The subscriptions you forgot are the easiest savings you have

A subscription is designed to be forgotten. That is the business model, not a side effect, and the research on it is blunt about how well it works.

The gap between the guess and the statement

In a 2022 survey of a thousand American consumers, people were first asked to estimate their monthly subscription spend, then walked through the categories one by one. The average guess was $86; the average total was $219, two and a half times as much. Streaming, cloud storage, apps, deliveries, memberships: each one small, each one billed to a card nobody reads line by line.

What the card data shows

Einav, Klopack and Mahoney went past surveys to the payment records themselves. They looked at what happens to subscription renewals when a customer's card is replaced, which forces the merchant to ask for new details, a moment when the subscriber has to actively decide to keep paying. Cancellations spike sharply in exactly those months, and the authors estimate that inattention, people simply not noticing the charge, is worth a substantial share of subscription revenue. The subscribers who cancelled at card replacement were not people who had decided to keep the service. They were people who had never decided anything.

The finding has a mirror image that matters for a household: the forgetting is per-card. If two partners have three cards between them, a subscription can sit on one card for a year without the person who reads the other two ever seeing it.

What a ledger that notices looks like

The defence is not willpower, it is a system that raises the charge at the moment it is worth a decision. Three signals do most of the work: a merchant that charges the same amount at the same interval is a recurring charge whether or not anyone declared it; a recurring charge whose price rose this month is a decision to remake; and a recurring charge with no other sign of use for a month or two is one to question. None of those requires a person to remember anything.

How PeakFinance does it

PeakFinance derives recurring-charge candidates from the ledger itself and asks the household to confirm or dismiss each one, once: a “not a subscription” is stored so the same question is not asked again next month. Confirmed subscriptions get their own tab, an annual cost, and alerts when a price rises or a service goes unused, like a streaming plan charged for 41 days with nothing watched. The annual saving opportunity on the insights screen counts only subscriptions that have gone quiet and price rises nobody agreed to, never the ones the household is plainly using, and it is stated as what dropping them adds to savings over a year, because that is where the money goes. Whose card it sits on is part of the row, so a charge on one partner's card is visible to both.