Productivity Software
Adobe
Adobe charges roughly the same real price for materially more product. What changed hands is control.
Adobe is deliberately a complicated CHI case, and this page holds both halves of it at once. The flagship Creative Cloud plan is cheaper in constant dollars in 2026 than at its 2012 launch, the bundle roughly doubled, and the upfront barrier to professional creative work fell by about 97%. That is the value, and it is unusually high. The hostility is real too, and it is not located in the headline price. It is concentrated in a 12-month commitment displayed as a monthly price with a 50%-of-balance exit fee that produced a $150 million federal settlement; a default migration of existing subscribers to the more expensive of two new tiers; a generative-AI allowance left unenforced until customers depended on it and then cut by 95%; and a 2019 demonstration that even installed software can be de-licensed by a dispute the customer never knew existed. Adobe did not become a story of charging progressively more for progressively less. It became a story of transferring control: the customer once held a durable, self-executing position, and now holds a set of defaults that resolve in Adobe's favour unless actively contested.
CFS = CVI − CHI (83 − 50 = +33). Assessed under CHI/CVI Methodology v2.0: dimensions score the underlying customer harm, patterns describe the behavior that produced it. CHI 50, CVI 83 and CFS +33 are the settled analytical values from the completed Adobe evidence dossier and are not presented as a range. Methodology →
The defining finding
The default direction reversed
Under perpetual licensing, inaction was safe: a customer who ignored Adobe kept working software indefinitely and spent nothing further. In 2026, inaction produces automatic renewal at a higher tier, a 50%-of-balance exit fee, monthly forfeiture of unused credits, and deletion of cloud content after an undefined period of inactivity. Fact May 2019 showed the same logic reaching software already installed. Strongly supported
The price is roughly the same in real terms. The default state is not.
Executive finding
A genuinely split verdict, held deliberately.
The evidence establishes meaningful customer-hostile mechanisms and unusually high customer value in the same relationship. CHI scores them separately, on purpose, and neither is permitted to erase the other.
The value - CVI 83
Creative Cloud All Apps launched at $49.99/month in April 2012; its 2026 successor, Creative Cloud Pro, is $69.99. Nominal +40.0% against cumulative inflation of +45.5%: the flagship is roughly 2-4% cheaper in constant dollars than at launch. Fact The bundle expanded materially over the same period, from roughly 14-15 desktop applications in CS6 to 20+, now including Acrobat Pro (a separate ~$449 purchase in 2012), the mobile line, Adobe Express, Firefly, cloud storage and Adobe Fonts. Fact A student today reaches the professional toolset for $240 in year one, against roughly $3,780 in 2026 dollars for CS6 Master Collection. Fact
Adobe delivers more product for slightly less real money than in 2012. The positive finding is not buried here; it constrains everything else.
The hostility - CHI 50
The documented extraction is concentrated in five places: the annual-paid-monthly architecture with its 50%-of-balance early termination fee, which produced a $150 million federal settlement and injunction in March 2026; the withdrawal of the cheap Photography tier for new customers, doubling the entry price; the June 2025 default migration of existing subscribers to the more expensive of two new tiers; a generative-AI allowance left effectively unenforced for roughly 17 months and then cut from 500 to 25 credits; and the May 2019 retroactive de-licensing of installed software. Fact Strongly supported
The hostility is not in the headline price. It is in what happens when the customer does nothing.
The relationship timeline
From ownership to defaults.
Under the perpetual model a customer paid once and received an indefinite right to run that version. Every entry below moves some part of that position onto terms that require continuing payment or affirmative action to preserve.
Upgrade Escalation · Hero exhibit
The default ran toward Adobe in both directions.
The same existing All Apps subscriber, June 2025, taking no action.
