How to stop people farming your waitlist
Fake referrals push real early adopters down your waitlist and spoil your launch data. What waitlist referral abuse costs, and which basic fixes break.
The fix you were proudest of before launch fits in one line of a release note: a referral only counts once the new person confirms their email.
Junk addresses had crept onto your leaderboard in week 4. You shipped the confirmation step the same afternoon, the junk went away, and you told a friend you'd dealt with waitlist referral abuse for the price of a few thousand emails. What follows is a scenario, with numbers you can check and replace with your own.
Four weeks later it's launch day. Your waitlist has 9,000 entries. Every confirmed referral moves someone up 25 places, and the top 500 get a founding seat: 12 months of your Team plan free, at $49 a month. You send the 500 invites at 10:15 a.m., the plans switch on, and the welcome emails say "12 months on us".
At 7:26 the next morning, someone you recognize writes in. They joined in week 1 and brought six colleagues. "I've been telling people about you since the first week. I'm #584. Who are all these people above me?"
You export the list and sort by referrals. The top entry has 212, all confirmed, so it climbed 5,300 places (212 × 25). The person who wrote to you climbed 150 (6 × 25). A few days of reading the list by hand leaves you with this:
- 24 people × 150 entries each = 3,600 repeat entries, 40% of your list
- 24 people × 7 seats each = 168 founding seats, about a third of the 500
- $49 × 12 months = $588 for each seat
- 168 seats × $588 = $98,784 of your product given away in one send
Say a Team account costs you $9 a month to run. That's 168 × $9 × 12 = $18,144 next year, or about 27 days of runway if you spend $20,000 a month. And the reward has already gone out. The email promised those 12 months in writing, so taking a seat back means deciding, account by account, who wasn't real, in the week you're meant to be welcoming your first customers.
Now look at who didn't get a seat. Remove those 168 entries and everyone from #501 to #668 moves into the top 500, including the person who wrote to you (584 minus 168 is #416). Those are the people you built this for. They found you early and told their team, and yesterday you emailed them the launch price.
Here's their place in line, week by week. A higher number is further back, and the confirmation step shipped at the end of week 4.
Your launch data went with it. You'd been reading 9,000 on the list, half from referrals (4,500 ÷ 9,000). Take out the 3,600 and you have 5,400 real entries and 900 real referrals, or one in six. You dropped a paid test and spent a month on a second referral campaign because of that 50%.
So now you choose. Honor 168 seats you know are wrong, revoke them and handle the replies, or open 168 more for the people who got pushed out, which doubles the running cost to $36,288. And you still haven't answered the person at #584.
Those numbers are made up so you can test them. These came from a real launch:
From our early beta
After a launch went viral on X, one team gained 6,400 new users in a month. 2,700 of them were duplicate accounts. About 32 became paying customers.
That's about 42% duplicates (2,700 ÷ 6,400). About 0.5% of everyone became customers, against about 0.9% of the 3,700 that weren't duplicates, and that gap changes every plan you build on the number.
Why waitlist referral abuse pays so well
A waitlist is built to have no friction: one field, no password, no account. That's right for a product nobody can try yet, and it means nothing connects one entry to the next. Your share page prints the exchange rate, "Refer 3 friends and jump 75 places", so anyone can work out what a seat costs in entries.
Each fake entry does two jobs. It lifts whoever referred it, and it takes a place in line. The payout comes last, in one batch on launch day, so for eight weeks all you see is a chart going up.
That's why the confirmation step didn't hold. It proves someone can open an inbox, which is all email verification proves anywhere. Seven seats are worth $4,116 (7 × $588), and at that price opening inboxes is easy work.
Basic waitlist fixes and how they break
Each of these catches something, and the cheap ones are worth keeping. Email is where the limits are easiest to see.
