Dropbox announced that on 1 January 2027 its terms of service would require users to be at least eighteen years old, reduce the period of inactivity after which a free account may be terminated from twelve months to six months, state that if one account linked to an email address is banned then other accounts using the same address may also be banned, limit refunds to cases where the law obliges the company to pay, and treat the mere continued possession of an account as acceptance of the new terms, a shift from the previous rule that acceptance required continued use of the service. These changes illustrate a recurring arrangement in which a service provider extracts assent from users not by asking for fresh agreement but by treating the user’s continued possession of the account, data, or access as sufficient consent, even when the user faces substantial costs to abandon the service.
The provider benefits because it can alter obligations, fees, or data practices without negotiating each change with every user. The user, meanwhile, incurs switching costs that include the loss of stored files, the disruption of collaborative workflows, the need to re‑establish sharing links, and the effort to migrate to a competing service that may offer inferior features or higher prices. Because the cost of exit exceeds the perceived cost of enduring the new terms, the user remains in possession of the account. The provider then interprets that continued possession as consent, closing the feedback loop that would otherwise allow users to reject unfavorable changes by leaving. The mechanism hinges on two asymmetries: the provider holds unilateral power to modify the contract, and the user lacks viable alternatives that preserve the same level of service, data integrity, or network connections. When these asymmetries coexist, the act of staying put becomes a coerced signal of approval rather than a genuine choice.
This pattern appears first in medieval craft guilds. Guild masters could raise apprenticeship fees or tighten admission rules while apprentices remained bound by the time already invested in their training and the reputation attached to the guild’s mark. Leaving the guild meant forfeiting the right to practice the trade in the town, losing access to guild‑protected markets, and facing social stigma. Masters therefore relied on the apprentice’s continued presence in the workshop as tacit agreement to the new conditions, even though the apprentice had no opportunity to renegotiate the terms. The guild’s control over the credential that certified skill functioned as the lock‑in that turned possession of the apprenticeship into forced assent to whatever fees or conduct rules the masters imposed.
A second example emerges from the railway industry in the United States during the late nineteenth century. After the federal government granted railroads extensive rights of way and limited competition, carriers could adjust freight rates with little effective oversight. Shippers who had built warehouses, sidings, and long‑term contracts around a particular line faced prohibitive costs to shift their goods to another carrier: constructing new sidings, renegotiating contracts with suppliers and customers, and absorbing potential delays during the transition. Railroads exploited this lock‑in by issuing rate increases that shippers had to accept or risk losing access to the market altogether. The continued use of a railroad’s tracks, despite the higher charge, was taken by the carrier as acceptance of the new tariff, mirroring the way a digital service treats continued account possession as consent to revised terms.
In the twentieth century, the evolution of end‑user license agreements for personal computer software provides a close analogue. Early software licenses were often perpetual and allowed users to run the program on any machine they owned. Beginning in the 1980s, firms such as Microsoft began to issue licenses that restricted the number of installations, required periodic renewal, or prohibited reverse engineering. Users who had already invested in learning the software, created files in its proprietary format, and built workflows around its features faced significant retraining and data‑conversion expenses to switch to an alternative product. The vendor therefore relied on the user’s continued possession of the installed software — evidenced by the presence of the program on the hard drive and the existence of compatible files — as implicit consent to the more restrictive license, even though the user had not been presented with a fresh choice at the moment the terms changed.
Telecommunications offer a further illustration. After the 1984 divestiture of AT&T, regional Bell operating companies gained the ability to modify long‑distance pricing and service bundles. Customers who had installed dedicated telephone equipment, integrated their private branch exchanges with the carrier’s network, and established calling patterns that depended on specific dialing prefixes encountered substantial rewiring and retraining costs to move to another provider. The carriers treated the ongoing use of their lines — indicated by the continued presence of the customer’s equipment connected to the network — as assent to new rate schedules or usage caps, despite the absence of a explicit renegotiation process.
Residential tenancy shows a parallel dynamic. Landlords may increase rent or alter lease conditions while tenants remain in the unit because moving entails the expense of hiring movers, paying application fees, undergoing credit checks, and enduring the disruption of finding comparable housing in a tight market. In many jurisdictions, a tenant who remains in possession of the premises after the lease term expires is held to have accepted a month‑to‑month tenancy under the same terms, unless the landlord provides notice of change. The tenant’s continued occupancy, therefore, functions as a proxy for agreement to whatever adjustments the landlord imposes, even when the tenant would prefer to negotiate or leave if the cost of exit were lower.
Across these cases the underlying causal chain is identical. First, the provider possesses the technical or legal ability to amend the governing rules without obtaining a fresh, explicit agreement from each user. Second, the user accumulates sunk costs — whether in the form of stored data, specialized equipment, learned skills, or social capital — that make exit costly. Third, the user’s continued possession of the service, account, or asset is interpreted by the provider as acceptance of the amended rules, thereby bypassing the need for a deliberate opt‑in. Fourth, because the cost of exit outweighs the cost of staying, the user remains in place, reinforcing the provider’s interpretation of possession as consent. The mechanism does not depend on any particular technology, legal regime, or historical period; it depends solely on the conjunction of unilateral amendment power and high switching costs that turn mere retention into a compelled signal of approval.
The signal from Dropbox’s 2027 terms of service change captures this mechanism in a contemporary digital setting: the age‑based access restriction, the shortened inactivity window, the collective ban on linked accounts, the narrowing of refund eligibility, and the rule that mere account retention constitutes acceptance all rely on the user’s inability to readily migrate their files, shared links, and collaborative history without incurring substantial loss. The provider’s invocation of age signals from app stores further illustrates how the service can infer user characteristics from the very act of possessing the account, using that inference to enforce a restriction that the user cannot easily contest without surrendering the account.
When the same logic is transferred to other domains — guild apprenticeship, railroad freight rates, software licensing, telecom pricing, residential leasing — the outcome is uniformly a shift of power toward the party that can alter the rules while the counterparty is locked in by non‑trivial exit costs. The historical precedents are not metaphorical parallels; they are concrete instances where the identical asymmetry of contractual amendment power and exit cost produced the same pattern of forced assent through continued possession.
The persistence of this arrangement reveals that any system which allows one side to change the terms of exchange while the other side’s departure incurs significant losses will inevitably generate consent that is extracted from inertia rather than deliberation. The only way to break the cycle is to reduce the asymmetry — either by limiting the provider’s ability to amend terms unilaterally or by lowering the user’s cost of exit — but as long as the lock‑in persists, the mechanism will continue to turn mere possession into forced assent.