t reading Canadian software filings as a kid in 2007. If I’ve botched a number or you read one of his letters and understood it differently than I did, please tell me in the comments. Half the reason I write these is to be corrected. Also, if you spot a grammar mistake, point it out too because English still isn’t my first language.

Thank you for reading this far.

Best regards,
Tiago

The Business

The company was founded in 2000, it’s based in Glasgow and is still run by one of it’s co-founders.
SpaceandPeople is a middleman. It doesn’t own shopping malls or shops in train stations. It own contracts. These contracts give them exclusive licenses from the actual owners of the spaces to sell, during short periods, the unused space for brands to do marketing and promotion of their products there.

The need for their service actually makes sense. Imagine you own a big shopping mall and you have something like 200 square meters of unused space that just sits empty and which you are not profiting on. You could try to sell that space yourself by hiring a salesperson, building a database of brand managers at Heineken or whatever, write the license contracts, send invoices and chase the people that don’t pay on time. Or, instead, you could hire someone like SpaceandPeople and they do all of that for you in exchange for a fee. In essence, that’s the business.

The group currently operates with three revenue segments:

UK Promotions: In 2025 it did £4.95m of net revenue. That’s a 21% growth over last year and represents about ~61% of total net revenue. Basically in this segment the company sells space and time (they might be in the aerospace business after all) to brands. Samsung was a big one in 2025 with 297 days present in a station, Rekorderlig installed a peach-shaped “cold sauna” at Broadgate (I had to look up all this places) and Nike did painted shipping containers in Manchester.

The real anchor of this segment is the rail network. They have systematically & successfully won contracts to put up these promotion spaces in almost every UK rail terminal. They used to be signed each year but in 2024 they signed it with exclusivity until 2029. I’m confident they will be able to do it again when the signing time comes again.
In 2025 they signed a new deal with London North Eastern Railway that added Newcastle, York and Doncaster on a three-year exclusive contract.
Together with Govia Thameslink, Northern, Southeastern, East Midlands and West Midlands (again, I had to search all of this, it sounds like Lord of the Rings), the company claims exclusive rights at “more than 1,600 stations representing over 1.6 billion passengers annually”. That’s a lot of eyes in which they can put products in front and charge for it.

I saw someone here on substack, on one of only short texts about the company here, saying they sold because the UK vape legislation coming out would hurt the business, but that’s not my opinion at all.
Vape companies make up only a tiny fraction of SpaceandPeople’s business. SpaceandPeople’s business relies on a massively diversified client base. Companies like Netflix (who famously rented major space at London King’s Cross and Edinburgh Waverley), Google and major auto manufacturers booking massive experiential advertising campaigns.
Telecom giants like Tesco Mobile and major energy utilities that use high-traffic kiosks to acquire new subscription customers.
Major retail and beauty names like LEGO, Space NK, Charlotte Tilbury, LUSH and Huda Beauty setting up physical pop-ups.

Sure, the few vape promotions that actually might be illegal to do now, but the next marginal buyer of space will pay a couple percentage points lower price and buy the time and space for their promotions. That small (at least that’s my view on it) loss will get diluted in the growth of other promotions.

UK Retail: Second segment with £0.56m net revenue in 2025, 7% growth over 2024. It has two sub-lines. The legacy “Retail Merchandising Unit” (RMU) business, older free-standing kiosks the Group rents out by the day, is being phased out. The growth product is “Rock Up and Pop Up” (RUPU) as they call it, which is a fully-designed, branded, end-to-end pop-up retail solution. The Group designs the kiosk, builds it (now at a new warehouse in Daventry), installs it, maintains it, places it in a premium centre and rents it short-term to a brand. They ended FY25 with 34 RUPU kiosks trading, up from 26 a year earlier, with brands including Thomas Sabo, Happy Socks, Lush, and Vieve.

German Retail: third segment with £2.52m gross revenue but only £0.99m net revenue, representing 16% growth over 2024. This is the legacy of the 2010 acquisition of Retail Profile Holdings Limited, which they paid £6.3m to acquire. The business is basically the same, mall kiosks placed inside ECE shopping centres (ECE is the largest German mall operator and the contract was renewed to 2029). At the end of 2025 the Group operated 130 kiosks in Germany (they had 118 in 2024). Note that this revenue is recognized gross (principal), unlike UK retail (agent), which is a piece of accounting plumbing that has caused much confusion historically. What this means is that basically, in Germany, the company buying the time and space from SpaceandPeople pays 1,000 and they register 1,000 in revenue, with the money the company actually has to pay the mall being registered as an expense. In the UK, only the fee enters as revenue, making it much simpler to analyze.

They then have their head office in Glasgow which consumes about £1.5m a year and houses the staff, IT, finance and group functions.

Revenue Model

The Group’s economics are dominated by the commission percentage and occupancy rates. They take a roughly 20-40% cut on UK agency revenue based on what I could understand studying the industry. It depends if it’s a Netflix presentation in a big space or a small demonstration of some retailer in a train station (in the German business, while it is calculated differently because of the revenue recognition model, after the expenses to the mall owners, the rate is approximately the same. 0.99m net revenue on 2.52m is about 40% rate which is pretty good ). They do not own the underlying real estate, so they have a very small hard asset base. They do own kiosks (£581k plant & equipment at year-end 2025, plus £490k right-of-use property, most of that is the new Daventry lease).

Cost Structure

The dominant cost is people. Of £6.28m administrative expenses in FY25, £4.31m (69%) is wages, salaries, social security and pensions for an average headcount of 66. Add £374k of depreciation and amortisation, £127k of office and warehouse rent and the remainder is selling cost, IT, audit and venue operating expense.

This is an asset-light, people-heavy business. Fixed costs dominate. Operating leverage is therefore high in both directions, a £1m swing in revenue at 30% incremental margin (which appears to be roughly the case in 2025) drops £300k to the bottom line. Equally, a £1m revenue decline takes most of operating profit with it.