Before renewal (to 17 Jun 2025)
$59.99per month, All Apps, annual billed monthly
- 20+ desktop apps, mobile and web apps
- Creative Cloud Standard existed at $54.99 - less than this customer was already paying
- Genuinely marketed and directly purchasable
After renewal, by default
$69.99per month, Creative Cloud Pro (+16.7%)
- Reaching the cheaper tier required affirmative action
- Standard strips mobile and web apps (non-AI functionality) alongside AI
- Students and teachers cannot buy Standard at all
The criticism is not that Standard was fake or hidden. Standard appears in the main plans grid, can be purchased directly by new customers, and has its own marketing page with buy-now calls to action; switching is free. The important finding is narrower and better evidenced: customer inaction resolved toward the more expensive option. Before the change landed, existing subscribers were given Pro-tier benefits at no charge and then billed for them at renewal - the mechanism that makes a default stick. What is supportable beyond that is de-emphasis: Standard sits at roughly position 23 of 26 on the plans page, below every single-app plan, with no badge and no promotional pricing. Intent is not established.
To customers the change was "new tools and creative capabilities"; to investors it was "value-based pricing" and "migration over to the Creative Cloud Pro plan, the higher-value plan." Both quotations are verified verbatim against Adobe's published transcripts. Fact
Price Creep · The verdict is three verdicts
One finding would be false. The index records all three.
"Adobe has Price Creep" and "Adobe does not have Price Creep" are both wrong. The CPI analysis disposes of the flagship claim; it does not touch the other two. A customer whose bill rose 16.7% in one move experienced that move regardless of where the thirteen-year trend sits.
1 · Flagship, long-term real
NOT ESTABLISHED
$49.99 at April 2012 launch; $69.99 in 2026. Nominal +40.0% against cumulative inflation of +45.5%: roughly 2-4% cheaper in constant dollars, while the bundle expanded materially. This is important counter-evidence, and publishing the opposite would be factually wrong. Fact
2 · Photography entry
SUPPORTED
The new-customer entry point moved roughly $9.99 → $19.99 following withdrawal of the cheaper 20GB tier in January 2025: about +100% nominal and roughly +40% real. The mechanism was tier withdrawal, not a headline price rise. Existing annual-prepaid holders still pay the 2013 price. Fact
3 · June 2025 discrete escalation
SUPPORTED
Existing All Apps customers were defaulted $59.99 → $69.99, roughly +16.7%, while Creative Cloud Standard existed at $54.99 and was genuinely marketed and directly purchasable. The finding is the default, not the existence of the price. Fact
The architecture behind the numbers
Adobe's default individual-plan presentation is an annual commitment billed monthly, displayed as a monthly price. Of Adobe's three payment options, two are twelve-month commitments; only the most expensive, at roughly a 1.6× premium, is genuinely cancel-anytime. This pricing architecture is the direct origin of the early termination fee examined below. Fact List price is also not paid price: Adobe discounts routinely, and at research time displayed Creative Cloud Pro at $34.99/month for the first year. Fact
The defensible criticism concerns the timing, size and communication of individual steps, not the thirteen-year price level.
Cancellation Friction · Temporally precise
Historical: confirmed. Current: unknown.
This finding must be split in time. The historical record is the strongest in the file. The current record, after the March 2026 injunction, is genuinely unresolved, and this page does not assert continued friction.
Historical cancellation friction - CONFIRMED
Adobe's annual-paid-monthly architecture historically included an early termination fee of roughly 50% of the remaining contract balance - in Adobe's own example, cancelling in month nine means paying half of the three remaining months. Fact No competitor in the control group was found to charge an early termination fee on a comparable individual plan. Strongly supported The United States alleged that ETF terms were buried "in small print or behind hover-over icons" and that cancellation processes "are designed to make cancellation difficult"; the complaint reportedly relays an internal characterisation of the fee as "heroin" for the company. Allegation
Customer reports describing the exact mechanism the government later charged predate the complaint by two and a half years, across independent platforms, from 2021 onward. Customer report The federal action resolved on 13 March 2026 in a $150 million settlement and injunction; Adobe denies wrongdoing and there was no adjudication of liability. Regulatory
The injunction's remedies target exactly the three behaviors customers had described since 2021: disclosure, reminders, easy cancellation.
Current post-settlement friction - UNKNOWN, leaning materially reduced
Adobe now documents a six-step, fully self-service cancellation flow at account level, with no phone call, chat or agent step and no retention interstitial, and most plans carry a full refund within 14 days. Fact That is the strongest single fact against a current-friction finding, and it is recorded as such.
