Working out what you pay for is the easy half. Deciding what to drop is where people stall — usually because they're asking a question that keeps everything.
The short answer
Sort by yearly cost, not monthly. Then apply one test per item — have I used this in the last 90 days? — and for anything you're unsure about, cancel it and find out whether you miss it. Resubscribing is cheap; a year of not-quite-using something isn't.
Rank by yearly cost first
Monthly prices flatten everything. £4.99, £9.99 and £12.99 all read as "a few pounds", so nothing stands out and you end up deciding by vibe. Convert to yearly and the list reorders itself — see calculating your real spend for the arithmetic.
The reordering is usually the whole insight. The most expensive line is very often not the one you'd have named.
The 90-day test
One question, asked honestly, per subscription: have I actually used this in the last 90 days?
Not "do I like having it". Not "might I use it". Those questions have no discriminating power — they keep everything, which is why people who ask them cancel nothing.
| Used in 90 days? | Yearly cost | Do this |
|---|---|---|
| No | High | Cancel now. This is the easy money. |
| No | Low | Cancel. Small ones are how the total creeps up. |
| Yes, rarely | High | Work out the cost per use — see below. |
| Yes, often | Any | Keep it. This is what the money is for. |
Cost per use, for the awkward middle
The genuinely hard cases are things you do use, just not much. Divide the yearly cost by the number of times you used it in a year.
- A £120/year service used twice is £60 a use. Would you pay £60 for that, each time, if it were priced that way? Usually not.
- A £120/year service used weekly is £2.30 a use, which is obviously fine.
This works because it converts a vague feeling into a price you can react to. Most people know immediately whether £60 a use is acceptable.
Ask whether you could just buy it when you need it
A lot of subscriptions replaced a thing you used to buy occasionally. If you use something four times a year, buying it four times might be cheaper than subscribing — and it moves the decision to the moment of need, where you'll judge it accurately.
When you can't decide, cancel and see
This is the most useful move on the page and the least used. If you genuinely can't tell whether something is worth it, cancel it and find out. Almost every service lets you come back, often with a returning-customer offer.
The asymmetry is stark: getting it wrong costs you a few minutes of resubscribing. Not deciding costs you a year of payments. People avoid this because cancelling feels final — it isn't.
Two traps worth naming
- Sunk cost. "I've paid for this for two years" is an argument for stopping, not continuing. The money is gone either way.
- Retention offers. Half price for three months is a real saving on something you want, and a way of keeping you on something you don't. If you'd already decided to cancel, the discount hasn't changed how much you use it.
What to do with the ones you keep
Cancelling is only half of it. For everything that survives the cull, put the renewal date somewhere that will interrupt you — otherwise you'll be doing this whole exercise again in a year, from scratch, having paid for another twelve months of whatever you meant to look at.
Annual plans especially: give yourself about a week's warning so the next decision is a real one. See tracking annual renewals.
Then don't lose the ones you kept
TrialBell tracks what survived the cull, works out each cancel-by date, and reminds you before the next renewal — so the decision comes round on purpose rather than by surprise. Free for 3 subscriptions; Pro is a one-time $19.99.
Coming soon to theApp StoreFrequently asked
How do I decide which subscriptions to cancel?
Rank everything by yearly cost, then ask one question per item: have I used this in the last 90 days? Anything expensive and unused goes first. For the ones you're unsure about, cancel and see whether you miss it — you can almost always resubscribe, and resubscribing is a much smaller mistake than paying for a year of nothing.
Should I cancel subscriptions I might use later?
Usually yes. "Might use later" is the reasoning that keeps everything, because it applies to almost anything. Most services let you resubscribe freely, so the real cost of cancelling something you later want back is a few minutes — against twelve months of paying for something you didn't use.
What's a good rule for how much to spend on subscriptions?
There isn't a credible universal number, and anyone quoting one is guessing. A more useful test is per-subscription: divide the yearly cost by how many times you actually used it last year. A £120 service used twice cost £60 a use — that framing decides it faster than any budget percentage.
Written by the TrialBell team and updated 29 July 2026.