Subscriptions: the spending you have already agreed to

What a subscription actually is, in payment terms

A subscription is a standing authorisation to take money from you until you withdraw it. One-off spending needs a fresh decision every time; this one needs a decision to stop. That single reversal explains almost everything about how these bills behave.

The word makes it sound like a relationship with a service. In payments terms it is narrower and more useful than that: somewhere there is a mandate — a record saying a named merchant may pull a stated amount from a stated account or card, on a stated cycle, until revoked. The service and the mandate are two different objects, held by two different parties, and they can be separated. Most of the trouble in this article comes from that fact.

So the useful question is never “what am I subscribed to?” It is what am I authorised to be charged for — a list that lives in your bank and card records, not in your memory. The two lists differ, and the difference is the money.

Why the list only ever grows

Three costs decide what happens to a subscription, and they are wildly unequal.

• Starting — one tap, on a screen built and tested to make that tap easy.
• Continuing — nothing. No action, no notice, no decision.
• Reviewing — finding what you pay, working out whether you use it, locating the cancel path, and revoking the mandate. Minutes, on a good day.

Set those three side by side and accumulation is not a character flaw; it is the arithmetic working as designed. The only step that removes a subscription is the expensive one, and nothing triggers it.

Now put a number on the first tap, because the number displayed is not the one you agreed to. Take a ₹199 monthly plan. On screen the comparison is ₹199 against one month of the thing. But nothing in the arrangement asks you again, so what you actually committed is ₹199 multiplied by however many months pass before you next think about it. Three years of not thinking about it is ₹7,164 — thirty-six times the number on the button. Those inputs are illustrative and yours will differ; halve the gap and the point survives.

This is a different failure from spending creeping up as income rises, which has its own mechanism in lifestyle inflation. Here the spending does not creep. It was set once and then simply never stopped.

The five rails an Indian recurring payment can run on

“Cancel the subscription” is one instruction covering at least five different mechanisms, each with the authorisation stored somewhere else. Knowing which rail a charge is on tells you where to go to stop it — and that is the only reason the distinction matters.

Read the last column downwards and one theme repeats: in four of the five cases the authorisation is not kept where you would naturally look. Delete the app and the store keeps billing. Close the account with the merchant and the mandate at your bank survives. Uninstall the payment app and the AutoPay mandate it created carries on.

That gap deserves a name, because it is the single most common way a cancelled subscription keeps charging. Call it the orphan mandate: the service is gone, the authorisation is not, and nothing in the system notices the mismatch. Nobody is being dishonest — the merchant and the mandate are separate records, and stopping one was never going to stop the other.

The one free signal you already receive

Recurring card and account debits in India sit inside a formal framework rather than being left entirely to each merchant to design. Two features of it matter to a household, and both are worth confirming for your own bank and card before you rely on them.

The first is a message sent ahead of a recurring debit telling you what is about to be taken — a pre-debit notification. The second is that above a stated amount, a recurring debit has to be authenticated by you each time rather than only once at registration, which is why a larger renewal sometimes asks for a one-time password and a small one goes through without a prompt. The amount and the notice period are set in the framework rather than by the merchant, and both can be revised, so treat them as things to check for the year you are reading this rather than as fixed facts.

What the notification is for is where most people get no value out of it. Treated as an alert, it is noise: a message about money that is going to leave whatever you do with the message. Treated as a prompt, it is the only per-subscription review trigger anyone ever gives you — it arrives before the money moves, it names the amount, and it comes once per cycle without you having to remember anything.

So the cheap habit is not a monthly audit. It is a single question asked at the moment one of these arrives: did I use this since the last one? Nothing else. The trigger costs nothing and fires on the merchant's schedule rather than on your resolve, which is the property that makes it work.

The honest limit: this only covers the rails inside the framework. A card sitting on file with an overseas merchant, or billing routed through an app store, may behave differently — which is precisely why the audit below reads the statement rather than trusting the alerts.

How to find what you are actually paying for

Start from the wrong end and the exercise fails before it begins. The instinct is to sit down and list your subscriptions from memory, then check the bank for the ones you forgot. That order cannot work, for a reason worth stating plainly.

Work from the record instead. The record does not know which services you like.

1. Pull twelve months, not three. Annual renewals are invisible in a quarterly window and they are usually the expensive ones. A year of statements catches every cycle exactly once.
2. Do it per instrument. Each bank account, each credit card, each debit card. A charge on a card you rarely use is the likeliest one to have been running unexamined.
3. Look for the repeat, not the name. The pattern is the same amount, to the same descriptor, at roughly the same date each cycle. Merchant descriptors often bear no resemblance to the brand you signed up with, so searching for the name you remember will miss things.
4. Then check the mandate lists — the AutoPay screen in your payment app, standing instructions in net banking, the subscriptions screen in each app store. This catches the mandates that have not debited yet, including ones from free trials that have not converted.
5. Reconcile the two lists. Anything on the statement with no matching mandate you recognise, and any mandate with no matching service you use, is the thing this exercise exists to find.

Doing this once a year, rather than resolving to be more careful, is the whole discipline. If you already run a categorised system of expense tracking, the recurring charges will fall out of it as a by-product — and a monthly budget that treats the total as one line called “subscriptions” is hiding the only detail that matters, which is which of them still earns its place.

Judging what you find, without the monthly price fooling you

The list is in front of you. Now the harder half, which is that the monthly price is almost useless as a basis for the decision.