Two issues remain unverified without an actual transaction: whether the early termination fee is disclosed on the live payment and order-review screens, and whether trial-conversion reminders are sent. The cancellation help page updated after the settlement does not mention the ETF, while the page that explains it accurately predates the settlement - not itself evidence of non-compliance, since the injunction requires disclosure before enrolment, which sits on screens that could not be observed without purchasing. Unknown No regulator, court, researcher or consumer group has audited compliance since March 2026. Fact
Do not read this page as claiming Adobe currently continues the historical prohibited behavior. It does not establish that, and neither does the available evidence.
Rentalization / Access Dependency · The strongest structural finding
Software access ended. File access, largely, did not.
Perpetual revenue fell from $3,416M in FY2011 to $89M in a single quarter of FY2026. There is now no Adobe creative product a customer can buy and keep. The finding is confirmed - and deliberately narrowed to software access, not files.
What continuing payment now controls
Under perpetual licensing, customers purchased software, decided whether to upgrade, and refusing an upgrade did not terminate already-purchased software; inaction generally preserved the customer's existing position. Under Creative Cloud, continued application functionality depends on continuing payment: Photoshop, Illustrator, InDesign and Premiere Pro cease to function on lapse. Fact Some working-state assets and services also depend on the subscription relationship: generative credits are forfeited immediately without proration, cloud storage drops to 5GB with files subject to deletion after an undefined "period of inactivity," and Lightroom cloud originals are retained roughly one year, then warned, then permanently deleted. Fact The vagueness of the deletion clock - Adobe's only governing language is "a period of inactivity" with no number attached - is itself the finding. Fact
Ending the commercial relationship eventually ends paid application functionality. That is the dependency.
The mandatory scope limit - this is not file hostage-taking
Adobe controls continuing access to tools and subscription-linked functionality. That is different from owning the customer's files. Locally stored .psd, .ai, .indd and .prproj files remain on disk, untouched, and openable in third-party software; no Adobe documentation describes any mechanism that deletes, encrypts or locks local files on cancellation. Analytical inference Lightroom Classic is the decisive counter-example: after cancellation the catalog remains fully functional and Export, Print, publishing and Quick Develop all keep working - Adobe deliberately preserved file egress while removing the Develop and Map modules. Fact
A lapsed customer can still get every photo and every edit out in standard formats. That is not the architecture of a system designed to trap files.
Adobe Fonts - the sharpest practical harm, correctly scoped
Confirmed for editable working documents: after cancellation, activated Adobe Fonts deactivate, and editable documents open with missing-fonts warnings and program-default substitution, which can change design layout or appearance - font substitution, text reflow, altered layout. Fact The scope limit is equally confirmed, and Adobe states it explicitly: already exported and static output remains intact, and files embedding font data "may be reproduced and distributed independent of your subscription status." Fact No permanent font licence is retained, and the Package command does not carry the font files.
The customer's file remains theirs, but an editable working document may no longer look as the customer created it.
Generative credits · Promise Reversal · Hero exhibit
Unenforced until load-bearing. Then cut by 95%.
The generative-AI allowance on Photography and Single App plans, before and after June 2025 enforcement.
Announced Jan 2024, unenforced ~17 months
500credits per month, effectively unmetered in practice
- Limits announced but not enforced for most Creative Cloud subscribers
- Customers built Generative Fill into production workflows
- Adobe's stated fallback: exhausted users "may be slower," not blocked
Enforcement, June 2025
25credits per month (−95%), at unchanged or higher prices
- Allowances cut 500 → 25 on Photography and Single App plans
- Promised throttling replaced by a hard stop, with no prominent retraction
- Top-ups sold as recurring monthly add-on subscriptions, not one-time purchases
The honest formulation: Adobe did not meter old features. It left new features unmetered long enough for them to become load-bearing, then metered them and cut the allowance by 95%. The graceful-degradation commitment was replaced by a hard stop visible only by comparing documentation versions. Promise Reversal is the finding this supports, and it is strongly supported. Credits also do not roll over, and unused credits are forfeited entirely on cancellation with no proration.
