Closing a Membership: How to End Something Other People Are Still Paying For

Published August 1, 2026

Part of: Digital Products — our full guide on this topic.

Disclosure: Some links below are affiliate links. If you sign up through them we may earn a commission at no extra cost to you. We only recommend tools we'd genuinely suggest to a friend. See our full disclosure.

Four endings have been written about here already. A platform can close down underneath you. A platform can remove you specifically. You can close the whole business on purpose. And you can take a finished product off sale while the business carries on around it.

This is the fifth, and it is the one where the roles swap all the way over. You are still trading. Nothing has gone wrong with your infrastructure and nobody has removed you from anything. But you have decided to switch off a membership, a subscription, or an ongoing service — and there is a group of people who are using it right now, who have built some part of their month around it, and who are still being charged for it while you work out how to tell them.

In the first two of those articles you are the one on the receiving end of somebody else’s decision. Here you are the platform. Everything that felt unfair when it happened to you is now a thing you are in a position to do well or badly.

The reason it needs its own instructions rather than borrowing from the product-retirement ones is a single structural difference. Retiring a product ends a selling relationship: the buyer already holds the file, and the last thing you owe them is one working download link. Closing a membership ends a service relationship that is still live. The thing you are switching off is something they have not finished using — and, unlike every other ending on that list, the money is still moving while you decide.

The quick version

• Turn renewals off first, before the announcement, before the join page comes down, before you tell anyone. Then the email can say you will not be charged again as a fact.
• Unpublishing the join page cancels nothing. New signups and existing subscriptions are separate things, and only one of them stops when the page goes.
• The closing date is usually discovered, not chosen — it is the day the last already-paid period runs out. Find it in your billing tool.
• Delay is not neutral here. Every week you sit on the decision is a week somebody pays for a thing you have already decided to end.
• Annual and lifetime members are a different case. Pro-rating solves the first. Nothing pro-rates the second.
• Hand over the archive permanently. It is done, it costs you nothing, and it changes the whole tone of the ending.
• Tell paying members before anybody public. Learning from a public post that the thing you pay for is closing teaches two things, and the second one stings.
• Do not run a last-chance join campaign into something you have decided to close.
• Pause and shrink are real options, and a pause only counts if you stop billing during it.

First, the boundary: four endings this is not

You are not closing the business. If everything is stopping — the shop, the site, the payment account, the domain — the sequence is different and larger, and closing down an online business covers it. Here the business continues; one offer inside it ends. That difference is what makes several of the instructions below possible at all, because you still have a working inbox, a live payment account and somewhere to send people.

You are not retiring a product. Retiring a digital product is a clean ending because the buyers already have the thing. Your only real obligation is to make sure everyone who paid has a copy before the listing goes. A membership has no equivalent moment — there is no file the member already holds that makes them whole.

They are not the ones cancelling. Handling subscription cancellations is about the member deciding to leave, which is normal, expected and largely their business. This is the exact inversion: you are cancelling all of them at once, and none of them asked.

It is not ending one client engagement. Ending a client project is a one-to-one ending with a contract, a scope and usually a notice clause that tells you what to do. Here you have a group, no individual notice period, and terms you probably wrote yourself and have not read since.

The clock is running, and it is running on their money

This is the part with no equivalent anywhere else, so it is worth stating bluntly.

From the moment you decide to close, every renewal that goes through is a charge for something you already know is ending. That is not fraud and it is not even unusual — most people take weeks to work out how to announce a closure, and the billing carries on in the background the entire time. But it is worth seeing clearly, because it changes what “I’ll deal with it next month” costs. In every other ending, delay costs you nothing but tidiness. Here it is somebody else’s money.

Which produces the first rule, and it is the whole reason the order of operations below is what it is: the decision and the switch-off should be close together, even if the announcement takes you longer to write. Turning renewals off is a five-minute administrative act. Writing an honest email to people who trusted you can reasonably take a couple of days. Do not let the second one hold up the first.

Turn renewals off before you announce

The instinct is to announce first — get the hard part over with, then handle the mechanics. Do it the other way around, for three reasons that are all about what the announcement does to a live billing system.

