Constellation Software has been one of the most successful compounders of the last decade. But every business eventually struggles with the law of large numbers, and Constellation is no exception. Mark Leonard himself has acknowledged that, from all-time high prices, future returns would likely settle around 8%. Even after the recent drawdown improves that math, the reality is that a $40 billion company simply can’t deploy capital the way a billion-dollar company can.

That’s what led me down a rabbit hole over the past few weeks. If the Constellation playbook works — and two decades of evidence say it does — the highest-return way to participate might not be through the mothership itself, but through its smaller offspring: spinoffs and subsidiaries running the exact same strategy at a fraction of the scale.

Today, we’re profiling four of them. Two are well-known members of the Constellation family — Topicus and Lumine Group. Two are lesser-known names being reshaped by Constellation’s people and culture — Sygnity and Asseco Poland. Each offers a different risk-reward profile, so I hope there will be at least one company you find interesting today!

Let’s dive in!

A company that’s a tenth the size of Constellation, by contrast, has decades of runway before the math of scale becomes a constraint. So the thesis is that you get most of Constellation’s advantages — the playbook, the incentive system, and in many cases the direct oversight — without the drag of deploying billions of dollars.

All three dynamics play to TSS’s strengths. It operates in 26 countries with decentralized teams that understand how to play in each region, it has Constellation’s capital behind it, and its permanent-home philosophy resonates with founders in a way that private equity never will.

One of the most important things in every programmatic acquirer’s playbook is the incentive system. Topicus/TSS mirrors Constellation’s. The two key drivers for management compensation are Returns on Invested Capital (ROIC) and net revenue growth. You need both because incentivizing only top-line growth degrades acquisition quality, while incentivizing only returns discourages capital deployment.

On top of that, 75% of after-tax bonuses must be used to purchase Topicus shares on the open market, held in escrow for a minimum of four years. Beyond that, there are no stock options or stock grants.

On valuation, I built a model around the same framework we used for Constellation: FCF2S (Free Cash Flow available to Shareholders) growth equals ROIC multiplied by reinvestment rate, plus organic growth.

I don’t want to make it too boring by going through all the assumptions in great detail. The main balance companies like Topicus, as well as all other companies we look at today, have to strike is between their ROIC hurdle and their reinvestment rate.

Obviously, you want to invest as much money as possible at the highest possible returns. However, those deals are difficult to find, so the higher the reinvestment rate, the more likely it is that ROIC drops. Especially for larger companies that can’t go as niche as some smaller ones.

Given the magnetism effect we discussed during our Constellation episode, I expect all companies to primarily defend their return hurdle rates. Thus, I keep ROICs relatively high and mostly play with the reinvestment rate assumptions.

Bear case: 15% ROIC, 60% reinvestment rate, 2% organic growth. At a 20x exit multiple in year ten, the implied return is about 6–7%. Not bad for a bear case, but it could mean another 20–30% downside in the near term.

Base case: 20% ROIC, 75% reinvestment, 4% organic growth. FCF2S per share compounds at nearly 20%. At 22x exit, that implies a mid-teens annual return over ten years.

Bull case (fairly unlikely): 25% ROIC, 90% reinvestment, 4% organic growth. In this case, the ten-year CAGR approaches 25%. As mentioned above, I consider this scenario to be rather unlikely since the 90% reinvestment rate and a 25% ROIC are somewhat at odds with each other — the more capital you deploy, the harder it is to maintain high returns.

If you assume a 25% ROIC, but only a 75% reinvestment rate, you would get to a CAGR of 20%.

That lumpiness shows up in the organic growth figures, too. Lumine’s organic growth is weaker than Constellation’s or Topicus’s, and there are years when it turns negative. The reason for that lies, once again, in the carve-out mechanics.

When Lumine acquires a carved-out division, that business has typically been cross-subsidized by its parent for years — sharing IT systems, back-office staff, and sales teams. Lumine has to rebuild all of that from scratch while simultaneously cleaning house on unprofitable customer relationships and low-margin revenue streams. The WideOrbit acquisition in particular, which was nearly the size of the rest of Lumine’s portfolio, caused significant organic drag in 2024.

On the returns side, Lumine’s ROIC since going public has averaged roughly 30% — materially higher than Constellation or Topicus in recent years. That’s largely because the carve-out model allows Lumine to buy businesses cheaply and then expand margins significantly, since these divisions were typically run inefficiently and undermonetized under their previous corporate parents.

I wouldn’t expect 30% ROIC to hold going forward, but 20-25% feels like a reasonable target for the years ahead, which is exactly where Constellation operated when it was Lumine’s size.

Since Lumine fits right into the Constellation universe with a focus on FCF2S, we use a similar model to Topicus. The assumptions are similar, with the difference of a lower organic growth rate, slightly lower exit multiples due to the lack of a similar track record as Topicus, but higher average reinvestment rates.

Bear case: 15% ROIC, 60% reinvestment, 1% organic growth. At 18x exit, the implied return is approximately 7%.