Capital Intensity

This is the favorable side. Capex requirements are very modest: £435k in FY25 was elevated by the new Daventry warehouse fit-out. In a normal year £150–£250k looks like the right number. The company really has limited capital expenditure needs.

Industry Structure

Cyclicality

This is a discretionary-marketing business attached to physical retail and physical travel. It got crushed by COVID. Revenue collapsed by 72% in H1 2020 (net revenue £1.1m) and full-year 2020 produced a £2.1m loss after tax. Pre-COVID, in 2019, the Group had earned a £0.7m profit. That is the worst drawdown on record and it nearly killed the business, the Group survived by massively reducing working hours, receiving CBILS-backed loans of £1.6m (if you are wondering what they are, they mean Coronavirus Business Interruption Loan Scheme loans, which is a fancy way of saying the British government secured the loans for them) and aggressive cost-cutting (headcount fell from 99 in 2017 to 52 in 2022). The german subsidiary also received direct money injections from the government as subsidies.

Also, in 2018 there was a downcycle driven by the collapse of Carillion (which was a major venue partner) and the loss of an Indian joint venture. Revenue fell 13% and profit before taxes fell 80%.

The current Q1 2026 trading update flags “increasingly severe economic headwinds, including inflationary pressures, higher interest rates and cautious consumer sentiment” and “more conservative purchasing across the broader brand market in the first quarter of 2026.” This is the third meaningful downturn signal in eight years. The business is cyclical.

Regulation and Disruption Risk

There is very little and very light regulation. The principal disruption risk is digital substitution. Brands have been shifting marketing budgets toward digital and programmatic for fifteen years. The counterargument, that is repeated by management and actually supported by industry data is that experiential, sampling and pop-up retail are the anti-digital product, prized precisely because they offer physical human contact. The trend is real on both sides.

Threat of New Entrants

The main barrier is the exclusive venue contracts. Anyone wanting to enter UK shopping-centre commercialization must (a) build a sales team, which is cheap and (b) win contracts away from the players already settled, which is hard but possible at the tender renewal dates. JCDecaux is the elephant in the room. JCDecaux already holds advertising contracts at Westfield London and Westfield Stratford City (8.5-year deal won in 2018) and has the scale to bid for almost anything. JCDecaux does not currently focus on fixed promotional kiosks and pop-up retail, which is SpaceandPeople’s specific niche. JCDecaux sells advertising panels. But the line is very thin.

Industry ROIC

Structurally moderate. JCDecaux with €765m operating profit on ~€8 billion of invested capital is running at a mid-single-digit ROIC. The economic problem with out-of-home advertising at scale is the high concession fees paid to venue owners (airports, councils, transport authorities). The structurally favorable position is to be a small and asset-light agent, which is exactly what SpaceandPeople is.

Management and Capital Allocation

Nancy Cullen is the CEO and a co-founder. She went to Glasgow University as an accounting graduate and assumed the charge as CEO in 2020 when the co-founder Matthew Bending retired, right in the middle of the crisis. She owned 3.48% of the company at the end of 2025, which is smaller than 2024 as she has been selling stakes on the market a few times. Even so, the truth is she took a difficult business in a dark period and turned it around. She may have been selling for personal reasons, since the hard-work on the business is evident. She has sold when the price was 100p and 200p, so I don’t think it’s a matter of her view of the intrinsic value of the business, but maybe you view it in a different way.

The list of the biggest shareholders is as follows:

The last acquisition was the Retail Profile one already mentioned, in April 2010 for the German business. Goodwill from that acquisition (£8.2m gross) has been impaired by £2.8m in 2020 because of the COVID impact, leaving £5.4m on the balance sheet today which exceeds 100% of tangible net assets. There have been no more acquisitions since then.

Regarding dividends and buybacks, the latter was never done, but dividend were once paid. They were cancelled in 2020 because of COVID and haven’t been resumed since. The Chair has said (on FY25 results) that dividend resumption “is not expected in the near term and will remain dependent on the continued delivery of strong and consistent financial performance.” Distributable reserves remain negative at -£1.37m retained earnings as of 31 December 2025 so that’s something to consider.

In terms of reinvestment, there were £435k of PPE additions and £111k of intangible additions in FY25. As I said before most of the PPE addition was the new Daventry warehouse fit-out. Maintenance capex appears to be around £150-200k per year based on prior periods.

As for cash and debt, the CBILS term loans of £1.5m (taken during COVID) have been fully repaid by end of FY25. They have a net cash position of £1.64m at year-end and an overdraft facility of £1.0m available which sits undrawn. This is a clean balance sheet and we love it.

The capital allocation track record over the past five years is essentially “survive COVID, repay COVID debt, do not pay dividends, do not buy back stock, invest modestly in the operational platform.” That is defensible and understandable but unimpressive. The big capital allocation decision in 2010 for the Retail Profile acquisition is hard to evaluate cleanly. The German business contributed £161k of segment operating profit on £2.5m of revenue in FY25 (6.4% margin), which is below the Group average so that means something. After a £2.8m goodwill impairment, the acquisition has clearly disappointed against its original case, but it has not been disastrous.

If you judge it by the earnings line it’s hard to evaluate if because of everything we have covered so far. They have taken big hits but not all of them in cash.

But FCF paints a clearer picture of how the story has been unfolding over the years:

It’s not crazy beautiful, but it definitely shows a different picture of the business and one where I see potential.

Red Flags

- Goodwill: £5.4m against a £3.9m net asset base. It’s sensitive to discount rate (which has been rising) and to RUPU growth assumptions (which the Group is forecasting at 32% in 2027 and being honest, that’s a heroic figure). I would expect a partial impairment in any year that materially disappoints.

- Trade receivables loss allowance: £522k against gross trade debtors of £1.95m, translates to a 27% allowance. That is a high allowance ratio for a B2B business. The Group explains this as the consequence of historical bad debt experience with smaller promoters. It suggests the customer mix on the brand side includes a tail of small, unreliable counterparties which is not great.