It removes the race. If members find out while their subscriptions are still active, the rational move for every one of them is to cancel immediately, before the next charge. That is a fortnight of people fighting your checkout to avoid giving you money you were not going to keep anyway.

It removes the accidental charges. Some members will not read the email for a week. Some renew tomorrow. Announce first and a handful of people get billed after being told the thing is closing, which is the single most avoidable bad moment in this entire process — and each one becomes a refund, a complaint, or in the worst case a chargeback that costs more to handle than the payment was worth.

It gives you the sentence that does the most work. Your subscription has already been stopped and you will not be charged again. If renewals are off before you write, that is a fact. If they are not, it is a promise you are asking somebody to trust from a person who has just told them bad news.

One caution, and it is the reason to verify rather than assume: turning off renewals and cancelling subscriptions outright are not always the same operation, and what each one does to a member’s access before the paid period ends differs between billing systems. You want the version where they stop being charged and keep their access until the period they paid for is over. Check what your own tool actually does — ideally on one subscription first, not all of them — because the version that cuts access immediately turns a courteous closure into people losing a month they already paid for.

Unpublishing the join page does not close anything

This is the mechanical trap, and it catches people who did everything else thoughtfully.

Taking the sales page down, hiding the join button, or unpublishing the product removes the entrance. In most systems, an existing subscription is a separate object from the listing that created it, and it does not know or care that the listing is gone. It has a schedule, a payment method and an amount, and it will keep running to that schedule indefinitely.

So the join page coming down is not a step towards closing. It is a step towards not making the problem bigger. Closing is a separate action performed on the subscriptions themselves, one that you have to go and find. This is the same failure that appears in the whole-business version — where it is one item on a long list — but here it is not a footnote. It is the main event, and the entire close consists of it.

Two things worth doing while you are in there. Count the active subscriptions, because the number is often not what you think — the figure people carry in their head is usually the one from the last time they looked at their membership numbers, and it drifts. And export the member list before you cancel anything, for the same reason you would keep a copy of your buyer records generally: you are about to need to email exactly this group, and the list is currently living inside the system you are about to start switching things off in.

The closing date is discovered, not chosen

Almost everybody picks a date first — the end of the month, the end of the quarter, a Friday — and then works out afterwards who that date treats badly. Do it in the other order.

Once renewals are off, nobody is paying you for anything beyond the period they are currently in. If you keep delivering until those periods expire, then by definition no member has paid for something they did not receive. So the closing date falls out of the data: look at the active subscriptions, find the latest date any current paid period ends, and close on that day.

This turns the hardest question in the whole process — who do I owe money to — into arithmetic. Usually the answer is nobody.

Two adjustments to that.

Monthly members are easy; the last one is at most a month away. If everyone is monthly, you have your date within about thirty days of the announcement, and the wind-down is short enough that it will not exhaust you.

Annual and long-term members can drag the date out absurdly. Somebody who renewed a twelve-month plan last week would have you delivering for another year. Nobody sensibly does that, and you do not have to: refund the unused part of what they paid and close on the earlier date. Work out how much of their term is left, return that share, and say so in their email specifically rather than in the general announcement. It is a small number of people, it is a small amount of money, and it is the difference between a fair closure and one that quietly relies on people not doing the sums.

If the numbers make refunding everyone with time remaining genuinely painful, that is real information rather than an excuse — and it is worth reading as feedback on annual plans generally, which are a loan from your future self that this moment calls in.

”Lifetime” is the case that will not pro-rate

If you ever sold lifetime access — to the membership, to the community, to “everything I ever publish” — closing ends a promise that had no end date written into it. There is no unused portion to calculate, because the term was open.

This deserves naming rather than hoping nobody raises it, because the people who bought lifetime access are, almost by definition, the members who backed you hardest and paid most up front. Three honest positions exist, and the wrong one is the fourth:

• Refund what they paid. Cleanest, most expensive, and completely defensible.
• Give them something of genuine equivalent value — the full archive permanently, plus whatever else you sell, or whatever the closest honest equivalent is.
• Explain plainly that you are doing neither, and why. This is a legitimate position if the money is long spent and the value delivered was real, but it has to be said out loud, in the email, rather than left for them to notice.

