The Sludge That Keeps You Subscribed
Cancelling takes more steps than joining because the extra steps earn money. In a randomised trial with more than two million newspaper readers, roughly half of those who took an auto-renewing trial went on to pay full price.
Answer to the question
Why is it so hard to cancel a subscription?
Because sellers add sludge, friction placed on the way out. A 2023 study of payment-card data estimated that inattention raised subscription revenue by between 14 per cent and more than 200 per cent. A 2014 rebate experiment found that information and reminders did little, while a simpler task halved the gap between intention and action.
Key takeaways
- In a 2018 newspaper experiment with more than two million readers, roughly half of those who took an auto-renewing trial went on to pay full price.
- Between 24 and 36 per cent of would-be subscribers avoided auto-renewal offers, and after a year auto-renewal earned the publisher no more than auto-cancel.
- In a 2014 rebate experiment by Joshua Tasoff, information and reminders made no reliable difference to the gap between intention and action; a simpler form cut it by about half.
- The US click-to-cancel rule finalised in October 2024 was vacated in July 2025; Britain's subscription rules are expected to start in January 2027.
Cancelling a subscription is hard because the way out was built to be hard, and the design works even on people who expect to see through it. The extra steps pay, and the remedy the evidence favours is fewer steps rather than better warnings.
Nudge gives this friction a name, sludge, and a chapter of its own, chapter eight, in the final edition by Richard Thaler and Cass Sunstein. Their founding story is a London newspaper that offered Thaler a one-month trial for a pound, renewing at twenty-seven pounds a month. Cancelling required fourteen days' notice, which the authors point out turned the month into two weeks, and could only be done by telephoning the London office during London business hours. Their definition reaches well beyond subscriptions: sludge is "any aspect of choice architecture consisting of friction that makes it harder for people to obtain an outcome that will make them better off (by their own lights)."
Those last four words do a lot of work. A confirmation screen before a large payment is friction that protects the person moving through it. Placed only on the exit, where the seller loses and the customer gains, that screen becomes sludge, so the same piece of design is protection in one place and a trap in another. A sign-up page with one button and a cancellation route with a survey, two retention offers and a phone number meets the definition on every count.
A confirmation screen before a large payment is friction that protects the person moving through it. Placed only on the exit, that screen becomes sludge.
Auto-renewal turns ordinary inertia into steady revenue for the seller
Auto-renewal runs on inertia, the tendency to leave things as they are that researchers call status quo bias, and the same pull that keeps most people on default settings for years. A large European newspaper let researchers randomise its trial offers, which gives a direct measure of what inertia is worth to a seller. Over one spring and summer, more than two million readers who hit its paywall were offered a trial of two or four weeks, either free or for under a euro. Half the offers rolled into a paid subscription unless cancelled. The other half simply ended unless the reader chose to continue.
Roughly half of the readers who took an auto-renewing trial went on to pay full price, which is inertia doing the seller's work. Many other readers saw the arrangement coming and stayed away from it, so that over the following two years the paper had fewer subscribers under auto-renewal than under auto-cancel. Auto-renewal brought in more money at first. After a year, according to the analysis led by Klaus Miller of HEC Paris, the two contracts were earning about the same.
The researchers also tried to separate readers who were inert from readers who were simply happy to keep paying, and to estimate how many of the inert ones knew it about themselves. Their answer was that fewer than half of the inert readers seemed unaware of their own habits. Most of the inert readers, then, were not fooled about themselves. They knew, and drifted anyway, because knowing does not remove a single step from a cancellation route.
Fewer steps did more than any warning or reminder email
A rebate experiment published in 2014, with Joshua Tasoff of Claremont Graduate University as first author, tested whether knowledge closes the gap between intention and action. Participants were asked how likely they were to send in a mail-in rebate form, and their predictions overstated what they actually did by about 49 percentage points. Three groups then got help. One was shown how a previous group had performed, one received reminders about the deadline, and one was given a simpler form.
Information and reminders reduced the gap by 7 and 8 points, and neither difference was statistically significant. Cutting the cost of redeeming reduced it by 26 points, roughly half. All three kinds of help nudged more people into sending the form, yet none shifted beliefs by a reliable amount. People went on expecting to act, and the one change that made a real difference was making the task easier. The authors of Nudge report the study in the same chapter and add their own gloss: "Telling people that some activity is dangerous… is like informing someone that people often oversleep. It is not much help."
