Why software is still worth studying — after the bubble burst

Rewind to the winter of 2021. Software was worshipped like a god. The market had convinced itself that SaaS was a business that almost couldn’t go wrong: revenue snowballed on itself, customers never left, and any multiple could be justified by “durable growth.” Then rates rose, the bubble burst, and names that had 10x’d on the way up gave most of it back on the way down. The faith was cruelly re-examined, and a lot of people who had confused a rising market for skill discovered which was which.

Out of that wreckage came a better question — not “will software go up?” but “which software companies were ever real compounding machines, and which just charged like one?” That single question is why this book exists, and why it starts from the parts of the boom that were never true.

Three uncomfortable truths sit underneath it. First, “recurring” revenue is not automatically “high-quality” revenue — a company that loses a fifth of its customers a year and buys growth with cash has a completely different business from one whose customers stay and expand, even if the two top lines look identical. Second, high gross margin is not the same as real cash return once you account for stock-based compensation and the dilution it hides. Third, and most important: a good product, a good company, and a good stock are three different things. An exceptional company whose price already discounts a decade of optimism can be a terrible investment; a mediocre one, priced for despair, can be an opportunity. The gap between them is called price — and, more precisely, the expectations baked into it.

Notice the through-line: none of these are accusations that software is a bad business. They’re the opposite — software can be one of the greatest business models ever built. The point is that the label “SaaS” guarantees none of it, and the boom’s fatal move was pricing every subscription company as if the label alone were the moat.

So this book will not hand you a list of tickers — any such list goes stale the moment prices, earnings, and the competitive map move. It hands you something more durable: a way to read any SaaS company, along four threads that recur in every chapter. Revenue quality over headline growth. Cash-flow truth over accounting profit. Moat position over feature lists. And valuation understood as a statement about expectations, not a number to be reverse-engineered into a “buy.”

And it does all this in the middle of the biggest shock to the model since it was born. AI is rewriting software’s margins (inference has a cost), its pricing (why charge per seat when one agent does ten people’s work?), and its moats (what’s defensible when everyone can call the same model?) — all at once. The goal here isn’t to predict how that ends. It’s to hand you the ability to form your own answer, chapter by chapter. In an era where software starts doing the work and everything moves faster, that ability is the one asset that doesn’t go stale.

This is the preface to “Compounding in the Cloud: The Business of SaaS, and How to Invest in It.”