What is not available is treating a lifetime member as an ordinary member and letting the word quietly stop meaning anything. People are far more forgiving of a clearly-stated ending than of a promise that just evaporates — which is exactly why “for as long as this runs” is a better phrase to put on a sales page than “lifetime”, and worth remembering the next time you price a membership.

Hand over the archive, because it is free and it changes everything

Nearly every membership is two products wearing one price tag: a library that already exists, and something recurring that only happens while you are running it — the new drop, the monthly call, the answers, the group.

The library can usually be handed over permanently. Zip it, export it, put it somewhere durable, and give every member a copy to keep, before the closing date rather than as a promise for afterwards. This is the highest-return action available in the entire close:

• The work is already done, so it costs you almost nothing.
• It converts the member’s experience from losing something I was paying for into keeping something I was paying for, which is a genuinely different feeling.
• It removes most of the refund pressure, because the complaint that drives refund requests here is not usually about the money, it is about being left with nothing.

Be straight about the half you cannot hand over. The recurring part is the part they were actually subscribed for, and pretending a folder of files replaces it is the one dishonest note available in an otherwise honest process. Say what they are keeping and say what is ending.

A practical note on the files themselves: whatever format you hand over should work without you. Links to videos hosted on an account you are closing, documents that live in a workspace you are about to cancel, and anything that requires logging into the membership platform are all things that will break shortly after you stop paying attention to them. Hand over the artefacts, not the addresses.

The community is a different object from the list

If the membership had a group attached — a forum, a chat server, a private space of any kind — it is usually the part people will actually miss, and it is the part with the least obvious answer.

Three routes, all honest:

Close it with the membership. Simplest and entirely defensible. Give a date, say it plainly, and let people exchange contact details in the last week if they want to — that final thread is often the most valuable one the group ever has.

Leave it running, unmoderated, and say so. Only if you genuinely will not be back, and only if you are honest that nobody is looking after it. An unmoderated space with your name on it is still a space with your name on it.

Hand it to a member who wants it. This is real and sometimes lovely, but only where the group lives somewhere that can genuinely change hands, and only with the members told before it happens, not after.

What does not work in any of those routes is treating the membership’s email list as a thing you can pass to somebody else. Those people gave you permission to contact them, and that permission is not an asset that travels — the same rule that applies when closing a business entirely. Recommending somewhere in your final email is a recommendation, and yours to make. Handing over the addresses is a transfer, and it is not.

Telling people, in the order that is not the instinct

The instinct is to announce publicly — it feels like getting it over with. Do it last.

1. Anyone still being charged, which after the steps above should be nobody, but check.
2. Lifetime and annual members, individually, with their specific situation addressed. This is a small enough group to write to properly, and they are the ones with the most at stake.
3. All current members, with the closing date, the archive, and the statement that no further charge will be taken.
4. Anyone who left recently but whose access has not expired, if your system has such a state. They are easy to miss and they will notice.
5. Your wider email list, who may have been considering joining.
6. The public — the sales page, the social profiles, the article that says “join my membership”.

A member who finds out from a public post that the thing they pay for is closing has learned two things: that it is closing, and that they were not important enough to tell first. The second one is the one they remember.

The announcement email itself does not need to be long, and it should not read like marketing. What it must contain: what is ending, when, that they will not be charged again, what they are keeping, and — if you want to say it — one honest line about why. “I am not able to keep this to the standard it needs” is a sentence that does you credit. Extended explanation, apology or self-criticism is not needed and tends to make people feel they have to reassure you.

Do not run a last-chance push

Selling new memberships into something you have already decided to close is the one action here that is straightforwardly wrong, and it is tempting for an obvious reason: closure announcements generate urgency, and urgency sells. A “last chance to join before it closes” campaign converts, and every person it converts is somebody buying a promise of next month from a business that has decided there will not be one.