Forcing a decision back on people has a visible effect of its own. A 2023 study of a large payment-card network followed ten subscription products from 2017 to 2021 and looked at the months when a cardholder's card expired and was replaced, which means the subscriber has to act to keep paying. Cancellations rose sharply in those months. Compared with customers who cancel as soon as a service stops being worth its price, the authors estimate that inattention raised sellers' revenue by between 14 per cent and more than 200 per cent, depending on the product.
| Setting | Change tested | Result |
|---|---|---|
| Mail-in rebates, 2014 | Information, reminders, a simpler form | Only the simpler form shrank the gap reliably, by about half |
| Newspaper trials, 2018 | Auto-renewal against auto-cancel | Fewer sign-ups under auto-renewal, equal revenue after a year |
| Card replacement, 2017 to 2021 | A forced fresh decision | Cancellations rose sharply in replacement months |
Across the three settings, a reminder or a warning did less than the number of clicks between a customer and the exit. Before signing up for anything that renews, count the steps on each side of the door. A large difference between the two is a statement of how the seller expects to make money.
Friction has a direction, and that direction shows who benefits
Sunstein turned the idea into a working tool. His 2021 book Sludge argues that paperwork, queues and baffling forms cost people jobs, benefits and care, and his paper "Sludge audits", published online in 2020 in Behavioural Public Policy, proposes that firms and governments regularly catalogue their frictions and remove those that cost more than they deliver. The paper argues that sludge often falls hardest on the most vulnerable. Nudge states the same aim in its preface: "Every organization should create a seek-and-destroy mission for unnecessary sludge" (p. xiv).
An audit has to allow for friction that earns its place. The authors endorse cooling-off periods for decisions people make rarely and while emotions run high (pp. 328 to 330), and they accept that the well-meant checks known in American government as program integrity generate some sludge as a by-product. The test is direction. A pause before a costly, emotional commitment serves the buyer. A pause that appears only when the buyer tries to leave serves the seller.
A pause before a costly, emotional commitment serves the buyer. A pause that appears only when the buyer tries to leave serves the seller.
The newspaper data also test the book's argument, made in the sludge chapter, that competition does not clear sludge away, because "No one opens a checking account intending to bounce checks." Between 24 and 36 per cent of the newspaper's would-be subscribers avoided auto-renewal offers, and after a year the paper's extra revenue had gone. The friction still worked on the people who said yes, and the card data show inattention running highest among customers who take cash advances on their cards. Self-protection works for the attentive and fails the people a sludge audit is meant to protect, which is why removing steps does more than trusting customers to notice them.
Laws on cancelling subscriptions lag behind the research on them
In October 2024 the US Federal Trade Commission finalised a rule requiring that cancelling be at least as easy as signing up. On 8 July 2025, days before it was due to take effect, the Eighth Circuit Court of Appeals vacated it because the commission had skipped a required step in its rulemaking procedure, a preliminary regulatory analysis. The ruling turned on procedure, not on the merits of easy cancellation. In February 2026 the commission restored its older rule, and in March 2026 it published an advance notice asking for comment on whether to try again. No replacement rule had taken effect when that notice appeared.
Britain has written more into law. The Digital Markets, Competition and Consumers Act 2024 contains a subscription regime that includes reminder notices, a 14-day cooling-off window when a trial converts or a contract of a year or more renews, and a requirement that a contract taken out online can be ended online. It is not yet in force. In August 2026 the government said it intends the rules to start in January 2027, with secondary legislation and regulator guidance still to come. Nudge records, in its sludge chapter, that California and New York already require online cancellation for subscriptions begun online.
These are descriptions of where the rules stood at the time of writing, and anyone with a dispute over a particular contract needs a consumer body or a lawyer rather than an article. The research is clearer than the rules themselves. When the route out is short, people leave the services they no longer value.
Common questions
Why is it so hard to cancel a subscription?
Because sellers add sludge, friction placed on the way out. A 2023 study of payment-card data estimated that inattention raised subscription revenue by between 14 per cent and more than 200 per cent, compared with customers who cancel as soon as a service stops being worth its price.
What is a sludge audit?