A unique-email rule compares text. Gmail ignores dots, so first.last@gmail.com and firstlast@gmail.com reach one inbox, and plus-tagged mail reaches the same inbox as well. That's how email works, and it's how one mailbox becomes several entries.
| Basic fix | What it catches | What it misses | Who it wrongly blocks | What it costs to keep working |
|---|---|---|---|---|
| Email verification | Typos and made-up addresses | Any second real inbox | People who never dig the email out of spam | Sending and deliverability work |
| Unique emails | The same address twice | Every different address | Almost nobody | Very little |
| Disposable-domain blocks | Known throwaway services | New domains and ordinary mailboxes | People using privacy forwarding | A list someone updates for the whole waitlist |
| Gmail dots and plus addressing | One Gmail inbox written several ways | A second real mailbox | People who sort their mail with plus addresses | Upkeep as mail services change |
| CAPTCHA | Simple scripts | Anyone entering by hand | People using screen readers or privacy tools | Fees and drop-off on a one-field form |
| Phone or card | Casual second entries | Anyone with another number or card | Early adopters who only want to hear when you launch | A charge per text, or payment code before you have a product |
| One per IP address | Repeat entries on one connection | Any other network | The colleagues you asked people to refer | Exceptions for every office |
| VPN blocks | Known VPN ranges | Everything off your list | Privacy-minded early adopters | Another list to maintain |
| Rate limit per referral link | Sudden bursts on one link | Entries paced over weeks | A real referrer whose post took off | Tuning, and apologies |
| Smaller founding reward | Some of the value per seat | More entries make up the gap | Every real early adopter | A weaker reason to join or share |
| Manual review | What a person can spot | Anything that looks ordinary | Whoever looks odd to a tired founder | 500 seats × 3 minutes = 25 hours |
Two of those hurt more on a waitlist. A real referral often means someone posted the link in their team's chat and six colleagues joined from one office connection, and a one-per-IP rule throws that away. A rate limit trips on whoever's post took off, while 150 entries over eight weeks is about 19 a week and never looks like a burst.
Stack everything and your one-field form now wants a verified email, a puzzle, a phone number and a card, for a product nobody can use yet.
Change what a fake referral can buy
You can't fix this at the form, so change the deal. These are changes to your own rules, and they limit how much a fake entry can win.
Cap what referrals can buy. Count the first 10 and stop. At 25 places each, nobody climbs more than 250 places, so 212 referrals earn what 10 do.
Keep most seats out of the race. Give 350 of the 500 by join order and 150 by referrals. Climbing can then reach 150 seats at most.
Pay the seat in stages. Two months up front and the other 10 after real use. If 168 seats still go wrong, you've exposed 168 × $98 = $16,464 instead of $98,784.
Make seats expire. A seat nobody starts using within 7 days goes to the next place in line, so #501 gets a real chance.
Read the top 100 by hand before the invites go. 100 entries × 3 minutes = 5 hours.
Put the rule on the page. State one entry per person up front, and write a fair message for the entries you remove.
Each of these takes something from honest people too: your keenest advocate hits the cap, and a real founding member waits for the rest of their reward. None of it stops a repeat entry that gets through from holding a seat.
Decide each founding seat when it's claimed
By launch week you've capped, split and staged the reward, and you're still sending 500 invites on faith. What's missing is an answer at the moment a seat is about to go out: does this claim look like a repeat?
Keeping a domain list alive and writing exceptions for every office is a second job, and AbuseZero means you don't have to build your own abuse prevention system. A single check looks at one thing. AbuseZero's multi-layer AI risk analysis examines 140,000+ signals, detects linked claims and flags abuse that single checks miss. There's a proprietary algorithm as well, with AI that checks the claim again. Where your email rules ran out, AbuseZero checks mailbox aliases, disposable domains and mail routing. Where the IP limit turned away a whole office, it sees beyond an IP address with VPN, hosting, network reputation, Tor and proxy intelligence. It recognizes repeat devices across new accounts, even when email addresses keep changing, and creating a fresh account isn't enough to erase previous activity.
Make the founding seat a benefit, call the API once when a seat is claimed, and you get a clear allow or deny under rules you control. Observe mode shows you what would be denied before anyone is blocked, and the quickstart gets your first benefit running.
Frequently asked questions
What is waitlist referral abuse?
It's one person creating extra entries on your waitlist, usually through their own referral link, to climb the line or collect a reward meant for someone else.
Can I take back a founding seat after it's gone out?
Start with what your page and welcome email promised, then read how to take back free credits after an abuse wave. AbuseZero's evaluation happens before the seat goes out, which is the easier place to decide.
How do I keep fake referrals out of my launch numbers?
Report referred entries apart from direct ones, and don't plan from the referral share until you've read the top of the list.
Protect your founding seats before launch day
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.