There is a legitimate version nearby, and the distinction is worth being precise about. Selling the archive afterwards as a one-off product is fine. The buyer is purchasing a finished thing they can see the boundaries of, not a subscription to a future. That is an honest transaction and it is a genuinely good afterlife for the work.

The afterlife: a membership makes an excellent product

You have just packaged the archive for members. That package is, with very little further effort, a product.

• Sell it as a one-off. Everything about selling a finished digital product applies, and the material is already made, already used and already improved by member feedback.
• Split it into a bundle or several smaller products. A three-year membership archive is rarely one coherent thing, and bundling the parts that belong together usually beats one enormous download.
• Turn the best of it into public content. Repurposing it into articles puts work that was locked behind a login in front of people permanently, and keeps sending traffic to whatever you sell now.

Keep the editable source files, not just the exports. A membership archive you can only ever republish as flat PDFs is one you can never revise, split or revive.

When not to close

Four situations where the instinct is wrong, and one of them is the most common reason people close a membership that did not need closing.

You are exhausted, not finished. A membership that has become a weekly obligation you dread is a schedule problem before it is a business problem. Before closing, ask what it would look like at half the frequency. This is the most common misdiagnosis in the whole category, and it looks identical from the inside to a business decision — burnout is very good at presenting itself as strategy.

You could shrink the promise instead. Weekly to monthly, live calls to written answers, four things a month to one good one — with the price adjusted honestly to match. Members who value the thing will usually take a smaller version over losing it, and you can say exactly that in the email and let them decide. Note that lowering a price is far easier than raising one, so this is a change you can actually make. Do read it as its own job rather than a softer version of this one: shrinking the promise is changing something people already paid for, and what you owe them is decided by what the sales page said you would deliver, not by the fact that the alternative was closing altogether.

A pause would do — if you actually stop billing. A stated break with a return date is a legitimate offer. A break during which members keep paying is not a pause, it is the thing this article is trying to help you avoid doing accidentally. And do not announce a return date you do not believe in; closing honestly beats a comeback you owe people and dread.

Somebody else could run it. Rarer, but handing over delivery — with the members told, and the money arranged — is a real option that people forget exists because it does not look like either quitting or continuing.

The order to do it in

1. Decide, then immediately turn off renewals. Do not wait until the announcement is written.
2. Verify in the billing tool that renewals are genuinely off and that access survives to the end of each paid period.
3. Export the member list while everything is still running.
4. Unpublish the join page and remove the buy buttons, so no one new arrives.
5. Find the closing date — the latest end date across current paid periods — and decide whether to refund any long terms instead of running them out.
6. Package the archive in a format that works without you.
7. Write to lifetime and annual members individually, with their specific answer.
8. Announce to all members: date, archive, no further charges, one honest line on why.
9. Deliver until the date. The wind-down is short by construction, and finishing it properly is most of what people will remember.
10. Send the archive, before the date rather than after it.
11. Close the community, or hand it over, having told people first.
12. Fix your own site: the sales page, internal links, and anything that still invites people to join. The URL question is the same one that applies to a retired product page — replace or redirect, never delete.
13. Audit your email automations. Any sequence that sold, welcomed or onboarded members is still sending, and closing a membership is exactly the kind of event that leaves an automated welcome email cheerfully greeting people into something that no longer exists.
14. Check your bio links and profiles. Social accounts you no longer post on are the most likely place a join link is still live.
15. Test the checkout you left behind. If the shop is still trading, run through it yourself to make sure nothing you switched off broke something you meant to keep.

The bottom line

Closing a membership is not a bigger version of taking a product off sale. It is the only ending on this site where you are switching off something people are actively using, on money that is still moving, and where you — not a platform, not a payment processor, not bad luck — are the one making it happen.

Which means the standard is different, and it is a standard you already know, because it is the one you would want applied to you in the article where the platform closes underneath you. Give people notice, stop taking their money the moment you decide, let them keep what can be kept, and tell them before you tell anyone else.

Almost none of that is expensive. The archive is already made, the refunds are usually zero if you get the order right, and the email takes an afternoon. What it costs is doing it deliberately instead of letting a subscription quietly bill people while you work out how to say it.