- Aged debtors. £1.19m net of provision is overdue at year-end 2025 (up from £0.90m a year earlier). This means working capital quality is deteriorating, not improving and that’s probably because of the headwinds they mention in the latest filing.

- Negative retained earnings. -£1.37m at year-end 2025. This prevents dividend payment until the deficit is eliminated, which would require another 3-4 years of current-level profit, which is a lock on unlocking the value on the balance sheet.

Risks That Could Permanently Impair Capital

As I do with every write-up, I will list the risks I see and rank them subjectively on probability (in a very simplified way, as all things should be) and the severity.

The main risk is the loss of rail network contract at the 2029 re-tender. Losing this exclusive contract would remove the most profitable component of UK Promotions segment.On the probability I rank it low-to-medium. The company has held the contract since 2014, has just won the most recent re-tender competitively and has built a unique operational platform around it. A loss is possible but not the base case in my opinion. As for the severity it would be very severe but not existential. UK Promotions would shrink by perhaps 40-60%. The company could survive as a smaller business with the German operation and the remaining rail operator contracts.

Second risk I see is a goodwill impairment. An impairment of even half the £5.4m would wipe out reported earnings for years and further hurt distributable reserves.
The probability I see is medium. The discount rate has been rising and the RUPU growth assumptions (32% in 2027, 10% thereafter) are aggressive. It could definitely happen. The severity is non existent in cash terms (impairment is non-cash), but severe in market terms since the equity is small enough that £2-3m of impairment would be visible.

Third easy risk to assume is something like COVID again. A new disaster creating a deep cyclical downturn. COVID showed the business can lose 70% of net revenue in months and a recession of 2009-magnitude could halve the revenue.
The probability I would say is medium over the next decade, by definition there will be at least one cyclical downturn.
In terms of severity it would be severe indeed. The Group survived COVID only because of subsidies and CBILS-backed government lending. A future downturn without state support would be more dangerous if they don’t manage to get to a more secure position before it happens.

The last main risk I see is key-man risk. The CEO is a co-founder and while I couldn’t confirm her age, she should be in the 55-60 range, if not already surpassed it. Succession is not visible in the filings so that is definitely a risk.
Probability I would say is low over the next five years, but medium over the next ten. Severity in my opinion is medium. While she was the one who took the reigns and turned it around, the main business of venue relationships, contract negotiation experience and the German business links I think can probably be done by the current existing team.

Intrinsic Value Framework

The FY25 result is the strongest in eight years. It included an “exceptional” H1 according to the CEO. Q1 2026 is already softer so FY25 is not the right anchor, I would say.

The post-COVID history is short, only three years (2023, 2024, 2025). Operating profit over those three years averaged £376k. Free cash flow averaged £640k. That is the best estimate of a mid-cycle annual earnings number I can defend from the information we have.

I can only see a way of normalizing this which is I will be slightly more generous on the basis that the Group has structurally added the LNER contract (December 2025), the Gropius Passagen contract (Q4 2025), the Southeastern Trains addition (2024) and the operational scale-up at Daventry. And I will be slightly less generous on the basis that Q1 2026 is showing the cyclical headwind that always seems to arrive in this business.

For me normalized annual free cash flow is in the range of £500k to £700k on a steady-state basis. That is a wide range because the business is small enough that single contracts and single quarters move the needle a lot.

At the 16th May 2026 share price of 170p and 1,976,457 shares outstanding, market capitalization is approximately £3.35m. Net cash is £1.64m and lease liabilities are £619k. If you count the leases as debt, then EV is £2.3m.

At that point, it gives us a FCF yield on the range of 21.7-30% which is very much considerable. Against a UK 10-year gilt yield of 4.97%, the spread is enormous.
This is the headline number that makes the position visible. A normalized yield of 25-30% in a business that pays no dividend (yet) and trades 4,000 shares a day in a £3m market cap is the kind of yield that screams either opportunity or trap.

Conclusion

In my opinion the business trades at a significant discount to a conservative intrinsic value guess, but it trades there for some very justified reasons. It is very very cheap, and it deserves to be cheap, but this cheap?

There is a chance to make a lot of money on this company. An intelligent owner of this stock would have to believe three things to hold this at today’s price: that Network Rail re-tenders in the company’s favor again in 2029, that the goodwill stays unimpaired through the next cycle and that the Engage and RUPU platforms scale as management projects. None of these is unreasonable. None of them is certain. The position size, if taken, must be small enough that being wrong about all three at once is a curiosity rather than a wound. It’s not difficult for me with so little money, going through the work of all this research for such a small position on the portfolio because if I’m right and it pays off, it still moves the needle. For most of you, this is probably just a thought experiment. I don’t think I should remind you that this is not financial advice, just my opinion and please note that I’m long on the company (if what I said in the beginning is already forgotten).

Thank you for reading this if you got this far. As usual, I welcome every criticism, advice or just an hello. Also, if you see any grammar mistakes, please point them out to me. English is still not my main language since the last post.

Best regards,
Tiago

A chaebol (재벌) is a large, family-owned business conglomerate in South Korea, characterized by diversified subsidiaries controlled by a founder or their family.

A chaebol is a just a large corporation that owns many (in full or controlling parts of) other companies. Kind of a Berkshire Hathaway but run by prestigious families in Korea. How many of those exist over there, you might be wondering? With large-enough scale to be considered one, roughly 80. The top 10 Chaebols represent 80% of Korea’s GDP ( This surprised me. Ten families control 80% of the country’s output).

I’m not going to dive deeper here on Chaebols and their meaning in Korea’s society, but if you are interested in those sorts of things, I recommend you go and research/read about it, it was very interesting for me. The kind of impact, recognition and respect that the families controlling them receive in their society is astonishing. They are kind of like royalty over there, it’s very strange from an outsiders perspective (at least for me) but fascinating at the same time.

