Back to Blog

Should you require a credit card for a free trial?

A worked trade-off for credit card required free trial abuse. What a card saves, which real customers it costs you, and the sum that decides it.

Trials started this month                 1,000
Cost to serve each one                       $6
Spent on trials                          $6,000
Customers won                                40
New monthly revenue, 40 x $39            $1,560

Trials from people who'd had one before     300
How many people that is                      75
Spent on those 75 people this month      $1,800
Spent on them over a year               $21,600
Customers won from them                       0

Those are example numbers. Read the bottom half the way you'd read it at 11:55 p.m. on the last Sunday of the month, with an investor update open on one side and a search for credit card required free trial abuse on the other.

Two hours ago you typed "4.0% trial conversion" into that update and felt fine about it. Then a ticket came in from someone asking why their trial had ended early. You knew the name. You'd answered that person in September from a different email address, and again in October from another. You went through your trial records, and now you have the sum above.

Your trial lasts 7 days, so four trials a month is one person starting again every week. 75 people × 4 trials = 300 trials, and 300 × $6 = $1,800. That's more than the $1,560 of new monthly revenue from all 40 customers you won. They've been using your product every week since the summer, and they've never paid you anything.

Now look at the 4.0% in your draft. 40 customers ÷ 1,000 trials is 4.0%, but 300 of those trials were never going to buy, and 40 ÷ 700 is 5.7%. You spent last quarter rewriting onboarding to lift a number those 75 people were holding down. You've also been quoting $150 to win a customer ($6,000 ÷ 40), when real people cost you $105 ($4,200 ÷ 40).

If your company spends $24,000 a month, $21,600 a year is 0.9 of a month, about 27 days of runway. Those trials also share a queue with the people who pay you. On the 14th you hit your model provider's rate limit at 2:10 p.m., and a customer's reports failed for 40 minutes. You refunded their $39 and apologized. Three in every ten trials on that limit belonged to someone who'd already had one.

Your co-founder's answer is already in the chat: require a card. But the last three customers you spoke to said they tried you because you didn't ask for one. A card is worth it only when what it saves beats the customers it costs you, and you can work that out before Monday.

Why free trials get restarted, and why a card looks like the fix

A trial is handed out per account, and a new account costs almost nothing. For someone who wants your product and doesn't want to pay, four new accounts a month is a small chore. That's why users keep restarting your free trial.

A card feels right because cards are harder to come by than inboxes, and it will cut repeat trials. But it asks everyone for the same thing, so it lands hardest on people you wanted: someone curious on their lunch break, an employee who'd have to ask finance, a student without a credit card. The person starting their fifth trial has already shown they'll go to some effort, and prepaid and virtual cards mean one person can hold more than one.

Card up front, card mid-trial or no card: the sums

Keep the example: 700 real trials and 300 repeat ones a month, $6 a trial, $39 a month. Say a customer stays 12 months, so each one is worth $468 (12 × $39). Every percentage here is illustrative, so change them to yours.

Card up front. Half of real people don't start, which leaves 350. Those who do are keener, so 10% buy, and that's 35 customers. Repeat trials fall by 70%, to 90.

Card mid-trial. Everyone starts without a card and gets the first part, which costs you $2. The remaining $4 needs a card. Say 240 real people add one and 15% of them buy, which is 36 customers, and 60 repeat trials add one too. The bill is (1,000 × $2) + (300 × $4) = $3,200.

Illustrative, per monthNo cardCard up frontCard mid-trial
What trials cost$6,000$2,640$3,200
Customers won403536
Saved against no card$0$3,360$2,800
Value of customers lost$0$2,340$1,872
Net against no card$0+$1,020+$928
Dashboard conversion4.0%8.0%3.6%
Net a month=(old trial cost−new trial cost)−(customers lost×customer value)\text{Net a month} = (\text{old trial cost} - \text{new trial cost}) - (\text{customers lost} \times \text{customer value})

Now move one number. If 60% of real people won't enter a card instead of 50%, you start 280 real trials and win 28 customers. Trials cost (280 + 90) × $6 = $2,220, so you save $3,780, but you've lost 12 customers worth $5,616. That's $1,836 a month worse than asking for nothing.

Customers won, illustrative01020304050No cardUp frontMid-trialWorse caseCustomers
Customers won, illustrative02040No cardUp frontMid-trialWorsecaseCustomers

At a $2,640 trial bill, your break-even is $3,360 ÷ $468 = 7.2 customers. Lose 7 and you're ahead. Lose 8 and you're behind. Your dashboard won't warn you, because the conversion rate just doubled to 8.0% while you won fewer customers.

