Subscription Fatigue and the Real Cost of Dark Patterns

Everything above in this series argues for recurring revenue. This note is the counterweight: the backlash is real, it's accelerating, and the tactics companies are using to fight it are quietly the most expensive mistake in the whole model.

The backlash, in numbers

The average American now juggles around 8.2 subscriptions and spends roughly $1,416 a year on them — and can't track it. There's a striking perception gap: people estimate they spend about $86/month and actually pay $219. In 2025, 47% of streaming subscribers felt they overpay, and video streaming posted a 44% churn rate in Q4 2024, the highest on record. Most telling for anyone setting prices: 60% say they'd cancel a favorite service over just a $5 increase.

That's the environment every new subscription launches into now. Each additional one faces more scrutiny than the last.

Dark patterns are regulatory liability now, not a growth tactic

When churn rises, the tempting move is to make leaving hard. That era is ending. The FTC's June 2024 action against Adobe is the template for what's coming:

• Defaulting users into the pricier "annual, paid monthly" plan while burying the early-termination fee in fine print.
• Making users hover over tiny icons to even discover that fee.
• Cancellation flows built as a maze of pages.
• Support reps trained to deter cancellation with delays, transfers, and dropped calls.

The FTC's "Click to Cancel" principle — cancellation must be as easy as signup — is the direction of travel, with the EU's Digital Services Act and Australia's ACCC moving the same way. The hall of shame is instructive: HP using DRM to disable third-party ink, BMW charging monthly for heated seats already installed in the car, and X Premium converting only about 827,000 of a claimed 540 million users — under 0.15% — after degrading the free experience to push people to pay.

The through-line: every one of these bought a little short-term retention and paid for it in trust, regulation, and brand damage worth vastly more.

Forced conversion churns worse anyway

Even setting ethics aside, coercion doesn't even work. 30–50% of all customer churn happens in the first 90 days. Users pushed into a subscription by an auto-converting trial — "forced converters" — churn at around 45%, while people who voluntarily upgraded were 70% less likely to churn. One company that switched from a trial model to freemium watched monthly churn fall from 4.2% to 3.1% — a gap that compounds enormously over a couple of years.

(There's also a quieter leak worth naming: about 11% of subscription revenue is lost to involuntary churn from failed payments alone. A decent dunning/retry system recovers roughly 70% of that. It's the rare retention win with no ethical downside — you're recovering revenue from people who wanted to keep paying.)

Why I make cancellation easy on purpose

I make it easy to cancel ResistGate and Amethyst Pro — not only because regulators will eventually require it, but because it's better business. Easy cancellation raises trial and signup conversion, because people feel safe to start when they know they can leave. And the users who stay are the voluntary kind who barely churn. The entire trust posture I've built — the reason someone picks a privacy-first blocker like ResistGate over Cold Turkey — evaporates the first time I make leaving feel like a trap.

The lesson: retention bought with dark patterns is a loan against your own future — it accrues regulatory risk, wrecks your reputation, and makes it impossible to ever raise prices. Retention earned through value is the only kind you actually own. In a world where people cancel over $5, being the company that's easy to leave is a competitive advantage, not a leak.

Part 7 of From Free to Premium. Next: what loot boxes and battle passes teach every builder.