The shift from selling products to renting access was the dominant business model change of the last fifteen years. It worked because recurring revenue is worth far more than one-off sales, and customers barely noticed individual charges.
The noticing has started.
What changed
Accumulation, mostly. One subscription is invisible. A household with fifteen of them, several forgotten, is looking at a meaningful monthly figure that arrives in fragments and never gets reviewed as a whole.
Rising costs elsewhere prompted the review. Once households started auditing spending, the subscriptions were the easiest line to cut, because cancelling requires no lifestyle change and takes minutes.
The company response
Predictable and mostly counterproductive. Harder cancellation flows, which generate regulatory attention. Cheaper advertising-supported tiers, which reduce revenue per user. Bundling, which is a return to the cable model that streaming was sold as an escape from.
The response that works better and is used less is simply making the product worth the money, which requires spending rather than optimising.
The category that survived
Anything where the alternative is genuinely worse. Software that a business depends on daily. Services where the subscription replaces something more expensive.
What is being cut is discretionary and duplicative – the fourth video service, the app used twice, the tier upgrade nobody remembers choosing.
The regulatory turn
Rules requiring cancellation to be as easy as signing up have appeared in several jurisdictions. That sounds minor and is not: a meaningful share of subscription revenue comes from people who intended to cancel and found it too tedious.
What comes next
Probably consolidation, with fewer larger bundles at higher prices. Which is where every media distribution model has ended up, repeatedly, and the interval between disruption and reconsolidation keeps getting shorter.