How to stop customers cancelling and resubscribing for a discount
Cancel and resubscribe discount abuse can empty a campaign in one night. Set eligibility windows and win-back rules, and see which basic fixes break.
Who pays you for 14 months and then cancels without a word? A customer on your $49 plan just did, on a Monday night, with the reason box empty. They've paid you every month, $686 so far, and when you send a short note asking what went wrong, nobody replies.
You don't dwell on it, because the evening is going well. That afternoon you emailed your list a new-customer price: $19 a month for 3 months, 40 spots. At 9:31 the first new subscriber comes through. By 11:40 p.m. all 40 spots are gone, someone has posted the signup chart in the team channel, and you go to bed pleased with the tallest bar you've ever had.
At 8:20 a.m. one of the new accounts writes to support, with an old email address pasted in: "Can you move my projects over from my old account? Same for the other 36. Thanks!"
You open the old account. It's the customer who left last night and said nothing. They run an agency with 37 client accounts on your $49 plan, and while you were watching the signup chart they cancelled each one and bought it back at $19. That's cancel and resubscribe discount abuse, and while it's happening it looks like the best night your company has had.
Nobody in this scenario is real, so swap in your own numbers. The arithmetic goes like this. Each spot gives away $30 a month for 3 months, which is $90, so 40 spots was a $3,600 budget. One person took 37 of them (37 × $90 = $3,330). The money you set aside to win 40 customers won you 3, at $1,200 each ($3,600 ÷ 3).
Your monthly revenue took the hit too. Those 37 accounts were paying $1,813 a month (37 × $49). All 40 new subscribers together pay $760 (40 × $19). Your biggest signup spike ever lowered your monthly revenue by $1,053.
And the reward has already gone out. Every new subscriber gets a 2,000-credit welcome pack when their first payment clears, which costs you about $6 in model usage. That's another $222 (37 × $6) sitting on balances, already paid for.
At 10 a.m. the newsletter placement you booked two weeks ago goes out. It cost $1,200 and it can't be pulled, so every reader who clicks lands on a page saying the offer's gone. Add it up: $3,330 + $222 + $1,200 = $4,752. If you spend $22,000 a month, that's about six and a half days of runway ($4,752 ÷ $22,000 × 30). At 11:15 a customer who has paid full price for two years forwards you the newsletter and asks why strangers get $19 and they don't.
Now you have to choose. Honor the price and you lose the $3,330. Pull the discounts and you're picking a fight with someone who has been paying you $21,756 a year ($1,813 × 12). Open 40 more spots for the newsletter readers and you've committed another $3,600 to a campaign that has brought you 3 customers.
That's the loud version. The quiet one is a single person on a single account who works out that leaving is cheaper than staying, and tells a friend. Say 20 people do it in your first quarter at $90 each, then 45, 80 and 120 as word gets around. That's $1,800, $4,050, $7,200 and $10,800, or $23,850 in a year. It never appears as a loss anywhere, because your dashboard counts every one of them as a new customer.
Why cancelling and resubscribing works
Your offer has a rule in it: new customers only. That rule lives in the small print, and your billing system can't read small print. It sees records. A new email makes a new record, a new record is a new customer, and so the checkout does exactly what it was built to do.
Price differences give people the reason. Whenever newcomers pay less than the people who stayed, the cheapest way to keep using your product is to leave and return. A come-back price emailed to people who left months ago can teach the same lesson, once customers notice that being gone is what earned it.
The obvious fix is to refuse any email you've seen before. It's an afternoon's work and worth doing. If your billing tool limits a code to once per customer, that's the same rule, and it ends in the same place: at the next email address. All 37 of the agency's addresses were new to you.
Write eligibility windows and win-back rules
You can't stop people leaving, and you shouldn't want to punish the ones who come back. What you can do is decide how often each discount goes to the same person, and make the answer hard to game.
Write down who counts as new. "Has never paid us, under any account, as a person or a company" covers the agency. Put it beside the offer in plain words, and back it with a one-account-per-person policy so support has something to point to.
Give every discount a window. A new-customer price is once, ever. A come-back price might be once every 12 months. Choosing between once and a repeating window depends on what the discount costs you.
Set a minimum time away. Someone who left last night hasn't lapsed. Send a come-back price only when the last payment was at least 90 days ago. Going 3 months without your product is a real price to pay for saving $90.
Enforce it on the server, at the moment the discount is applied. Refuse if the account has a live subscription or a claim inside the window, and record the claim in the same transaction.
Pay the reward late. Let welcome credits land after the second invoice. Put the biggest discount in month 3. Someone who cancels after month 1 has then collected very little.
Split the budget by channel. Give your list 20 spots and the newsletter 20, each with its own single-use codes. A bad night on one channel can't empty the other, which also helps when a code leaks.
