Adobe Sold You a Subscription. The Real Product Was the Difficulty of Leaving.

Try to cancel your Adobe subscription mid-year and you discover the part of the deal nobody read aloud: the plan you almost certainly signed up for - 'annual, paid monthly' - carries a 50% early termination fee, and that fee was, regulators alleged, tucked behind a hyperlink while the friendly monthly price sat in big type up front.⁷ The monthly number was the bait. The thing you actually bought was the cost of changing your mind. For a decade that hidden cost was a feature, not a bug - and it built one of the most reliable cash machines in software.

The official story is that Adobe moved to the cloud in 2013 to give creatives a cheaper, always-updated, more flexible product. The real story is that Adobe discovered the most valuable thing it owned was not Photoshop. It was the difficulty of leaving Photoshop - and it learned to charge for that difficulty on a recurring basis, forever.

Why killing perpetual licenses was the whole point

Adobe had sold subscriptions alongside boxed software since 2011, hedging. Then in May 2013 it stopped hedging: Creative Cloud went subscription-only, and CS6 became the last version of Creative Suite you could ever own.⁵ Read that as a pricing change and you miss it. It was a change in the nature of the transaction. A perpetual license is a sale with an exit - buy CS6, skip the next upgrade, keep working for years on what you already own. A subscription has no exit, because the software stops the moment the payments do. Adobe didn't just raise the lifetime price. It removed the option to stop paying while keeping the tool, which is the only leverage a customer ever had.

And it was not obviously cheaper for the customer, which is the tell. A user on the common two-to-three-year upgrade cadence accumulates well over a thousand dollars across that span - enough, at many historical price points, to meet or beat what a perpetual upgrade would have cost. The 'subscriptions save you money' line was true only for the heaviest buyers, the ones who'd dutifully purchased every annual release at full freight. For everyone else, the deal traded a lower headline price for a permanent meter. The meter is the moat.

Creatives saw it instantly. A Change.org petition demanding the end of the mandatory model gathered roughly 38,000 signatures by mid-August 2013 - a real revolt, if a smaller one than the inflated 50,000 figure that later retellings settled on.⁶ It didn't matter. The switching costs were already too high: years of muscle memory, client files in proprietary formats, a workflow built around tools no rival fully replaced. The petition was the sound of people realizing they were locked in - and signing it anyway, from inside the lock.

The math that turned grumbling into $17 billion

Switching costs plus a meter you can't pause equals recurring revenue that only grows, and the financials read like a machine working exactly as designed. Adobe's fiscal-2023 revenue was $19.41 billion, with non-GAAP operating income of $8.92 billion - the margin profile of a company that captured a customer base and now simply collects.² Creative revenue alone reached $12.68 billion in fiscal 2024, still growing 10%, and Digital Media annual recurring revenue exited that year at $17.33 billion.¹ Annual recurring revenue is the cleanest possible name for what was built: a base that renews by default, because not-renewing means losing the tool you've organized your livelihood around. Even the latest filing credits the growth to its flagship apps and Acrobat doing the same patient work.⁸

When the friction became a federal case

A moat built on how hard it is to leave invites a specific kind of trouble: someone eventually asks whether the leaving was made hard on purpose. On June 17, 2024, the DOJ filed a complaint on behalf of the FTC alleging Adobe failed to clearly disclose the material terms of its annual-paid-monthly plans and made cancellation unfairly difficult, even naming two executives as individual defendants.³ The complaint's spine was the design itself: the annual plan pre-selected as the default, the monthly price displayed proudly, the 50% termination fee buried in fine print or one click away.⁷ In March 2026 Adobe settled for $150 million - a $75 million cash penalty to the government plus $75 million in free services - while denying any wrongdoing, with court approval still pending.⁴

Isn't this just a great business getting punished for being great?

The honest objection is that Adobe earned its lock-in fair and square. Photoshop, Premiere, and the rest are genuinely the best tools in their categories; the switching costs are high because the alternatives are worse, not because Adobe rigged a maze. By that read, $150 million is a parking ticket against a base producing nearly nineteen billion in revenue, and the company will pay it and move on.¹² Fair - up to a point. The flaw in it is the distinction the settlement quietly draws. There is a difference between a moat made of real product superiority and a moat made of obscured terms and a buried exit fee. The first is defended utility; the second is behavioral friction monetized in the dark. Adobe has plenty of the first. The case was about the second - and the second is precisely the part regulators have now signaled they will normalize away. 'Easy cancel' as a baseline expectation strips out the friction premium without touching the software's quality at all. The product moat survives. The dark-pattern surcharge does not.

Adobe spent a decade collecting two tolls at once: the price of the best tools in the business, and the price of being unable to easily put them down. The settlement doesn't touch the first toll - the software is as good as ever, the recurring revenue as fat. It targets the second, and in doing so marks where the line now sits. The lock-in that made Adobe a money machine was always two things wearing one face: a great product people won't leave, and an exit quietly made hard to find. The first one was the moat. The second one was the bill - and it just came due, at $150 million, with the regulator promising it won't be the last company to pay it.