Base case: 20% ROIC, 80% reinvestment, 2% organic growth. Cash flow compounds at 18%, implying roughly 16% annual returns at a 20x exit.

Bull case: 22% ROIC, 100% reinvestment, 2% organic growth. The cash flow CAGR hits 24%, and at a 22x exit multiple, annual returns approach 23%.

If you asked me to do the type of consumer research that Peter Lynch describes in One Up On Wall Street, I’d just look down at the apps I use most (which I’ve done a few times before), and so that’s why I’m a bit embarassed to admit that we haven’t even looked at the business behind the app I use most: Spotify.

(Almost) Everyone agrees Spotify is a great product, but investors can’t decide if it’s a great business because Spotify isn’t a typical software company where gross margins can expand dramatically as revenue scales.

Instead, it’s an audio platform built on licensed content, and the suppliers who own that content are about as powerful as suppliers get. Meaning that, for the entirety of Spotify’s existence, the company has faced capped margins, yielding rather modest profit margins for a “tech giant”.

Lately, though, the story has started to change. Operating margins have finally inflected upward in a real way.

So, as always, let’s build a mental model of what Spotify is, why it exists, where value accrues, and what needs to be true for the stock to be attractive from here.

Spotify’s value proposition to users is obvious to anyone who has ever used it. The company took the world’s catalog of music and put it at your fingertips, available to stream as much as you want, whenever you want, either for free with ads or for a modest subscription fee that is still, in my opinion, one of the best-value subscriptions for entertainment on the internet.

What’s less obvious, and what drives the entire investment debate, is that the music industry is a market with a significant amount of “supplier power.” While Amazon, Uber, and Airbnb run marketplaces where they tower over the individuals and small-to-medium-sized businesses that supply their platforms (e-commerce brands, drivers, and hosts, respectively), Spotify doesn’t tower over its supply partners.

In fact, the music industry is a market dominated by a small group of music labels that control most of the world’s rights, as we learned when we studied Universal Music Group a few months ago, and those labels have had the leverage to dampen Spotify’s earnings power ever since the company’s inception.

This is why the popular narrative that “Spotify exploits artists” is inherently incomplete. There’s a separate debate about how labels split royalties with artists and what “fair” should mean, but Spotify isn’t operating against the industry; it’s intertwined with it and always has been.

On the surface, there’s a good argument for music streaming being a commodity. Press play, and you hear a song, with streaming platforms effectively all having the same catalog of music available.

But Spotify’s stickiness comes from everything around the song. For me, the cost of switching from Spotify is emotional as much as it is functional, since I have years’ worth of playlists hosted on Spotify corresponding with different phases of my life.

Beyond that nostalgia, which may not be as impactful an argument for others, Spotify’s recommendation system is what makes the product so sticky. It blends your habits with broader listener behavior and has become extremely good at discovery, and as alluded to, it has LOTS of data on what you have enjoyed listening to at various times in your life. Honestly, much of the music I listen to today didn’t come from friends but rather from Spotify’s suggestions, AI DJ, and its autoplay feature that recommends similar music after your selected playlist or album has finished playing through.

And then there’s Spotify’s hardware agnosticism. Translation: Spotify is everywhere, no matter what device you’re using, from iPhone to laptop, your car, smart TV, or smart watch, Spotify (mostly!) works very well.

As obvious as that sounds, streaming competitors that sell hardware have incentives to optimize performance inside their product ecosystem (Apple, Alphabet, and Amazon), while Spotify optimizes for ubiquity.

This is why I think specialization has benefited Spotify, because for people who treat music as a form of companion media — something that follows along with them almost no matter what they’re doing — you need a music streaming app that’s available across as many devices as you may use on a daily basis. As you switch between playing music through Apple Play in your car, your Alexa at home, and on your Samsung phone at the gym, knowing that the Spotify app will work as expected and is synced up across devices is a subtle but important point.

Spotify’s business model is simple to understand, geared around two engines, with one of the two doing significantly more of the heavy lifting.

Premium subscriptions drive about 90% of revenue. Advertising drives the other 10%, as the Robin to Premium’s Batman. Yet, of Spotify’s 750 million monthly active users, around 290 million are premium subscribers, and 460 million are free users (supported by ads).

The free tier is the top of the funnel, as mentioned. It expands reach, reduces customer acquisition friction, and converts some portion of users over time as Spotify increases perceived value. People tend to become reliant on Spotify over time, hence why “perceived value” fluctuates.

And the free product is good enough that most users are happy to stay free, but the model still works because enough people are willing to pay to bypass ads, and evidently, this is where Spotify makes all its money.

The structural problem for Spotify’s profitability is royalties to artists and labels. Spotify isn’t a classic fixed-cost software business, because roughly 2/3 of every dollar in incremental revenue, whether from ads on music or premium subscriptions, has had to be paid out to music-rights holders.

So from a margin perspective, the business looks more like Uber, where growth carries high variable costs that don’t disappear with scale. In other words, there’s a ceiling on the gross margins that Spotify can achieve, limiting how “tech-like” the profit profile can become, as I touched on in the intro.