What this all means is that the company I’m going to talk about is controlled by one of those Chaebols: Hyundai Department Store Group.
This one originated from the main Hyundai Chaebol, when it was spun-off in 1999 but it’s not affiliated to it in any way now (except for the fact that the family controlling it is heavily related to the main family who controls Hyundai Motor Group. Super big families, controlling most of Korea’s businesses and therefore it’s society… kinda crazy. Sorry, I couldn’t control myself on not talking more about it).

The Chaebol own’s 50% of Hyundai Ezwel and then there are a multitude of lower-stake shareholders starting at 8% (owned by Tempered Investment Management Ltd, an investment firm based in Canada) and then immediately falling to lower than 2% for each one of the others. I estimate free float to be around 35%, but I might be slightly wrong on that. This is a big risk which will be explained, but the company is so incredibly cheap with such good economics that it’s still worth to look at it.

Let’s talk about the business.

Business Description

The business is simple to understand. While in the West companies reward their employees with general one size fits-all models (like gym memberships and dental care), over there a big part of the companies are starting to do something different. They give out points to employees that they can use however they like in a digital mall. Anything from books, travel, eyeglasses, health check-ups, gift-sets and so on. This means the employee that wants gym memberships can get one and the one who wants to go visit Lisbon (shameless Portugal plug) can also get what he/she wants. For the employer it’s also great since they receive a tax-deductible welfare expense.

Hyundai Ezwel is the company running the digital mall for more than 2,700 companies. Now, note that it doesn’t buy or resell goods. It just operates a platform. A corporate client signs a multi-year contract and pre-pays welfare points (this creates a sort of float by the way, which I will mention in the valuation part). When an employee buys a TV from Samsung on the Ezwel “mall”, the employee’s card is debited, the merchant ships the TV, the corporation is invoiced for the points used and Ezwel takes a comission spread between what the client pays and what Samsung receives for the TV.

The pricing model was harder to understand. The price is the “take-rate” on Gross Merchandise Value (GMV). In 2025, GMV was KRW 848.8 billion and revenue was 142.65 billion. This is an implied take-rate of ~17%. I found it high for a marketplace, but the reason for this is that Ezwel is not really a “marketplace”. It’s a closed system operating B2B2C where the corporate buyer pays up-front for the points and Ezwel earns margin on the products and on platform fees.

Welfare budgets in Korea grow with payrolls and with adoption of selective-welfare schemes. This selective-welfare market grew at a ~10% CAGR over the past five years and Ezwel’s management expects 7-8% growth going forward.

From what I read in the fillings, client’s switching is very rare, which would make sense. I imagine it would be hard and costly to switch provider. Your employees already know the platform, they have leftover points that you as the employer already payed for in advance, the whole billing system would have to be adapted to a new partner… It’s just a mess and so, even if the guy next door charges a take-rate of 1% less than the current one, it’s just not worth it. This is my interpretation of it.

There is substantial operating leverage for Ezwel. It comes from running more transactions on fixed tech and call-center cost.
Let’s look at it closer and the cost structure.
For 2025, on 142.65 billion of revenue, gross profit was 79.67 billion (~56% gross margin). SG&A was 55.48 billion yielding 24.2 billion of operating income (17% margin).

The largest costs are personnel (380 employees, mostly engineers and costumer-service staff), platform maintenance and marketing. These are predominantly fixed. So when volume of merchandise grows 6-8%, revenue grows 8-10% and operating profit grows 15-20%.
The 2025 numbers showed it: volume +2.1%, revenue +8.8%, OP +19.5%. This is real operating leverage.

Looking at the unit economics, with about 3.42 million employees using the platform, that equates to KRW 41,710 of annual revenue per user of the platform.
With 17% operating margin, that is KRW 7,090 of operating profit per user.
At KRW 6060/share and 22,477,903 shares outstanding, the business is being valued at 136.2 billion. Considering cash and short-term investments of 99.3 billion and total debt of 6.6 billion, that is an EV of 43.5 billion. That is an EV per active user of roughly KRW 12,700 when each of these users produces 7,090 of operating profit annually… More on this later.

The capital intensity is very low. On end-2025 total assets were ~ 241.7 billion of which roughly 100 billion is cash and short-term investments and another large chunk is operating receivables and prepayments tied to the welfare-points cycle.
Tangible operating assets are essentially the office in Seodaemun-gu, the call-center hardware, and the platform code,call it KRW 50–60 billion. With the business throwing off 24 billion of operating profit on that base, you can do the math yourself for ROIC.

There is no supplier concentration because it aggregates ~5,700 merchants. If one falls, there’s another one to replace it. Costumer concentration I’m not entirely sure. They mention flagship costumers like the Korean Police, Korail (the national rail operator) and KB Kookmin Bank. They also mention the public-sector client book is very significant and that tends to generate very long-term contracts. I would consider costumer concentration a risk, but a minor one.

The geographic concentration, just like Dino, is total. They only operate in Korea.

Industry Structure and Economics

Ezwel has done her own market sizing and estimates that Korea has over 20,000 companies with more than 100 employees. Of these, they expect that only about 25% of companies have shifted to the this selective-welfare model. But adoption has been climbing steadily because: (a) employees prefer choice and (b) the tax treatment is favorable when administered through a registered welfare provider.

The cyclicality is less than I expected when I first started looking at it. Welfare budgets are a function of payroll and salaries in Korea rarely decrease, this means the welfare bill keeps going up on average. In 2020, the year COVID compressed Korean GDP, Ezwel's merchandise volume actually grew, because public-sector clients were directed to disburse welfare points early to support consumer spending. I have not seen a year of revenue decline in the company’s history.

Regarding regulation, the selective-welfare framework rests on Korean tax law treating in-kind benefits as deductible to the employer and non-taxable to the employee within limits. Both have been in place for two decades and have been tightened, not loosened, over time, generally in ways that pushed more spending through registered platform operators (which is why Ezwel’s market grew). The risk vector here is a tax-policy reversal that turns these points into ordinary income to employees. It would not zero the business, because corporatations still want the administrative outsourcing, but it would compress GMV growth materially.