Put cancellations beside signups. Report new subscribers minus cancellations. In the scenario that's 40 minus 37, or 3, and it's a very different chart to post in the team channel.
The window is what limits your loss, and you can work out the most it could cost before you launch:
At $30 off for 3 months, with someone claiming every quarter, that's $30 × 3 × 4 = $360. They pay $228 a year instead of $588. A 12-month window makes it $30 × 3 × 1 = $90, and they pay $498.
What every window depends on
A window only counts claims it can tie together. A new email arrives with an empty history, so it passes each of these rules the first time and every time after.
The basic checks when the customer is paying
Most of the usual checks were designed for free signups. Here the person has a working card and fully intends to pay you, which changes what each one is worth.
Email rules have limits of their own. A unique-email rule compares text, and one inbox can be written many ways. Gmail ignores dots, so jane.doe@gmail.com and janedoe@gmail.com reach the same inbox, and mail to jane+news@gmail.com arrives there too. That's how email works, and aliases are a topic of their own.
| Basic check | What it catches | What it misses | Who it wrongly blocks | Cost to keep working |
|---|---|---|---|---|
| Email verification | Typos and dead addresses | Every real address | People whose code lands in spam | Deliverability work |
| Unique emails | The same address twice | Any other address | Almost nobody | Very little |
| Disposable-domain blocks | Known throwaway services | Ordinary mailboxes, new domains | People using privacy forwarding | A list someone updates forever |
| Gmail dots and plus addressing | One inbox written several ways | A second real mailbox | People who sort their mail with plus addresses | Upkeep as mail services change |
| CAPTCHA | Simple scripts | A person at a checkout | People using screen readers | Fees and lost sales |
| Phone requirement | Casual repeats | Anyone with a second number | People without a reliable mobile | A charge per text |
| Card requirement | Little, since buyers already have one | Anyone happy to pay $19 | Nobody new | Nothing, and it adds nothing |
| One per IP address | Repeat checkouts on one connection | Any other network | Offices, campuses and households | Exceptions and tickets |
| VPN blocks | Known VPN ranges | Everything off your list | Staff on company VPNs | Another list to maintain |
| Rate limits | Bursts from one account | 37 separate accounts | Teams joining together | Tuning per route |
| Usage caps | Overuse inside one account | A discount, which isn't usage | Anyone at the cap | Deciding when the product stops |
| Shorter discounts | Some value per claim | More claims make up for it | Every honest customer | Fewer people convert |
| Manual review | What a person can spot | Anything at 11 p.m. | Whoever looks odd to the reviewer | Hours every campaign |
Run that Monday night against the table. Thirty-seven ordinary addresses, one per client, so every email row passes. A human at a checkout, so the CAPTCHA passes. Reviewing 40 new subscribers at 3 minutes each is 2 hours of reading (40 × 3 = 120 minutes), and the first invoices had already been paid by the time anyone could have started.
The work also restarts with every campaign. Lists go stale, exceptions pile up in support, and the person you refuse may be someone who really did leave for a year and wanted to come back. Every row above assumes a new record means a new person, and your records have no way to tell you when it doesn't.
One answer before the discount is applied
You don't have to build your own abuse prevention system to find out whether a claim looks like a repeat. AbuseZero exists for the exact spot where every row above ran out: a fresh account isn't enough to erase previous activity. AbuseZero flags abuse that single checks miss and detects linked claims, with multi-layer AI risk analysis examining 140,000+ signals. There's also a proprietary algorithm, with AI that checks the claim again. A unique-email rule compares text. AbuseZero detects repeated payment fingerprints across users, even when new accounts and emails hide the connection. Mailbox aliases, disposable domains and mail routing get checked too, so those lists stop being your job.
Before a discount is applied, your server makes one API call and gets a clear allow or deny, under rules you control. A new-customer price can be once per user, and a come-back price once every 365 days. Observe mode shows you what would be denied before anyone is blocked, so you can watch one full campaign first. The quickstart covers your first benefit.
Frequently asked questions
What is cancel and resubscribe discount abuse?
It's someone ending a subscription and starting a new one, usually under a new email, to collect a price meant for new customers or for people who really left.
How long should a win-back window be?
Longer than someone would happily go without your product. A reasonable place to start is 90 days away and one claim every 12 months, then adjust based on what the formula above says you can afford.
Should I honor a discount that shouldn't have gone out?
Usually yes for the current period, because a surprise price change costs more trust than the discount did. Fix the rule, then explain the next refusal clearly.
Is every returning customer abusing a discount?
No. People leave for budget or timing and come back for good reasons. Decide what you want to give them, then write that down as a window.
Protect your next campaign
Start with Builder
Pay as you go. Evaluate your first claim today.
See pricing
Builder, Pro, Scale and Enterprise.
Talk to us about Enterprise
High volumes and custom terms.