For most of the last decade, until recently (and just barely!), Spotify has had lower operating profit margins than Amazon, a business that requires enormous capex and labor costs to support its logistics empire. If you asked the average person which business should be more profitable, I’m going to guess 99 in 100 would say Spotify!

But also, that’s why Spotify’s recent margin inflection matters so much, because we’re finally starting to get an idea of how profitable this business can be after many years of losses as the business scaled.

Shown above, Spotify’s operating margin improved from roughly -5% in 2022 to nearly 13% in 2025. 13% is incredibly modest compared to the Mag 7, besides Amazon, but it does signal that scale is finally showing up in earnings, and that Spotify’s investments outside of pure music, like podcasts and audiobooks, are starting to contribute.

Spotify has always had more pricing power than it has used, underlining investors’ debate over its earnings power. In most markets, Spotify went from 2011 to 2023 without any price hikes. Even today, U.S. pricing is only about 30% higher than it was 15 years ago, which means Spotify’s inflation-adjusted prices fell while the product improved.

Recently, Spotify has started taking price more aggressively in North America and Europe, but international expansion complicates the ARPU (average revenue per user) story. To gain traction in lower-income markets, Spotify prices lower. When you also consider the introduction of family and student plans, you can see why premium monthly ARPU fell from a peak of $6.84 to about $4.29 in 2021. Since then, after price increases globally, premium ARPU has grown again, compounding around 2% per year.

So the business is subscription-driven, and the long-term earnings power depends on three things: Pricing, conversion, and higher-margin layers on top of music.

Everyone values music, and correspondingly, Spotify’s service, differently. Some people would pay twice as much if they had to, and Spotify wants to capture that willingness-to-pay without blowing up churn across the broader base. That’s why they’ve experimented with the idea of a super-premium tier, offering perks like exclusive playlist tools and early access to tickets or album releases. For a while, people thought lossless audio would anchor that tier, but Spotify rolled it into standard premium, which suggests they’re still figuring out what “superfan” actually means in product terms.

Nevertheless, especially in North America and Europe, an opportunity remains to add a higher-priced premium tier, with the potential to significantly raise premium ARPUs. I’m still trying to wrap my head around why they haven’t done this yet, honestly, after talking about it for years. One thought is that Spotify is/was hoping to earn better margins on super-fan subscriptions, but the labels have protested (this is speculation).

Podcasts were phase one of Spotify’s expansion beyond music, and they pursued them in a splashy way. The $100m Joe Rogan exclusive deal in 2020 was…eye-popping, to say the least. The deeper strategy stems from podcasts expanding time spent on-platform and expanding ad inventory, while the unit economics can be better than music because podcast revenue isn’t constrained by the same royalty structure. There are no massive podcast labels demanding a cut.

It’s a subtle point, but premium users listen to music ad-free, yet podcasts aren’t ad-free for anyone. Spotify can monetize podcast listening through its ad network when shows opt in, and the revenue split on marketplace advertising can look closer to 50/50 with podcast hosts, rather than the two-thirds pass-through dynamic of music.

Spotify has also evolved its podcast strategy. It moved from trying to win through exclusives to trying to win by owning the creator toolchain, hosting, analytics, ad-tech, and marketplace products, so creators can grow and monetize efficiently on Spotify even if their show is available elsewhere. And bringing video podcast functionality onto Spotify is part of that shift, partly because YouTube has become a massive podcast platform (actually, the biggest!), and Spotify needed to meet creators where attention is going. P.S., you can ‘watch’ The Intrinsic Value Podcast on Spotify each week!

Audiobooks are phase two, and they’re still early. Spotify’s catalog of English audiobooks grew from roughly 150,000 to roughly 400,000 titles since the premium audiobooks launch, and they introduced Audiobooks+, an $11.99 add-on that unlocks an additional 15 hours of audiobook listening each month on top of the 15 hours included in premium. Daniel and I both felt the packaging is off, though.

Having the hours that you paid for expire each month is frustrating and disappointing, and 15 hours isn’t enough for serious readers (listeners?), either. My wife, who averages 1-2 audiobooks a week, doesn’t even bother with listening on Spotify due to how limiting and expensive the hours are.

Interestingly, Spotify also partnered with Bookshop.org, so you can buy physical versions of books you’re listening to, and they have a Page Match feature that lets you jump between a physical page and the matching point in the audiobook. I’m skeptical this will be a meaningful revenue driver, especially when Amazon is much cheaper in the price comparisons I ran, but it’s still evidence that Spotify wants to use audio to bleed into broader consumption habits and even e-commerce.

This matters because Spotify’s future growth is expected to be increasingly international, and YouTube is incredibly popular in emerging markets. Nearly 500 million people use YouTube in India alone! And for price-sensitive consumers, the YouTube bundle can be compelling, especially when YouTube is already where their entertainment lives.

Personally, I’d rather pay a bit more to keep my Spotify subscription, and if I want ad-free YouTube, there’s a tier called Premium Lite that removes most ads but doesn’t bundle in YouTube Music.