As for technology disruption I just can’t see any near-term existential threats.
The platform is software, but it is integration software deeply tied to Korean HR systems, Korean card rails and Korean tax-receipt formats. A pure software play (a SaaS startup, an LLM-driven shopping agent etc.) cannot realistically replicate the regulatory plumbing, the merchant network and the seven years of HR-integration history with a ministry client. The far-future risk is that Korean firms abandon points-based welfare entirely in favor of cash allowances. That is a culture-change risk on a 10–20-year clock, not a tech risk.

I would say the business has a moat but not a big one. There are basically three big-players (Ezwel ~50%, SK M&Service ~30% and e-Genadu ~15%) in the space who cover ~95% of the market as per Ezwel’s analysis. That is a stable oligopoly and take-rates have not collapsed. It is a B2B switching-cost moat with a side-helping of scale advantage in the merchant network. Take it source by source.

The only “moat-sources” I see are the switching costs with HR/payroll re-integration, card re-issue, employee re-education and 6–12-month implementation cycles.
It also has a modest cost-advantage by being the biggest player in the space, which means bigger operating leverage over the fixed costs, but that’s it.

Management and Capital Allocation

This is where I have the most reservations. Hyundai Ezwel is a controlled subsidiary with Hyundai GF Holdings owning 50.0%. The CEO is Park Jong-Sun, appointed by the parent itself. I couldn’t get much info on the man, but he used to run another company for the Chaebol.
There is no significant CEO equity stake of the kind we look for in an owner-operator. Also, compensation disclosure on KOSDAQ is less specific than other exchanges. The available filings indicate a fixed-plus-bonus structure with bonus tied to operating profit and revenue growth. Gameable metrics if one is unkind. There is no evidence of an owner mentality at the top. The owner mentality, such as it is, sits at the parent.

Regarding capital allocation, there was an acquisition of Vendys (paid KRW 37.1 billion in November 2022). Vendys is a small operator and it does the same model for restaurant meal vouchers, turning paper meal tickets into a phone-tap payment with around 3,300 corporate client locations as of Q4 2025. On entry, Vendys lost KRW 1.0 billion of operating income in 2023 and KRW 0.97 billion of net income. It reported a tiny KRW 0.04 billion OP in 2024 and only earned KRW 0.40 billion OP in 2025.

Goodwill on the deal was impaired by KRW 16 billion in 2023 and 6.3 billion in 2024 which is a clean admission that the price paid was high. By 2025 the trajectory turned positive (Q4 2025 alone delivered KRW 0.36 billion of OP). The first three years of this deal returned roughly nothing on KRW 37 billion deployed, and the cumulative goodwill impairment of KRW ~24.3 billion since 2023 is the company telling you the same. I will not dress this up: the Vendys deal was a mediocre acquisition that is now starting to work.

This impairment affected earnings, but not cash generation.

The company announced in early 2025 a 3% buyback for 2025 followed by 2% annual buybacks-and-cancellations every year from 2026 to 2028. At today’s price (~KRW 6,060 as I’m writing this), with EV implied around 32% of market cap, this is exactly the right use of capital.

The pace is constrained by chaebol convention, they will not get aggressive, but the direction is sound. At least the money is being distributed to shareholders.
The dividends also play a big role. Dividend payments have rose alongside the cash generated.

Dividends and Buybacks in 2025 totaled 12.32 billion, which on a current market cap of ~132 billion, that’s a return of about 9.3%, which will keep rising if cash generation keeps rising too.

Risks of Permanent Capital Impairment

As I do for all companies, I will list the bigger risks that I see and assign the probability and severity that I see on them. This is highly subjective and you will probably see them with different probabilities and severities. Possibly even more risks. Hopefully not an obvious one which I should have spotted out.

A tax-policy reversal that turns welfare points into ordinary employee income could very well permanently impair the company as I mentioned above. I see it as low probability but very severe if it happens.

There are also many risks related to the Chaebol. Minority shareholders systematically receive less than fair-value pricing on intra-group trade, M&A or equity issuance if it happens. The biggest risk I see is the attempt to take it private at a low-ball price. I would say it’s medium in probability and high in severity. On a 20y basis, I can definitely see it happening somewhere along the road.

A loss of a flagship costumer, or a number of them, could also pose a risk. But considering everything we saw before, I would give it a low probability score but a medium level of severity.

I can also see the possibility for the society as a whole to just prefer cash rewards instead of a point-based system. I would say it has a low probability in the next 10 years, but a bigger one in a span of 20 years. Severity would be high.

The last one I can imagine is disastrous M&A under chaebol direction, where the parent company directs Ezwel to pay a premium for what it might consider a strategic acquisition of an unrelated asset. The Vendys deal was a mild version of this. I can definitely see it happening again, but not on a disastrous level, given the company’s overcapitalization. I would give it a low probability score and medium severity.

Since the company’s revenues depend on salaries, if there aren’t any workers, there isn’t any welfare on salaries being paid. I don’t know how the future of AI & Robotics will progress. I can see a universe where in 30 years ~99% of the work has been automatized and humans either live a life full of leisure or oppression. I just don’t know if this is that universe. I would give it a medium probability and high severity.

For now, I would rank the chaebol-governance risk as the single most important one for a minority shareholder. Not because Hyundai Department Store has a particularly bad reputation among the Chaebol, from what I can see they don’t, but because any 50%-controlled Korean subsidiary carries a structural discount for a reason.

The Vendys acquisition is exhibit A: the parent had a strategic vision (build a total welfare platform), Ezwel paid the price and the minority took the goodwill impairment. That kind of thing tends to happen once per decade if you are unlucky.

Valuation Framework

First let’s establish normalized earnings power.