Adobe's own documentation states that credit consumption depends on "the generated output's computational cost and the value of the generative AI feature used" - by Adobe's own words, pricing is set partly by value, not solely by compute. To investors, credit consumption is a headline growth KPI that routes users into "higher-value Creative Cloud offerings or acquiring Firefly credit add-ons." Mechanism and outcome are established; motive is not asserted. Fact
Artificial Scarcity · Scoped narrowly
The broad accusation fails. A narrow one survives.
Where the scarcity claim fails - and must be conceded
Inference genuinely costs money: every generation consumes GPU time, video diffusion is orders of magnitude more expensive than image diffusion, and the credit ratios between modalities track real cost ratios closely. Metering video is economically rational, not artificial. Fact Metering is universal in this market and no major vendor rolls credits over. Strongly supported Adobe provides significant included generative capability: standard image and vector generation is unlimited on Creative Cloud Pro, and Firefly standalone plans are competitive-to-generous against peers. Fact
Broad Artificial Scarcity is not supported. The claim that Adobe's allowances are unusually restrictive is false for Firefly standalone and for image workflows on Pro.
Where it survives - precisely
The bundled low tiers and video. Adobe sells 2,000 credits for $9.99 standalone, so at Adobe's own retail rate the 25-credit allowance is worth about twelve cents - a $54.99/month plan includes roughly $0.12 of generative capacity. That number cannot be explained by inference cost, and it was reached by cutting an existing entitlement. Fact In video, the 20-100 credits-per-second burn rate means even Pro's 4,000 premium credits buy about 200 seconds of lowest-tier video per month; an allowance that sounds ample in images is trivial in video. Fact
The scarcity finding survives only in the specific low-credit and video contexts. It is not published as a general charge.
May 2019 · Historical exhibit
Installed software, de-licensed by a dispute the customer never knew existed.
Retroactive de-licensing - STRONGLY SUPPORTED as a composite event
Around 8-13 May 2019, Adobe notified some Creative Cloud customers that they were "no longer licensed to use certain older versions of the applications or deploy packages containing these older versions," and warned that continued use could put them "at risk of potential claims of infringement by third parties." The dispute arose from third-party royalties in which the customer had no part and no visibility. Contemporaneous press connected the dispute to Dolby, which had sued Adobe over royalties in March 2019; Adobe itself named no third party in any located communication, and this page does not state the Dolby connection as Adobe-confirmed fact. Strongly supported
Adobe narrowed the notice within twelve days, clarifying on 20 May 2019 that CS6 perpetual-licence customers were unaffected and could continue using their software as normal. Fact One component is confirmed at the highest tier: Adobe's still-live support-policy page states that deploying older, unsupported Creative Cloud versions excludes customers from third-party claim coverage under their contract. Fact No first-party Adobe artifact for the "no longer licensed" wording was recoverable, which is why the composite event is strongly supported rather than confirmed.
This incident is historical evidence, corroborating rather than load-bearing for the current Access Downgrading finding - which rests on four current, documented mechanisms: fonts, forfeited credits, the storage cut, and Lightroom cloud deletion.
Information & privacy · The viral version was wrong
Adobe does not train Firefly on your Creative Cloud files.
The evidence does not support the most repeated accusation, and CHI's credibility depends on saying so plainly. The narrower concerns that do survive are recorded with equal care.
Rejected: training on customer content
The claim that Adobe trains Firefly on ordinary Creative Cloud customer files is not established, and it has been contradicted by Adobe's contract since June 2024: "We will not use your Local or Cloud Content to train generative AI models except for Content you choose to submit to the Adobe Stock marketplace," with the sublicence loophole expressly closed. Fact No credible contrary evidence has emerged in over two years, and no privacy regulator anywhere has taken action on AI training on customer content. Fact The 2024 terms-of-service firestorm was a drafting-and-communication failure - the disputed terms expressly disclaimed ownership of customer content and contained no grant of AI-training rights. Fact
The revised terms converted an implicit absence into an explicit prohibition. It took public pressure, and the resulting covenant is stronger than the industry norm.