A regular review in which an organisation lists the frictions it imposes and removes those that cost more than they deliver. Cass Sunstein proposed it in a paper published online in 2020, and Nudge (2021) calls for "a seek-and-destroy mission for unnecessary sludge" in every organisation.
Is the US click-to-cancel rule in force?
No. The Federal Trade Commission finalised it in October 2024, and the Eighth Circuit vacated it on 8 July 2025 on procedural grounds. The commission asked for public comment on a fresh attempt in March 2026. Britain's subscription rules, passed in 2024, are expected to start in January 2027.
Sources
Thaler, R. H., & Sunstein, C. R. (2021). Nudge: The Final Edition. Penguin Books. Read in full by this publication, with a reliability audit. Ch. 8 "#Sludge", pp. 151-176: definition p. 153; the London newspaper trial p. 155; California and New York online cancellation p. 157; the rebate study pp. 158-159; competition p. 161; seek-and-destroy p. xiv; cooling-off periods pp. 328-330; program integrity p. 154.
Miller, K., Sahni, N. S., & Strulov-Shlain, A. (2023). Sophisticated consumers with inertia: long-term implications from a large-scale field experiment. Working paper, version of 27 January 2023 (first circulated as Becker Friedman Institute Working Paper 2022-44). Full text read (January 2023 version, NBER conference copy): 2,092,861 readers of an anonymous large European daily who hit its metered paywall, April to August 2018, follow-up to April 2020; 2x2x2 design (auto-renewal or auto-cancel, two or four weeks, EUR 0 or EUR 0.99); roughly half of auto-renewal takers continued to full pay; 24 to 36 per cent of potential subscribers avoided subscribing in the first weeks after an auto-renewal offer; share of subscribers 10 per cent lower under auto-renewal over two years; revenue equivalent after one year. A later talk abstract (Strulov-Shlain, February 2026) reports different headline figures from a revised version, so this article uses the January 2023 figures only.
Einav, L., Klopack, B., & Mahoney, N. (2023). Selling subscriptions. NBER Working Paper 31547. doi:10.3386/w31547. Abstract read, plus the NBER Digest summary (October 2023): ten subscription products, payment-card network data from August 2017 to December 2021, cancellation rates much higher in months with card replacement, inattention raises seller revenue by 14 per cent to more than 200 per cent relative to fully attentive consumers, higher inattention among cash-advance users.
Tasoff, J., & Letzler, R. (2014). Everyone believes in redemption: nudges and overoptimism in costly task completion. Journal of Economic Behavior & Organization, 107, 107-122. doi:10.1016/j.jebo.2014.08.011. Abstract read: predictions overstated redemption by about 49 percentage points; information and deadline reminders reduced overoptimism by 7 and 8 points, not significant; reducing transaction costs reduced it by 26 points, significant; all three nudges increased redemption but had no statistically significant effect on beliefs.
Sunstein, C. R. (2022). Sludge audits. Behavioural Public Policy, 6(4), 654-673. doi:10.1017/bpp.2019.32. Published online 6 January 2020. Abstract read: sludge as excessive or unjustified frictions; institutions should conduct regular Sludge Audits; sludge can hurt the most vulnerable.
Sunstein, C. R. (2021). Sludge: What Stops Us from Getting Things Done and What to Do about It. MIT Press. Abstract read: publisher's description as given on the Harvard Kennedy School book page.
Federal Trade Commission (2024). Amendments to the Negative Option Rule ("click-to-cancel"), finalised October 2024; vacated by the US Court of Appeals for the Eighth Circuit, 8 July 2025, before the planned 14 July 2025 effective date, for failure to issue a preliminary regulatory analysis; advance notice of proposed rulemaking published in the Federal Register on 13 March 2026, comments due 13 April 2026. Full text read of practitioner summaries (Morgan Lewis, July 2025; Greenberg Traurig, July 2025; Sidley, February 2026; Subscription Insider, March 2026); the opinion and the Federal Register notice were not opened.
Digital Markets, Competition and Consumers Act 2024 (UK), subscription contracts regime. Not in force; on 9 August 2026 the government announced a target start of January 2027, brought forward from spring 2027. Full text read of practitioner summaries (Browne Jacobson, updated August 2026; Stephenson Harwood, 24 August 2026; Latham & Watkins); the Act and the government announcement were not opened.