From Disc to Cloud: the birth of a "compounding machine"

In 2013, Adobe handed its users a near-insulting ultimatum: the disc you bought outright is finished — from now on you pay every month, or the software goes dark. Users revolted, signing petitions that accused Adobe of holding hostage the very tools they earned a living with. Analysts warned that forced subscription would drive users toward piracy and rivals. And financially, Adobe did the thing Wall Street fears most: it deliberately let revenue fall. Because one large upfront purchase was now spread into small monthly payments, annual revenue slid from ~$4.4B before the transition to just over $4B. Adobe chose that decline with its eyes open.

Why? Because a perpetual license is revenue once. A subscription is revenue that recurs — and, as customers add seats and prices rise, compounds. Time delivered its verdict: by FY2025 Adobe’s revenue had crossed ~$23.8B (from ~$19.4B in FY2023 and ~$21.5B in FY2024) — more than 5× its pre-transition level, and higher-quality on every axis: predictability, retention, margin. In the early transition alone, paid subscriptions grew from ~300k to ~1.4M, and Creative ARR rose from ~$155M to ~$768M.

Fourteen years earlier, in a small San Francisco apartment, Salesforce had already raised a deliberately provocative banner: “No Software.” Of course it made software — the real meaning was: don’t buy a disc, don’t have IT deploy and patch it across server after server. Open a browser, log in, and use a full CRM, paying by the seat, by the month. That flip changed the economics, not just the delivery. A license company starts each quarter near zero — last quarter’s deal doesn’t recur — while a SaaS company keeps last year’s revenue and stacks new customers on top. Revenue takes on the character of a stock: a reservoir that keeps filling, not a series of rains that must fall anew each quarter. Predictable, cumulative, high-margin (multi-tenant cloud makes the cost of serving one more customer near-zero), and self-financing (customers prepay, funding growth cheaply) — stack those four properties and you get the narrative the whole market fell in love with: software is a compounding machine.

But the story has another half, and it hides inside the users’ anger. For customers, subscription isn’t always the better deal: someone who upgrades once every few years often pays more over time, and software has gone from “an asset you own” to “a service you rent” — stop paying and you can lose access to years of your own work. That lock-in is a moat for the vendor and a loss of control for the customer. Adobe’s users weren’t wrong to be furious — and it’s exactly that lock-in that kept them paying once bound. The moat and the anger are the same fact seen from two sides.

So subscription reveals its true nature: a magnifying glass. It makes good businesses better and leaves bad ones nowhere to hide. For a business without stickiness, subscription only exposes churn, quarter after quarter, in the renewal line; “recurring revenue” propped up by relentless price hikes into a saturated market is fragile and borrowed from the future. Same income-statement line, potentially two completely different businesses. So this book sets one discipline from its very first page: recurring revenue ≠ high-quality revenue. Telling the two apart is where the whole book begins.

This is Chapter 1 of “Compounding in the Cloud: The Business of SaaS, and How to Invest in It.”