Preserved: the narrower concerns
Individual-account content analysis is opt-out rather than opt-in, while business and enterprise accounts are automatically opted out - Adobe applied the protective default to customers with legal departments and the permissive default to individuals. Fact Content analysis powers auto-tagging and similar features; conflating it with generative-AI training would be inaccurate, and this page does not. Fact Separately, partner-model workflows send the customer's prompt text and reference content to external providers, where the selected partner's own terms govern; Adobe's no-training covenant does not necessarily govern third-party providers identically, and the gap is disclosed in a fact-sheet PDF rather than the Terms of Use. Strongly supported
The strongest ethical grievance belongs to Adobe Stock contributors, whose work was the training corpus, compensated by a discretionary bonus with no meaningful opt-out - and Firefly's "ethically sourced" training set was reported to include roughly 5% AI-generated images from rivals.
Pattern verdicts · Final state
What the evidence actually supports.
Confirmed
Strongly supported
Partial / narrow
Inference only
Not established / rejected
Inference is not promoted into fact anywhere on this page. Where the dossier records UNKNOWN, this page preserves the uncertainty rather than resolving it in either direction.
Where Adobe gets it right
The evidence behind the CVI 83.
This section carries equal weight to the patterns above, not a footnote. Positive evidence receives the same evidentiary seriousness as criticism, and several of these items are independently sufficient to defeat the most common accusations.
Flagship value: Creative Cloud All Apps/Pro is roughly 2-4% cheaper in constant dollars than at its 2012 launch, the annual-prepaid option about 9% cheaper, Creative Cloud Standard about 23% cheaper than the 2013 flagship in real terms, and Single App about 20% cheaper.
A materially larger bundle: 20+ applications against roughly 14-15 in CS6, now including Acrobat Pro (a separate ~$449 purchase in 2012), the full iPad and mobile line, Adobe Express, Fresco, Aero, Premiere Rush, Character Animator, Firefly, cloud storage, collaboration and Adobe Fonts.
A dramatically lower upfront barrier: CS6 Master Collection cost about $3,780 in June 2026 dollars; a student today reaches the same toolset for $240 in year one. Subscription made professional Adobe software accessible to people who could never have afforded Creative Suite outright.
Eleven years of Photography-plan price stability, with the annual-prepaid 20GB price never raised - still $119.88, about 30% cheaper in real terms than in 2013 - and when Adobe raised the monthly variant in 2025, it publicly told customers how to avoid the increase by switching to annual prepay.
Continuous product improvement instead of an 18-24 month wait, on the corrected chronology: Lightroom Super Resolution (June 2021), Lightroom AI masking (October 2021), Select Objects and Select People (October 2022), Lightroom AI Denoise (April 2023), Premiere Text-Based Editing (May 2023), Premiere Enhance Speech (February 2024).
Interoperability and egress against short-term self-interest: Adobe publishes the PSD specification explicitly so third parties can read and write the format, donated PDF to ISO as an open standard, and preserves meaningful file egress in important workflows, including Lightroom Classic's retained Export, Print and publishing after cancellation.
A generous cloud grace period by industry standard: Lightroom cloud originals are retained for roughly one year after lapse, with a further 30-day email warning before deletion, and a dedicated downloader tool.
Content authenticity: Adobe co-founded the C2PA standards effort in February 2021, and Content Credentials are applied automatically to Firefly output.
AI contractual protection: following the 2024 controversy, Adobe converted its customer-content no-training commitment into contractual language, closed the sublicence loophole, and made the local-device exclusion explicit - terms stronger than the industry norm.
Firefly IP indemnification on qualifying plans, underpinned by licensed training-data provenance - a legally meaningful, competitor-differentiating benefit, noted with its limit: it is framed as an enterprise benefit, so most individual subscribers do not receive it.
Substantial free offerings at real scale: Adobe Express free for K-12 schools, students and educators with no paywalls or ads; Firefly and Express free tiers; Acrobat Reader; creative freemium monthly active users past 90 million in FY2026, up about 70% year over year.