Cost per use is the measure that separates them. A ₹499 monthly plan opened four times in the month costs just under ₹125 a session; opened once, it costs ₹499. Compare that against what the same thing costs one-off — a cinema ticket, a single delivery fee, a one-month plan bought when you actually want it. Where a subscription costs less than buying the thing as you need it, the arithmetic has already settled the question. Where it does not, you are paying for availability rather than use — a real thing to buy, and one worth buying knowingly rather than by default.

Three shapes recur, and they need different questions.

Then the annual-versus-monthly trade-off, which is genuinely two-sided. Twelve months for the price of ten saves you about a sixth of the cost — real money, and the discount is not a trick. What it costs is eleven chances to notice: the monthly plan asks the question twelve times a year, the annual plan once. For something you know you use weekly, that is an excellent trade. For something you are keeping because you might use it, the discount buys a year of not deciding, and the discount is smaller than a year of not using it.

Two more distinctions worth keeping straight. A subscription is not a sinking fund — one is a recurring cost you can stop, the other is saving for a lumpy cost you cannot. And whether a given service is a need or a want is not the question here; plenty of wants survive this test comfortably, and the point of cost per use is to stop that argument being had on feeling alone.

Stopping one, and the traps in the way

Cancelling has two steps, and almost everyone does only the first. A third step is what tells you the other two worked.

1. Cancel at the merchant, so the service knows and the billing relationship ends properly. Keep the confirmation.
2. Revoke the mandate at the rail — the AutoPay screen, the card issuer, the bank, the app store. This is the step that removes the authorisation itself.
3. Check the next cycle’s statement. Not the next day; the date the debit would have fallen. A cancellation is confirmed by an absence, and the absence has a date.

Four traps account for most of the failures.

A free trial is the product itself, not a sample of it. A trial that requires a mandate at signup has already created the authorisation; conversion needs no further action from you and, by design, no further decision. If a trial is worth taking, the mandate revocation date is worth putting in a calendar the same day.

Second, pause is not cancel. A pause keeps the mandate alive and moves the debit, which is fine if you meant to pause and useless if you meant to leave. Merchants offer the pause first because it works — for them.

Third, a card reissue is not a cancellation. Blocking a card, letting it expire, or getting a replacement after a loss may or may not break an existing mandate, and relying on it as an exit is planning around a mechanism you do not control. Revoke the mandate deliberately instead.

Fourth, closing the account with the merchant does not touch the rail. That is the orphan mandate again, and it is the reason step two exists as a separate step rather than as a consequence of step one.

If a debit arrives after all that, the route depends on a distinction worth getting right. A charge under a mandate you did create and then forgot is not an unauthorised transaction — it is a transaction you authorised and stopped watching, and the dispute machinery is not built for it. A charge taken after you revoked the mandate is a different claim, and there the rules on unauthorised electronic transactions come into view: a debit reported as unauthorised is credited back within 10 working days of the report, value-dated to the transaction date, while the bank investigates — the reversal runs ahead of the finding rather than after it. Whether a debit under a revoked mandate qualifies is not something to assume — ask the bank to classify it before you rely on the timeline, and keep the revocation confirmation, because that document is what the classification turns on.

Where it goes wrong

1. Auditing from memory. It finds the subscriptions you use and misses the ones you do not, which inverts the purpose of the exercise.
2. A three-month window. Annual renewals never appear, and they are the largest single charges on the list.
3. Cancelling in the app when the billing runs through an app store or a bank mandate. The service stops; the money does not.
4. Treating the pre-debit notification as noise. It is the only per-subscription review prompt anyone sends you, and it is free.
5. Judging by monthly price. Two plans at ₹499 are not comparable until you know how many times each was opened.
6. Taking the annual discount reflexively. It is a good trade for something you use weekly and a poor one for something you are unsure about, because it removes the eleven moments you would have reconsidered.
7. Signing up for a trial with no revocation date. The mandate exists from day one; the calendar entry is the only thing that expires.
8. Counting the cancellation as a saving without moving the money. An amount freed inside a spending account is spent, not saved. It has to be routed somewhere on the same day, which is the mechanism behind separating spending from reserves.

The pattern underneath all eight is the same one the article opened with. Every mistake here is a place where doing nothing produces a charge, and doing something is the only way to stop it.

What to do with the amount you recover

An audit of this kind produces a specific and unusually reliable figure: a monthly amount that was leaving the account, is no longer needed, and did not come out of anything you were enjoying. It is the cleanest input to how much can go out on payday there is, precisely because nothing had to be given up to find it. It also has a habit of disappearing if it is left in the account it was already sitting in — which is why the amount and the redirection are one decision, made the same day, rather than two.

Whether a recurring amount of that size is worth anything over a long horizon is a question of arithmetic rather than opinion, and it is testable rather than arguable. FNOTrader’s Mutual Funds app runs a contribution schedule against the full published record of daily per-unit prices — the net asset value, or NAV — kept by AMFI, the mutual fund industry body, around 34 million rows of it. It reports invested against value, the worst peak-to-trough fall along the way, and the return measure built for money arriving on irregular dates, XIRR. Past performance is a record of what happened, not an indication of what will happen.

One closing observation, which is really the article in a sentence. The subscription and the monthly investment contribution use the same mandate machinery — a standing authorisation, a fixed date, no decision required at the moment of debit. The mechanism that quietly drains an account is the same mechanism that quietly fills one; the only difference is the direction, and which of the two you set up on purpose. FNOTrader is not a SEBI-registered investment adviser and does not give investment advice.

Common questions

Continue reading

More in Budgeting & Saving · App: Mutual Funds · Definitions: glossary · Free tools: calculators · All: every article