I take 2025 as roughly representative of mid-cycle, with adjustments. Operating income of KRW 24.2 billion. Net income of KRW 24.0 billion includes a KRW 2.64 billion Vendys-related tax refund, so cash earnings excluding that one-off would have been roughly KRW 21.4 billion. Strip out the remaining drag from the next-generation platform-build capex (running through H1 2026 per management’s discussion in the fillings), and a normalized through-the-cycle operating earnings figure I would estimate is in the range of KRW 22–25 billion of operating income and roughly KRW 20–25 billion of free cash flow at the consolidated level, before any growth from here. I believe this is very conservative. The business converts earnings to cash extremely well. Working capital needs are very small because of the float created by the prepayment model I mentioned earlier: corporate clients pay upfront for welfare points before employees spend them, meaning Ezwel is effectively running on other people’s money to fund its operations This means that every won of net earnings turns to cash, and often more than that.

With an EV of 43.5billion, that’s roughly a ~2.1x EV/FCF multiple taking the lower-end of that range of FCF.
There is something on the balance sheet which I left out on purpose. The company also owns stakes in other companies like KBank, HD Marine and Doosan which are valued at 14.1 billion. Depending on whether or not you wish to discount this, that multiple could be lower.

That is crazy cheap and the honest reading is that the market refuses to credit minority holders for the cash on the balance sheet, because under chaebol control the cash may never come out, it sits there, earns the Korean policy rate and gets occasionally redirected into deals like Vendys.

So the more useful comparison is at the equity level:

- Normalized FCF / Market Cap KRW 20bn / KRW 132bn ~15.15%

- Trailing dividend & buybacks / Market Cap KRW 12.3bn / KRW 132bn ~9.3%

9.3% shareholder yield on the equity level with the potential to grow at 8-10% for the next decade it’s not terrible when you pair it with the possibility that many other things can happen to unlock value. The pressure’s it is receiving from the KSE exchange can make it so that it has to ramp up the pace of cash distribution. There could be a buyout on the line with a decent premium.

Conclusion

I see all of it like this: It’s a very small company ( if we take it in USD terms, it’s a 90 million market cap company), operating in an oligopoly with decent margins, where the market it operates in is more likely than not to keep growing at a decent rate.
It has a stupid amount of cash on the balance sheet which most likely won’t get distributed BUT it might. And it might also get acquired by the parent at a nice premium (or a bad one, but I don’t see Korean culture doing that and wanting to spend years in litigation with minority shareholders, but yet again, I don’t know Korean culture very well). And that’s what it is. A business yielding shareholders 9,3% annually on the basis that one day it might hand you a nice amount of cash, with an underlying business that is quite profitable. That’s a nice bet to make for if you’re one to build a statistically cheap basket of securities and wait for the math to math it out.

Last disclaimer: I am long on Hyundai Ezwel, and please, for the love of god, do your own due dilligence. I’m just a new-to-the-game investor who is still learning a lot everyday and most of all, one that doesn’t understand Korean. I trusted the machine who translated the documents for me, you probably shouldn’t.

As with every post, I hope you enjoyed it and if you read it, thank you very much for your time and attention. If you have any critique or advice, please let me know in the comments. If you spot any grammatical error, please also let me know as English is still not my main language. (or just forgive me and move on).

Best regards,
Tiago

Lets get into it. The company is Dino Polska, based in Poland and founded by Tomasz Biernacki. This is an interesting guy. He didn’t show up to his own company IPO and has never sold a single share since then. More on him later.

Business Description

Dino runs a chain of small, identical grocery stores in the small towns and outskirts of Poland. Each store is roughly 400 squared meters, about a third the size of a Lidl, with a small parking lot and solar panels on the roof. Fresh products (meat, bread, dairy) are about 41% of sales, packaged grocery is 47% and “non-food” is roughly 12%.

They opened 345 new stores in 2025, ending the year with 3,003 locations and twelve distribution centers (two more are being built). The economic engine is unusual, most grocers lease, but Dino owns. Since 2010 the company has bought the land and built the building for practically every new store.

Once a store is built and stocked, the unit economics are simple: low gross margin of about 23%, low ticket size, high frequency of visits and no rent expense.
Over time, the depreciation runs out, the land appreciates with Polish nominal GDP and the store becomes essentially a perpetuity of fresh-food cashflow.

New stores start off with lower margins and ramp up over 2/3 years. This means the group’s margins are always depressed by the 11-13% of the network of stores that is brand new. That is the same dynamic Sam Walton enjoyed at Walmart in the 1970s and is genuinely rare.

Revenue grows from two sources and only two: new stores and what management calls like-for-like (LFL) growth in the existing base of stores. There is no e-commerce, no franchising and no international operations (yet). They do one thing and they do it well.

Their cost structure is also simple: cost of goods sold (~76% of revenues), wages ( the largest operational expenditure and the one currently compressing profits), energy (partly offset by their solar panels in every store) and depreciation. Fixed costs are modest because they don’t pay rent. The operating leverage is real, but somewhat fragile: at this margin level, every 1% of LFL growth above wage inflation drops to the bottom line; for every 1% below it is painful and that’s exactly what happened in 2024 and parts of 2025.

Capital intensity is high, there is no way around it. Total capex was ~2,1B PLN in 2025 against revenue of 33,6B, roughly 6,2% of revenue. By comparison, a See’s Candy’s type business runs closer to 1-2%.
Franchise or commodity? This is the most important question and the honest answer is mixed. Dino has real franchise value, in any given village it is often the only fresh-meat counter for 10 kilometers, but at the corporate level it is a price taker against Biedronka and Lidl.

Costumer concentration is zero and supplier concentration is moderate, but the company owns its own meat plant (Agro-Rydzyna, acquired progressively from 2003 to 2010). Geographical concentration is total, since it only operates in Poland.
I see some analysts inside and outside substack hinting at possible future geographical expansion to nearby countries with similar rural contexts as Poland, but management has never hinted at it, so that is something to not rely on.

Industry Structure and Economics

Polish food retail is a roughly PLN 350B+ market that has consolidated quickly over the last 15 years. Branded chains (Biedronka, Lidl and Dino) have been taking over the small local shops.

