Vivian Voss

The Subscription You Did Not Ask For

licensing law saas architecture

In the Net ■ Episode 01

In 2012, Adobe shipped Creative Suite 6. A studio could buy the Master Collection once, for around two thousand five hundred euros per seat (US list price 2,599 dollars), and run it for years. Thirteen years later, the same studio leases Adobe Creative Cloud All Apps for around 743 euros per seat per year (Adobe Germany list, May 2026). The tools have not got dramatically better. The architecture under the licence rather has.

This is the first episode of In the Net, a series on the documented mechanics of vendor lock-in. The premise is simple. Every platform tells you how to come in. The architecture tells you whether you can leave. We will look at how each major platform's promise was built, what the lock-in mechanisms actually are, why the documented exit does not work as advertised, and what the realistic escape routes look like.

We start with Adobe because the pattern is at its clearest. The promise was good. The promise was kept for thirty years. The architecture under the subscription, however, is a separate story.

The Promise, Honestly

Adobe Creative Suite was, for three decades, the most defensibly chosen tool stack in design and publishing. Photoshop appeared in 1988. Illustrator in 1987. InDesign in 1999. Premiere in 1991. The applications were excellent, kept in active development, and produced the file formats that the entire commercial design industry settled on.

A studio buying CS6 in 2012 made a rational decision. Pay once, run forever. Major upgrade cycle every few years. Predictable cost, predictable workflow, the same key bindings their staff had learned over a decade. Nothing about that was a trap.

This matters. Lock-in stories are most useful when they begin with the promise that was real. Adobe's promise was real. The mechanism that came after did not undo the original promise. It changed the architecture under it.

The Day the Architecture Changed

In May 2013, Adobe announced that CS6 was the last perpetual-licence release of Creative Suite. Going forward, the only way to use Photoshop, Illustrator, InDesign and the rest would be Creative Cloud, a monthly subscription. The applications continued to ship. The licence did not.

The transition was managed elegantly. Existing CS6 owners could keep using their software. New customers and upgrades, however, were on the subscription path. Within two years, the perpetual market was effectively closed.

Three lock-in mechanisms were wired into the new architecture.

Three Independent Hooks, One Enclosure the studio tools, files, contract File formats PSD • AI • INDD proprietary, rich, archived Cloud Libraries fonts, brushes, assets decay on cancellation Cancellation fee 50 per cent of remaining annual-plan ETF

The First Hook: File-Format Binding

Adobe's file formats are proprietary and rich. PSD carries Smart Objects, Layer Effects, Adjustment Layers, and a depth of state that no open format mirrors completely. AI carries Illustrator's specific path semantics. INDD, more than the others, has effectively no portable equivalent: IDML exists as an interchange format, but loses live links, master-page state, and complex layout work.

The judgement is not that Adobe should publish their file formats. The judgement is that the file format is the lock. A studio that has spent a decade building PSDs is not just buying software; it is paying rent on its own back catalogue.

The Second Hook: Cloud-Library Binding

Creative Cloud syncs fonts, brushes, asset libraries, colour palettes, and shared resources between the applications and across teams. Cancel the subscription, and the synchronised assets become unavailable. Some assets remain locally; others depend on the licence. The working environment is not portable.

This is a more recent hook than the file format. It came in over the years between 2014 and 2020, with Creative Cloud Libraries, Adobe Fonts (formerly Typekit), and the deeper integration between Photoshop and Lightroom Cloud. Each individual feature was a real productivity gain. The cumulative effect was that the user's working environment lived inside Adobe's account, not on their own machines.

The Third Hook: Cancellation Architecture

Adobe's Annual Plan, billed monthly, carries an Early Termination Fee equal to 50 per cent of the remaining months on the contract. A user three months into a twelve-month plan who cancels owes nine months at half rate.

In June 2024, the US Department of Justice, on behalf of the Federal Trade Commission, filed suit against Adobe alleging that this fee was hidden behind layered cancellation flows and that the company had violated the Restore Online Shoppers' Confidence Act. The case (FTC v. Adobe) is, as of this writing in 2026 and to my knowledge, still in motion. Whatever the eventual outcome, the architecture itself is documented in the agreements.

A subscription that costs more to leave than to keep is not, by any honest reading, a subscription one can leave at will. The architecture says what the marketing copy does not.

