How much free plan abuse is acceptable?
There's no universal acceptable abuse rate for SaaS. Work out yours by weighing what extra accounts cost against the real signups each fix turns away.
7,200 of your 24,000 free users aren't new people, and you don't know that yet. Your co-founder has just sent you this:
The fund's analyst wants to know how many of our 24,000 free users are different people. They need it by Thursday. Do we know?
You don't. You know the 24,000, because it's on slide 4 of the deck and in last month's board update, and it's the number that got your second support hire approved in September. Until this message, nobody had asked how many humans that is.
So you stay late. This is a scenario, so swap in your own figures as you read. By 10:30 p.m. you've read through 300 accounts by hand, about a minute each: 200 picked at random, 50 from June and 50 from October. You mark an account as an extra when it's plainly someone back for another free allowance. Of the 200, you mark 60, and 30% of 24,000 is 7,200 accounts that aren't new people. June comes in at 9 of 50, so 18% of its 1,100 signups were extras, leaving 902. October comes in at 25 of 50, leaving 1,000 of 2,000.
Your free plan costs $14,400 a month. The extras use every credit they get, at $1.50 each a month, so 7,200 × $1.50 = $10,800 of that bill is theirs.
| What the deck says | What your sample says |
|---|---|
| 24,000 free users | 16,800 people |
| Signups up 82% since June (1,100 to 2,000) | Real signups up 11% (902 to 1,000) |
| A free user costs $0.60 a month ($14,400 ÷ 24,000) | A real one costs $0.21 ($3,600 ÷ 16,800) |
Every row was a decision. The 82% is why the investor took the meeting, and why you hired that support person at $6,500 a month. The $0.60 is why you cut the free allowance by more than half in April, and your real users took that cut over a cost that was mostly never theirs.
The money is worse. You have 480 customers at $39, so $18,720 comes in each month, and $10,800 of it goes straight back out to accounts that were never going to pay. That's 58 cents of every dollar your customers give you. You have $420,000 in the bank and spend $35,000 a month, which is 12 months of runway. Without the extras you'd spend $24,200 a month, which is 17.4 months. More than five months of your company went to people who already had an account.
It's 11:40 p.m. You're going to send the real number, and you know it. What you don't have is the next sentence, the one that says how much of this is normal and how much you plan to stop. So you search for acceptable abuse rate SaaS, hoping for a figure that makes 30% sound survivable. Nothing you find can settle it, because the answer depends on two numbers only you have.
Why no single abuse rate is acceptable
An abuse rate tells you what share of your accounts are extras. It doesn't tell you what they cost. At 30%, an app where an extra costs $0.05 a month has a rounding error, and an app where it costs $8 a month has an emergency.
The obvious fix is to tighten signup until the rate looks better. But every restriction also stops some real people, and they leave no record when they close the tab. You never get an angry email from the customer you didn't get. So ask which setting makes your total cost lowest, counting the abuse you let through and the people you turned away. Whatever abuse rate sits at that point is your acceptable one.
How to work out your acceptable abuse rate
Three costs add up:
Abuse left is the extras you don't stop, times what one costs over its life. Work out your cost per free account first. Here, $1.50 a month for 6 months is $9.
Friction is the real signups you lose, times what one of them is worth:
Your 480 customers out of 16,800 people is 2.9%, so call it 3%. If a customer brings in $500 of gross profit over their life, one real signup is worth 3% × $500 = $15. Use the conversion rate with extras removed, or you'll undervalue every real person.
Upkeep is fees, support tickets and the hours someone spends on lists.
Now take October: 1,000 extras and 1,000 real signups. Every percentage and upkeep figure below is illustrative.
| Setting | Extras stopped | Real signups lost | Abuse left | Friction | Upkeep | Total | Abuse rate |
|---|---|---|---|---|---|---|---|
| Off | 0% | 0% | $9,000 | $0 | $0 | $9,000 | 50% |
| Light | 35% | 1% | $5,850 | $150 | $300 | $6,300 | 40% |
| Medium | 55% | 6% | $4,050 | $900 | $800 | $5,750 | 32% |
| Strict | 70% | 18% | $2,700 | $2,700 | $1,500 | $6,900 | 27% |
| Card | 85% | 45% | $1,350 | $6,750 | $1,800 | $9,900 | 21% |
Medium wins at $5,750, $3,250 a month better than doing nothing. It still lets in 450 extras beside 940 real signups, so in this scenario your acceptable abuse rate is about 32% (450 ÷ 1,390). Card has the lowest abuse rate in the table and the highest bill, because a rate can fall while the business gets worse.