Cancellation is easy at the mechanical level today - a documented six-step self-service flow with a 14-day full-refund window on most plans - and Adobe deferred planned FY2026 pricing actions, pivoting to freemium acquisition.
Adobe exhibits meaningful customer-hostile mechanisms, particularly around customer control and dependency, while simultaneously delivering unusually high customer value. CHI and CVI are scored separately, on purpose: CVI 83/100 sits well above CHI 50/100, and CFS +33 reflects that gap directly. High CVI does not erase CHI, and documented hostility patterns do not suppress legitimate Adobe value.
What remains unresolved
Open research questions, not published claims.
Claims explicitly rejected
CHI is not simply compiling complaints.
This section is the product of adversarial research, including two claims this investigation actively expected to confirm and could not. Rejecting them strengthens the findings that survive.
What Adobe didn't do alone
- The subscription transition genuinely hurt Adobe before it helped: FY2013 revenue fell 7.9%, GAAP operating margin collapsed from 26.8% to 9.5%, and GAAP net income fell about 70%. This was a real bet with real downside, not a costless money grab.
- Generative-AI metering is universal in this market, no major vendor rolls credits over, and inference genuinely costs money - the scored hostility is in the entitlement withdrawal, the throttle-to-block switch and the 25-credit floor, not in the existence of limits.
- The industry converged on subscriptions: Apple introduced a creative subscription in 2026 and made Pixelmator Pro for iPad subscription-only, Capture One went subscription-first, and Affinity abandoned paid perpetual licensing entirely. It did not converge on Adobe's specific terms: no competitor in the control group was found to charge an early termination fee, and two profitable firms still sell professional creative tools outright.
- In June 2026 Adobe deferred planned further pricing actions and pivoted to freemium acquisition - evidence of constrained pricing power that cuts against a thesis of unchecked annual extraction, weighted as a postponement rather than an abandonment.
These concessions strengthen rather than weaken the assessment: CHI scores Adobe's particular implementation choices - the annual-commitment-as-monthly-price architecture, the default-to-Pro migration, the credit allowance left unenforced until it was load-bearing - against the realistic alternative of the same commercial strategy executed with honest defaults and disclosure, not against an idealized company with no commercial pressure at all.
Methodology note
This page is built from the corrected Adobe Customer Hostility Index Evidence Dossier, revision 2 (compiled 8 August 2026), under CHI Methodology v2.0. Evidence labels - Fact, Strongly supported, Adobe claim, Customer report, Analytical inference, Allegation, Regulatory, Unknown - preserve the dossier's evidentiary distinctions.
The dossier applies a tiered source standard: Adobe filings, legal and help documentation, earnings-call transcripts verified verbatim, government complaints and settlements, and BLS CPI data at Tier 1; contemporary trade reporting at Tier 2; forums and complaint aggregators at Tier 3, never as sole proof of a material allegation. Where evidence was insufficient the dossier states UNKNOWN rather than inferring - including in instances where the unsupported claim would have favoured a hostility finding. Settlement allegations are recorded as allegations; Adobe denies wrongdoing and no liability was adjudicated.
CHI 50/100 (Concerning), CVI 83/100 (Exceptional) and CFS +33 (Customer-Friendly) are the settled analytical values for this assessment and are not presented as a range. Current pricing, allowances and plan details should be reconfirmed immediately before any republication.
Final finding
Adobe is why CHI and CVI are separate numbers. The company delivers unusually high customer value: a flagship slightly cheaper in real terms than at launch, a bundle that roughly doubled, a 97% lower barrier to entry, and standards work that cut against its own short-term interest. None of that is diminished by the score above it.
And the hostility is real, concentrated where control lives rather than where price lives: an exit fee architecture that drew a $150 million federal settlement, a default migration to the more expensive tier, an AI entitlement withdrawn after it became load-bearing, and a demonstration in 2019 that installed software can be reached by disputes the customer never sees. Adobe did not become a story of charging customers progressively more for progressively less. It became a story of transferring control.
Adobe charges roughly the same real price for materially more product.
The exchange did not become worse in price. It became worse in control.