The Standing: Market Power and How the User Is Treated

Two further dimensions matter, and neither is reducible to price. The first is market power. The second is how the user is treated in the contracts the user has signed.

Market position. Adobe holds, by industry aggregator estimates, more than eighty per cent of the creative-software market. Photoshop alone accounts for around forty-two per cent of the graphic-design segment, InDesign for around twenty-six per cent, Illustrator for around twelve per cent (multiple market-research aggregators citing 2024 figures). The consequence is that Adobe's three core file formats, PSD, AI and INDD, are the de-facto industry standard. A studio that has migrated to Affinity, Krita or any other alternative still receives PSDs from clients, suppliers and partners. The lock-in is not only at the studio level. It is at the level of the entire industry's file-exchange protocol. An agency that holds out alone has not actually escaped; it has only insulated its own machines, while still negotiating in someone else's format.

This is the difference between a Lock-in and a Quasi-Monopoly. The Lock-in keeps the individual customer. The Quasi-Monopoly keeps the customer's customers, and through them, the customer.

The Quasi-Monopoly: Adobe vs Everyone Else Photoshop ~42% InDesign ~26% Illustrator ~12% Adobe stack > 80 per cent Affinity, Pixelmator, Resolve, Krita, GIMP, Inkscape, Capture One: the rest

How the user is treated. In February 2024, Adobe quietly updated its Terms of Service to include language that, in many readings, granted Adobe broad rights to access user content, including to train generative AI models. The change went largely unnoticed for months. In June 2024, Adobe pushed a routine re-acceptance pop-up to Creative Cloud and Document Cloud users, and many users encountered the new language for the first time as a forced confirmation: accept, or lose access to your tools. The artist community responded with a sustained backlash. Adobe issued clarifications on its own blog on 10 June 2024, and on 24 June 2024 published revised Terms that explicitly state Adobe does not train generative AI on customer content (with the exception of submissions to the Adobe Stock marketplace), and that the company never has.

The clarification is, as far as one can verify, accurate. The clarified Terms are clearer than the prior wording. The question that remains is what was being granted before the backlash, and why a company with this market position needed to be told by its own users that this is not the language a creative tool's licence ought to contain. The original wording was not a slip. It was the architecture of the contract a corporation believed it could propose to its captive user base. That belief, and not the words themselves, is the part of this story that does not get corrected.

These two dimensions, market position and contractual stance, change the reading of the price. The price is not only what the studio pays in money. The price is also what it accepts, by signing the contract, about the rights it retains over its own work, and about the position from which the contract is offered. A subscription is the lower-cost end of the price. The other end is harder to put a number on.

The Exit That Isn't

Adobe will tell you the exits exist. They are correct, technically. They are wrong, operationally.

Export your PSD to a portable format. Smart Objects flatten. Layer Effects sometimes lose their parameters. Adjustment Layers may bake into the underlying pixels. The file opens elsewhere; it is no longer the file you built. Anyone who has tried to recover an old PSD in another tool has met this in person.

Cancel your subscription. The Early Termination Fee triggers. The flow is layered. Customers report that the cancellation requires multiple confirmations, optional retention offers, and live-chat negotiation. The exit is documented. It is also a fence.

This is a Lock-in by design. Not in the sense that someone in a boardroom said "let us trap our users". In the sense that the architecture, taken as a whole, produces the outcome that users do not leave even when they would prefer to. That is what an architecture is: the outcome the structure makes likely, regardless of intent.

The Price

For a studio of ten seats on Creative Cloud All Apps, the recurring cost is in the order of seven thousand four hundred euros a year, every year, with no terminating event. Compare to the perpetual model: ten CS6 Master Collection licences in 2012 would have cost around twenty-five thousand euros and remained usable for the years the team chose to run them. The subscription model has been more profitable for Adobe and more expensive for studios across any horizon longer than a few years.

Ten Seats, Thirteen Years — Perpetual vs Subscription 100k EUR 0 2012 2026 years CS6 perpetual: ~25k EUR, flat CC subscription: ~7.4k EUR / year, cumulative break-even ~year 4

For a studio that wants to leave, the cost is migration. Months of file conversion, weeks of retraining, and a tail of legacy work that will, rather inevitably, not transfer cleanly. This is not a fee Adobe charges; it is a cost the architecture imposes.

The Escape Route

The alternatives are real and professional, but the landscape has shifted in 2024 and 2025. Two of the strongest names have changed owners. The migration paths are still good. They simply require attention.