Change one input and the answer moves:
| Lifetime cost of one extra | Cheapest setting | Total a month | Abuse rate you'd accept |
|---|---|---|---|
| $0.30 ($0.05 × 6 months) | Off | $300 | 50% |
| $9 (this scenario) | Medium | $5,750 | 32% |
| $48 ($8 × 6 months) | Card | $15,750 | 21% |
The signups and the fixes stayed the same, and the target still went from 50% to 21%. Run your own abuse side here.
- Repeat claims per month
- 1,200
- Monthly cost of repeat claims
- $2,400
- Yearly cost of repeat claims
- $28,800
- Reported trial-to-paid rate
- 5.0%
- Rate among first-time claims
- 5.9%
How to choose your target
Divide the lifetime cost of one extra by the value of one real signup. The three rows above come to 0.02, 0.6 and 3.2. The cut-offs come from the illustrative table, so rerun them on your numbers.
Two things stay outside the trade-off. Free usage should never be able to lock out paying customers, so give it its own budget and a hard spending cap. And your reported numbers get no tolerance at all. An extra that costs $0.30 may not be worth stopping, but it still inflates your growth slide, so take extras out of your metrics before anyone decides anything on them. That's your next sentence to the analyst: about 16,800 people, about 11% real growth, a target and a cap on free spend.
What each setting buys you, and how it breaks
Light: email rules
Email verification catches typos and dead addresses. It can't tell you whether someone has been here before, and real people lose the code in spam. A unique-email rule stops the same address twice and misses every other address. Personal Gmail ignores dots and delivers mail with a plus sign and a word before the @ to the same inbox, so one person can look like several. And the rule needs updating every time you learn another mail provider's quirks. Disposable-domain blocks catch known throwaway services, miss new domains and ordinary mailboxes, can turn away people using privacy forwarding, and need a list someone updates forever.
Medium: CAPTCHA, IP limits and VPN blocks
CAPTCHA stops simple scripts. Someone opening accounts by hand gets past it, and people using screen readers or privacy tools can get stuck, so you pay in fees and abandoned signups. A one-per-IP limit catches repeat signups on one connection and misses any other network. It also lands on offices, campuses and households sharing an address, and each of those becomes a ticket. VPN blocks catch only the ranges on your list and lock out privacy-minded people and staff on company VPNs.
Strict and Card: phone numbers and cards
A phone requirement slows casual second accounts. Anyone with another number gets through, people without a reliable mobile number can't sign up, and you pay for every text. A card requirement stops casual repeats too, misses anyone with another card, and turns away students and people who wanted to try first.
Limits that only shrink the damage
Rate limits stop bursts, ignore 40 accounts that each stay under the limit, and catch real users in a busy hour. Usage caps hold one account to its allowance, and the next account arrives with a fresh cap. Shorter trials and smaller allowances lower your cost per account, but every honest user pays for that, and someone after 1,000 credits now opens 5 accounts at 200 instead of 2 at 500. Manual review catches what a person can spot, and at one minute each, October's 2,000 signups are more than 33 hours of reading.
Then all of it ages. Say Medium's 55% slides to 40% as the lists go stale. Abuse left becomes 600 × $9 = $5,400, your total becomes $7,100, and the setting you picked now costs more than Light's $6,300. Nothing alerts you. You find out by spending another evening reading accounts.
When your cheapest setting still costs too much
Every row in that table makes the same trade: stop more extras, lose more real people. A better answer needs something that stops more extras without that matching loss, and you don't have to build your own abuse prevention system to get it. AbuseZero puts 140,000+ signals through multi-layer AI risk analysis, which detects linked claims and flags abuse that single checks miss, and those signals are read together, never one at a time. An IP limit or a domain list can't do that. There's also a proprietary algorithm, with AI that checks the claim again.
AbuseZero also checks mailbox aliases, disposable domains and mail routing, and 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. Creating a fresh account isn't enough to erase previous activity.
You make one API call when someone claims a benefit and get a clear allow or deny, under rules you control. Observe mode shows what would be denied before anyone is blocked, so you can fill in that table from your own traffic. The quickstart protects your first benefit.
Frequently asked questions
What is an acceptable abuse rate for a SaaS free plan?
It's the rate at the setting where your total cost is lowest. In the worked example it ranged from 21% to 50%, depending only on what one extra account costs.
Should the goal be zero free plan abuse?
On your bill, usually no, because the last few extras can cost more to stop than they cost you. In your reported numbers, yes.
How do I measure my abuse rate without a tool?
Read a random sample by hand. With 200 accounts and 30% marked, the true figure is probably around 24% to 36%. Measuring your abuse rate covers the method.
Find your number before the next board update
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.