Software was historically bought once. The dominant model is now recurring payment, and the transition changed the relationship between user and product in ways worth examining.
What changed
Under the purchase model, you bought a version. It continued working indefinitely. Upgrades were optional purchases.
Under subscription, you pay for access. Stop paying and access ends, in most cases including access to a version you'd already been using.
The distinction matters most for anything where you've created content. Documents, projects and files created with subscription software may become inaccessible without an active subscription, depending on format and on whether alternatives can read it.
The case for subscription
Stronger than the complaints suggest, and worth stating fairly.
Development funding. Continuous development requires continuous revenue. The purchase model produced a pattern of feature-stuffed major releases timed to drive upgrades, which isn't obviously better for users.
Lower entry cost. Professional software costing a large sum upfront excluded people who could afford a monthly amount. Subscription genuinely widened access.
Security maintenance. Ongoing patching costs money, and the purchase model provided no revenue for it after the sale.
Alignment of incentives. A vendor dependent on renewal has reason to keep users satisfied continuously, rather than only at the point of the next release.
Predictability. For businesses, an operating expense is frequently preferable to a capital purchase.
The case against
Also stronger than dismissals suggest.
Total cost. Over several years, subscription frequently costs considerably more than a purchase would have. For users who don't need continuous updates, this is straightforwardly worse.
No stopping point. A purchased tool remains available indefinitely. A subscription must be maintained forever, including through periods of reduced income.
Price increases. Subscription pricing can be raised, and users with content locked in the format have limited ability to refuse.
Feature removal. Functionality can be moved to a higher tier after users have committed.
Discontinuation. If a vendor stops the product, subscribers lose access to the tool entirely rather than retaining a working version.
Requirement to be online. Many subscription products require periodic authentication, which means connectivity failures or account problems prevent access to your own work.
The lock-in mechanism
The element that determines how much bargaining power users retain.
Where a product uses a proprietary format that only it can read, your existing work is hostage to continued subscription. Leaving means abandoning it or accepting a lossy conversion.
This is the single most important thing to evaluate before adopting any tool professionally.
Products using open or documented formats, or offering complete export to standard formats, preserve your ability to leave. Products that don't have effectively priced their future increases into your dependency.
Middle positions
Several models exist between the poles and receive less attention.
Perpetual fallback licensing. Subscribe, and if you stop, you retain a working licence for the version current at the time you stopped. Some vendors offer this and it addresses the main objection.
Buy a version with optional updates. Purchase the current version outright, with a subscription for future updates. If you stop, you keep what you have.
One-time purchase with paid major upgrades. The traditional model, still used by some independent developers.
Open source. No licensing dependency at all. Quality varies enormously by project and in several categories the open alternatives are genuinely competitive.
What to evaluate before adopting
Practical questions, particularly for anything you'll depend on professionally.
What happens to my files if I stop paying? Can I open them with anything else?
Can I export completely, in a format something else can read, without losing structure?
Does it require connectivity to function, and what happens during an outage?
What has the vendor's pricing history been?
Has the vendor discontinued products before, and what happened to those users?
Is there a perpetual fallback, or does access end immediately?
The practical position
Subscription isn't inherently exploitative and it isn't inherently good value. It depends on usage.
If you use something intensively and benefit from continuous updates, subscription is frequently reasonable and sometimes cheaper.
If you use something occasionally, or need a stable tool that doesn't change, it's poor value and a perpetual licence or open alternative is better.
The thing worth resisting is the lock-in rather than the pricing model. A subscription you can leave is a commercial arrangement. One you can't leave because your work is trapped is something else, and that distinction is visible before you commit.
What happens when a vendor is acquired
A scenario worth planning for, because it is common and the outcomes follow a pattern.
Acquisitions frequently produce price increases, migration to the acquirer's platform, feature consolidation, or discontinuation of the acquired product entirely.
Users generally have no say and limited notice, and the terms permit it.
Which is another argument for evaluating export capability at adoption rather than at the point it becomes urgent. The moment to establish that you can leave is when you have no reason to want to.