Affinity, by Serif (now Canva). In March 2024, Canva acquired Serif for approximately three hundred million pounds. In October 2025, the three Affinity applications (Photo, Designer, Publisher) were consolidated into a single application called Affinity by Canva, version 3.0. The new model is free with a Canva account. AI features are locked behind a Canva Pro subscription, around one hundred and ten euros a year. The original Pledge given at acquisition included a commitment to perpetual licences; that commitment was, by the publisher's own communication at the version 3.0 launch, replaced when the new model launched. Affinity remains a strong one-to-one replacement for the Adobe core, with direct PSD and AI import. It is also itself an example of the pattern this series tracks.

Pixelmator Pro, by Pixelmator (now Apple). Apple announced its acquisition of the Pixelmator team in November 2024 and completed it in February 2025. The standalone perpetual licence is still listed on the Mac App Store at around fifty euros. Newer features, such as the Warp tool added with the iPad release in January 2026, are exclusive to Apple Creator Studio: a subscription bundle priced at 12.99 dollars per month or 129 dollars per year, which ships Pixelmator Pro alongside Final Cut Pro, Logic Pro and the iWork suite. The standalone path remains, but the rate of new feature development sits inside the bundle.

DaVinci Resolve, by Blackmagic Design. Free for video editing, with a Studio version available as a one-time purchase. The free version includes colour grading that has, afaik, been used on theatrical productions; Blackmagic Design lists notable feature-film credits on its own product pages. The most stable of the alternatives in this list, with a perpetual non-subscription model that has held since 2009.

Krita. Free and open source, professional digital painting application maintained by the KDE community. Strong brush engine, deep tablet support, scriptable. Stable.

GIMP and Inkscape. Free, photo and vector. Robust and feature-complete for most workflows. The user experience is, honestly, dated, and migration cost from Adobe is real work because the muscle memory does not transfer cleanly. They are the right choice for budget-constrained teams or FOSS-aligned organisations that can absorb the UX cost.

Capture One. A direct Lightroom alternative. Individual perpetual licences are still offered (US list 299 dollars). Multi-user studios, however, were moved to subscription-only in 2024 (PetaPixel reported a 344 per cent multi-user pricing change). This is the same pattern as Adobe: perpetual quietly retained for individuals, removed for the larger commercial users.

The maths still works. Adobe Photoshop and Lightroom Photo plan in Germany is around 17.99 euros per month (annual commitment, billed monthly) or about 142 euros per year prepaid; Photoshop Single App and Creative Cloud All Apps cost more. Affinity is free with an account. Pixelmator Pro standalone is around fifty euros, perpetual. DaVinci Resolve is free. The break-even against any Adobe subscription happens within months, not years.

The harder question is which of the alternatives stays an alternative. Two of the largest names have changed owners in eighteen months, and the change has, in both cases, moved the architecture in directions worth watching. This is not an argument against the alternatives. It is an argument for the only sustainable strategy in this landscape: attention.

What to Take Home

If your tools come with a recurring bill, your back catalogue lives in someone else's licence. If your tools come from a vendor that holds eighty per cent of the market, the industry's standard formats are theirs as well, and your migration is incomplete even when your studio has finished it. If your tools come with Terms of Service that granted broad rights over your content until your community noticed, the architecture of the contract is a separate question from the price tag.

The pattern repeats. The market holds the lock-in. The contract holds the user. Even the alternatives need watching, because two of the largest names have changed owners in eighteen months. The subscription you did not ask for is the architecture you signed up for: the rent has gone up, the formats have stayed standard, and the wording in the agreement has, for at least one entire spring, said something the publisher later had to take back.

The tools have not got better. The rent has, the standing has not, and attention is the only price that scales.

Adobe CS6 Master Collection in 2012: ~2,500 EUR per seat, perpetual. Adobe Creative Cloud All Apps in 2026: ~743 EUR per seat per year, in perpetuity. Three hooks: proprietary file formats (PSD, AI, INDD), cloud-library binding, 50 per cent Early Termination Fee on annual plans. US DOJ filed in June 2024 over hidden ETFs; case still in motion. Eighty per cent market share makes the lock-in industry-wide. Affinity (Canva, March 2024), Pixelmator (Apple, February 2025), Capture One Multi-User (subscription-only, 2024) all changed shape in eighteen months. Attention is the only price that scales.