Why the Small Subscription Is Harder to Cancel Than the Big One

A $50 charge on a card statement tends to get read. A $5 charge tends to get scrolled past, and a charge that is never read is a charge that never has to justify itself. That is the argument of this piece: the smaller the subscription, the fewer times anyone asks whether it should still exist, so small subscriptions outlive large ones for reasons that have nothing to do with how useful they are. The pattern is backwards from how spending feels, and there is reasonable research behind each piece of it.

What the research says about unnoticed recurring charges

None of the studies below measured “$5 versus $50” directly, and I am not going to pretend they did. What they establish is the machinery: people lose track of recurring charges, a recurring charge mostly survives because nobody is prompted to reconsider it, and a price quoted in small periodic units is judged against a different mental yardstick than the same money quoted as a lump sum. Sources were read on 2026-09-21.

Source	What was studied	Finding relevant here	Limits
C+R Research, 2022	Survey of 1,000 US consumers, 22 April – 2 May 2022	First-guess monthly subscription spend averaged $86; after itemising by category it was $219, a $133 gap. 74% said recurring charges are easy to forget; 42% said they had kept paying for something they no longer used; 72% pay by auto-pay	Self-reported survey by a market-research firm, not audited statements
Einav, Klopack & Mahoney, “Selling Subscriptions” (NBER w31547)	Card-transaction data for ten popular US subscription services	In the months a payment card is replaced — which forces the subscriber to actively re-enter it — cancellation rates are much higher than in ordinary months. The authors estimate inattention raises seller revenue by between 14% and more than 200%, depending on the service	The size of the effect varies a lot by service; the paper does not break results out by price
Gourville, “Pennies-a-Day” (Journal of Consumer Research, 1998)	Laboratory studies of how a price is framed	A cost framed as a small ongoing amount leads people to compare it with other small ongoing expenses (a coffee, a newspaper); the same cost framed as an aggregate leads them to compare it with large infrequent purchases	Lab setting, hypothetical transactions

Put together: the first guess in the C+R survey was off by a factor of about two and a half, and it was off in the direction of forgetting charges rather than inventing them. The Einav paper shows that what ends a subscription is often not a change in how much the person values it but an interruption that forces an active choice. And Gourville explains why the small charge is the one that gets waved through: “$4.99 a month” is filed next to a coffee, where it looks harmless, and is never filed next to “$60”, where it would have to compete with things a person actually weighs.

Attention is spent by size

A large charge is a decision; a small charge is background noise. This is not laziness. Nobody can treat every line on a statement as worth a real review, so the lines that look insignificant get skipped. The trouble is that “insignificant per month” is not the same as “insignificant”. The arithmetic is dull and worth doing once:

Monthly price	Per year	Over three years
$2.99	$35.88	$107.64
$4.99	$59.88	$179.64
$9.99	$119.88	$359.64
$50.00	$600.00	$1,800.00

The $50 subscription is obviously the most expensive row. But it is also the row most likely to be cancelled in month three, because every statement puts the question in front of its owner. If it is cancelled after three months it cost $150. A $4.99 subscription that nobody looks at for three years cost $179.64. The small, ignored charge can outspend the big, scrutinised one without ever once feeling expensive — not because of its price, but because of its lifespan.

“I might use it” is cheapest to believe at $5

The third thing keeping the small subscription alive is optionality — the just-in-case subscription in its purest form. A big subscription has to keep proving its worth. A small one survives on the vague sense that cancelling now would mean signing up again later. At $5, the cost of being wrong about cancelling feels lower than the nuisance of re-subscribing, so the default is to do nothing, and doing nothing is the subscription’s whole business model.

A long-running small account also accumulates something that feels like sunk cost: settings, history, uploads. None of that is a reason to keep paying — what has been spent is spent — but it is a reason the decision feels heavier than the price suggests. It helps to separate the two questions. “Can I export what is in there?” is a practical question with an answer, and what a service keeps after you cancel varies more than most people assume. “Should I keep paying?” is a different question, and the export answer usually removes the only real objection.

Monthly billing is the forgettable kind

A small charge billed monthly hides inside every statement and never triggers a review, while a bigger one billed annually arrives as a single invoice that is hard to miss. This is the same mechanism by which a forgotten free trial bills you anyway: the running total is never presented all at once. With 72% of the C+R respondents on auto-pay, there is no monthly moment at which anything has to be approved.

Then there is the cancel flow. Log in, find the account page, click through a retention offer, confirm a second time, sometimes email a human. Each step is a chance to give up, and on a small charge giving up is cheap: the reward for finishing is five dollars. The friction is the same size whether the plan costs $5 or $50, which means that relative to the saving it is ten times larger on the small one. That asymmetry, more than anything about the product, is why the small plan wins these standoffs.

The reframe: quote it to yourself as a lump sum

Gourville’s result runs in both directions, and that is the useful part. If a small periodic price makes people reach for small comparisons, then restating it as an aggregate makes them reach for large ones. So do not ask “is this worth $5 this month?” Ask: “would I sign up for this today, cold, if it cost $60 upfront for the year?” — the annual version of the cost-per-use question. The per-month price is an anaesthetic; the annual sum is the real number. I would not claim the question settles every case, but it puts the charge in the comparison set where it belongs.

How to run the check

The practical move is an audit done in one sitting rather than a fight conducted monthly, because catching each small charge in real time is exactly the contest that attention-by-size loses.

1. Pull three months of statements for every card and for the app-store accounts (Apple and Google bill subscriptions that never appear by the vendor’s name). Three months catches quarterly billers; a once-a-year search for “annual” and “renewal” in email catches the rest.
2. List everything that recurs, in one column, with the annual figure beside it. The C+R gap — $86 guessed, $219 itemised — is what this step corrects. Individually the small charges are invisible; stacked, they are a number that gets a reaction.
3. Apply the lump-sum question to each row under about $10 a month. Those are the rows that have never been questioned.
4. Cancel in the same sitting. “I’ll cancel it next month” is the sentence the small subscription depends on, because next month it is still five dollars and still invisible. If the vendor makes it a multi-step flow, that is an argument for doing it now while the list is open, not later.
5. Set one calendar reminder to repeat the list. The Einav finding is that subscriptions end when something forces an active decision. A recurring reminder is a way of manufacturing that interruption on purpose instead of waiting for a card to expire.

Where this does not apply

The reframe is not an argument for zero subscriptions. A password manager opened daily, a backup someone relies on, a family plan other people use — those pass the “would I sign up again today at the annual price” test easily, and they should. The goal is only subscriptions that would be chosen again. The small ones are where the dead weight gathers, not because they are worse products but because they are the ones nobody looks at.

One caveat about the evidence: the survey is self-reported, the transaction study does not split its results by price, and the framing study is a lab experiment. The claim that small charges specifically survive longest is my reading of how those three findings combine, not something any one of them measured. It is cheap to test on your own statement, which is the only dataset that matters for this decision.