In 2025, Biedronka generated 108B PLN in revenue with ~3,900 stores, Lidl around 45B with ~930 stores and Dino 33B with 3,003 stores. There is another “big” player called Zabka, owned by CVC but they operate with a different model with convenience stores, around 12,300 of them, but much lower per-store revenue. Eurocash, Auchan, Kaufland, Aldi and Carrefour fill out the rest.

Cyclicality is usually something I address in my analysis, but Polish grocery is more inflation-sensitive than cycle-sensitive. Food inflation ran around 17%+ in 2023, fell to deflation in parts of 2024 (compounded by the reinstatement of the 5% VAT on food, per the management report) and ran at low single digits in 2025 and expected around 4% for 2026.

There is a single downturn in modern Polish grocery history, the old 2020 Covid days (seems like it was ages ago and last year at the same time), and Dino’s revenue grew 32% that year. The risk is not GDP, the risk is wage inflation surpassing food inflation, which is exactly what has been happening since 2024 — Polish wages rose 12-14% while Dino’s net prices were slightly negative (deflation), hence why EBITDA margin fell from 8,6% in 2023 to 7,9% in 2024 and again to 7,5% in 2025.

Regarding the regulatory environment, three items matter:

• Sunday trading is banned for most weeks of the year. This is already in force and I don’t expect it to get worse.

• The retail sales tax was reinstated in 2021 and is paid on a progressive scale. Dino paid 445.4 million in 2025 versus 387.3 million in 2024. This is essentially a permanent sales tax of 1.3-1.4% on the big chains, which is not gigantic, but it hurts the bottom line

• VAT on food returned to 5% in 2024.

None of these are existencial.

As for potential technological disruption, online grocery in Poland is small — about 5% of food retail and it’s concentrated in major cities.
For a basket of the typical items one would buy from a grocery store, the unit economics of online delivery don’t work for a 1,500-person village in some rural area. This is a form of genuine protection, although I would not count on it forever.

There is no customer power and the supplier power is very small. There’s not much that many tiny, unorganized producers can do versus a big 3,003-store buyer.

The barriers to entry are easy to understand. They are physical.
To replicate Dino today in Poland one would need to: (a) acquire land parcels in towns of 1,500-10,000 people, (b) build or hire all the logistics to be able to supply the stores and (c) accept many years of negative cashflow. No rational competitor will do this. Biedronka and Lidl however, do compete with Dino. They are progressively opening larger stores in the 10,000-25,000-people towns where Dino was the only modern grocer.

The industry ROIC has been pretty great in Poland for the last 10 years. These stores have been able to capitalize on the consolidation gains of the traditional mom-and-dad stores.
On top of that, polish wages have been lower than the rest of western europe on average and real estate was cheap in the outskirts of towns and cities. In my view, I don’t see how this won’t fall by a few percentage points in the medium-long run.

Competitive Position and Economic Moat

Let me be honest here. There was a time when one could say the moat around Dino’s profits were higher. Those days have gone by and the moat is not widening, in fact, in my opinion it is shrinking.

There is no pricing power since the company is a price-taker. Even though food inflation was 3,3% in 2024, Dino had to run negative price increases. Real pricing power would mean one could pass inflation costs to consumers and they did not.

Switching costs are non existing. Since discount grocers are price takers, the consumer will just choose the cheapest and closes one to him or herself. Even so, if the switching cost is a 5min drive to the nearest Biedronka, that’s not a switching cost on my books.

There are some cost advantages present and this is the real moat, even though it’s narrowing. By paying zero rent, at maturity that is a 3 to 5 percent gain versus a competitor paying a lease. Included in these cost advantages are the fact they have in-house construction through Krot-Invest, which was also founded by Biernacki in 2013 to exclusively build Dino stores. It’s supposed to speed up construction times and reduce costs. Meat-processing is also done in-house through Agro-Rydzyna. But as we saw before in the EBITDA section, these cost advantages are being eaten by wage inflation, and Biedronka and Lidl have bigger scale to absorb it.

The brand has a name to it, but competitors have bigger names in the heads of polish consumers. I don’t consider this a structural advantage.

Network effects are absent.
The company does however benefit from scale effects on a village level. A 3000-person village will not support two modern grocers. Whoever buys the land, builds the building and stocks the store first, wins that cashflow in “perpetuity”.

Management and Capital Allocation

This is the section Dino looks best.
Tomasz Biernacki, the founder, is 53 years old and Chairman of the Supervisory Board. He owns roughly ~51% of the company — roughly a 4,4B USD stake.

He hasn’t sold a single share since the IPO in 2017. I read that a former executive said that Tomasz himself chose and approved even the smallest details, like the cheapest garbage can supplier.

The CEO position has been vacant since 2021 when the previous CEO left for personal reasons. Even after all this time, the company still hasn’t appointed a new one. Tomasz is effectively CEO. It’s legal under polish law but not the standard I would prefer. There is no clear succession plan. If Tomasz was to get hit by a bus tomorrow (knocking on wood 3 times), I don’t know who would run the company.

Regarding capital allocation, there has only been one material deal, in 2024 — eZebra — but it’s small (the revenue it produces is less than 1% of total revenue), not transformative and the rationale is not obvious to me. I will skip this.

There have been zero dividends and zero share buybacks. Which makes sense, considering the company has had consistent 20%+ Return on Equity (this fell below 20% for the first time in 2025)

Organic reinvestment is essentially the entire capital allocation story. Capex was ~2,1B against cash from operations of 2,7B. That leaves out 600Million of FCF.
The capex figure has risen consistently with the increase in operating cashflow.

Net debt at year end 2025 was 200 million against equity of 8,6Billion. To put it into perspective, that’s roughly one-eighth of one year’s earnings. I think that’s more than fine, debt has been used as a temporary working capital tool, not as leverage.

Some owner-economic metrics:

FCF has been erratic, as previously stated, management prioritizes reinvestment to open more stores.

But there is an interesting metric to look at — Net Income per square meter of selling space.
In 2020 it was 640million of Net Income by roughly 540,000 square meters. This equates to around 1185 PLN per square meter. In 2025, with Net Income of 1,54B by ~1,201,200 square meters is 1282 PLN per square meter. This means profitability per square meter built and operated has increased 8% in five years which is not bad.

Risks That Could Permanently Impair Intrinsic Value

The way I analyze risk is: I enumerate them in my head and rank them by the probability I can imagine and picture mentally and by the severity of that risk. I know this is highly subjective in a way, but please bear with me.

The main risk I see in the foreseeable future is saturation of the polish market. There is not a lot of space left (even though management believes there is). With the competition raging on with Biedronka and Lidl, every new store location will probably be fought between them, because there is winner-take-all dynamics in most of the small towns. After saturation, growth capex at 18-20% return on capital stops. I see this as high probability in 10 years and high severity.
Some argue that they can expand internationally — while this may very well be true — management has not hinted at it and so I wouldn’t count on it. After all, that’s a form of margin of safety as well.

The second main risk I see developing is continuing margin compression from wage inflation and further price wars between the discount grocers. If steady-state EBITDA margin settles at 6%, the per-store and per-share owner economics are significantly worse than what the track record hints at. I see this with high probability and already materializing. Level of severity is moderate.

Third risk is key-man risk on Tomasz. If the proverbial bus comes for him, I do not know how this company looks without him. The rest of the management has no significant skin in the game and no succession plan announced. Low to medium probability over a 15-20 years holding period. Severity is high.

Fourth risk is regulatory action against big-chain grocers, including Dino. I can see the sales tax being raised in the future. I see it with a medium probability over a 10 year period and low-to-medium severity.

Fifth risk is currency exposure. This is a non-issue to polish readers but to us outsiders, it poses a real risk that one has to consider. PLN has appreciated vs. the EURO (my currency) in recent years, but a 25% drop is well within the realm of possibilities.

Intrinsic Value Framework

The way I view intrinsic value “calculations” tends to be different than most people and it keeps mutating a little bit from time to time. But the essence of it is that I don’t use spreadsheets as I see that as creating a false sense of precision.
As Buffett stated multiple times, if it requires you to look at it to the decimal place, it’s not fat enough. I tend to think alike. If napkin math is not enough to justify investing, there’s not enough margin of safety on it.

I look at what I could extract from the business if I owned it in it’s entirety (this only works if management is shareholder friendly, or you have enough capital to take a position in the company big enough to enforce decisions). Then I compare that to the price the market is offering me that business, and determine if the return is sufficient. This depends greatly on what my opportunity cost is at any given time. If I can deploy capital in another opportunity yielding a bigger return, that becomes my cost of opportunity.

At the time of writing, my opportunity cost is roughly 18%, so let’s save that for the rest of the analysis.

Let’s look at Dino’s normalized earning power. FCF in 2025 was ~610million. That number is misleading since most of that capex is directed for growth. If we assume maintenance capex to be somewhat around depreciation & amortization , we can estimate it around 500-600 million on total capex of ~2,09B.
If we ask: what would FCF be if Dino stopped growing, freezing store count at 3,003, the answer is roughly 2025 operating cashflow of 2,7B minus 500-600 million of maintenance capex, equating to around 2,1-2,2B PLN.

That is the right “earnings power” anchor. Today’s business run as a perpetuity would throw off around 2,1B in cash.
I will pause here, because I need to state something: the maintenance capex is most likely understated in this form. A more skeptical analyst, also applying a bigger margin of safety, would say that to maintain 3 thousand stores, distribution centers, truck fleets, solar panels and a meat plant, maintenance capex is closer to 800 million. This would bring the cash we could take out from the business to around 1,9B (and let’s not consider the fact margins could decline further).

How much does Dino cost to acquire? After the stock split, it has 980,4million shares outstanding with a price at the time of writing of about 32PLN. This means the market cap is roughly 31,3Billion. If we add 200 million of net debt it goes to 31,5billion.

So what’s the return we would get if it stopped growing? 1,9Billion on a purchase price of 31,5B gives a return of ~6%. The Polish 10-year bond is paying ~5,7%. This means that at this price you’re being paid 6% by Dino when the government will pay you almost the same for lending it money for ten years. One’s safer than the other. That’s not what I would call a great deal, and I don’t think anyone reading would either.

Some will say, what about growth?
Indeed, let’s talk about it. If management is able to deploy ~1,4B of capex for growth projects at a pre-tax return on capital of 20% for the next six years (reaching about 5k stores, the number everyone on substack bases the analysis on), that capex earns roughly 280 million of incremental operating income. At a 19% effective tax-rate, that’s ~227million flowing to the bottom line, compouding each year. Over six years, steady-state FCF that an owner could extract would grow from 1,9B to around 3-3,5B. That’s a yield on cost of ~9,5-11,1%. Not great either, discounted to the present is even worse.

After that point, one has to accept the business matures and grows roughly in line with Poland’s GDP growth, or one goes to the closet, whips out the cristal ball and starts projecting international expansion, margins and growth. That’s out of my league.

Even the assumed growth and steady-state FCF I’m projecting is contingent on a number of things going right: (a) Dino is able to reinvest capital at great returns like in the past; (b) Biedronka and Lidl don’t crowd the rest of the geography available and (c) that margins stabilize towards 7%+ and not 6%.

Closing Summary

This is a fine merchant business run by a great operator, but selling at a price that is simply not attractive. I might very well be proven wrong, but right now this goes into the “great business, bad price” pile.

The company will announce 2026 Q1 results on May 14th. I’ll be looking.
If it the stock falls to the ~25PLN range, it becomes more interesting but still not great compared to my opportunity cost.

This type of post will be rare. I don’t think it adds much to my audience to “pitch” companies that I think are overvalued, but my contrarian gene forced me to go against the consensus on here. Maybe you learned something new and that’s better than nothing. If not, refer to the line below:

If you have any criticism, suggestions or notice any grammatical errors (english is not my first language), please comment below and let me know.

Regards,
Tiago