Now see where the saving came from. Of the $3,360, the 350 real people who never started account for $2,100 (350 × $6), and the 210 repeat trials you stopped account for $1,260 (210 × $6). So 62.5% of what the card saves is real people not trying your product. The 90 repeat trials that got through still cost $540 a month, or $6,480 a year.

Compare that with the line you'd really like: stop the 300 repeat trials and nothing else. That saves $1,800 a month and loses no customers, which beats both card options here.

A card also adds work: refunds and disputes from people who forgot to cancel, failed payments to chase, and a form that attracts people testing stolen cards. That last one is a job for your payment processor's fraud tools.

The other basic fixes, and how each one breaks

If the card fails your sums, these come next.

FixStopsLets throughTurns away by mistakeOngoing cost
Email verificationDead addressesAny real second inboxPeople whose code lands in spamDeliverability work
Unique emailsThe same address twiceAny different addressAlmost nobodyClose to nothing
Disposable-domain blocksServices on your listDomains not on it yetPrivacy forwarding usersA list that goes stale
Phone requirementPeople without a spare numberAnyone who has onePeople without a steady numberA fee per text
CAPTCHABasic scriptsSomeone signing up by handScreen reader usersFees, abandoned signups
One trial per IP addressRepeats from one connectionAny other networkOffices, campuses, householdsExceptions and tickets
VPN blocksRanges on your listEverything elseStaff on company VPNsAnother list
Rate limitsBursts from one sourceFour accounts a monthBusy teams at peak hoursTuning per route
Usage capsOne trial overrunningThe next trial's fresh capAnyone mid-task at the capPicking where you stop
Manual reviewWhat a person can spotAnything spread over monthsWhoever looks oddHours every month

Unique emails run into how email works. Gmail ignores dots, so sam.lee@gmail.com and samlee@gmail.com reach one inbox, and mail to samlee+april@gmail.com lands there too. Your database sees three customers, and the rules differ by provider.

A shorter trial lowers your cost per trial and works against you here. Cut 7 days to 3 and someone who wants continuous use starts about 10 trials a month instead of 4 (30 ÷ 3).

Add all of this to your form and someone who came for a quick look faces an email code, a puzzle, a phone number and a card before seeing a single screen. The 75 people starting four trials a month already know that form by heart. Your real prospects are the ones who leave.

How to decide whether to require a credit card

Measure your repeat share first. At 3% of 1,000 trials, repeats cost $180 a month (30 × $6), less than one lost customer at $468. Then run the formula with a pessimistic guess at lost signups, and test on a slice of new visitors, judging it on customers kept.

NoYesYesNoYesNoDoes a trial costreal money to serve?Keep it open,no cardCan the costly partwait until later?Ask for a cardmid-trialCustomers lost underyour break-even?Test a cardup frontNo card, limitrepeat claims
NoYesYesNoYesNoDoes a trial costreal money to serve?Keep it open,no cardCan the costly partwait until later?Ask for a cardmid-trialCustomers lost underyour break-even?Test a cardup frontNo card, limitrepeat claims

A card up front suits buyers who already know what they want. For everyone else, the $1,800 line is the one to chase, and a card can't reach it. Your database can tell you one account had one trial. It has no column for the person behind the account, which is why a one-account-per-person policy is easier to write than to apply.

Keeping the trial open without paying for repeats

AbuseZero is built for that $1,800 line. You can leave the form as short as it is today, make one API call when someone starts a trial and get a clear allow or deny under rules you control. Where a unique-email rule gave up, AbuseZero recognizes repeat devices across new accounts, even when email addresses keep changing. Your one-per-IP limit stopped at an address, and AbuseZero sees beyond it with VPN, hosting, network reputation, Tor and proxy intelligence. If you do decide on a card, AbuseZero detects repeated payment fingerprints across users, even when new accounts and emails hide the connection.

AbuseZero's multi-layer AI risk analysis examines 140,000+ signals, detects linked claims and flags abuse that single checks miss, and those signals are read together, never one at a time. There's also a proprietary algorithm, with AI that checks the claim again. Creating a fresh account isn't enough to erase previous activity.

Observe mode shows you what would be denied before anyone is blocked, and the quickstart protects your first benefit.

Frequently asked questions

Does requiring a credit card stop free trial abuse?

It reduces it and doesn't end it. Someone with another card can start again, and real people leave along the way.

Is a $1 trial safer than a free trial with a card?

A real charge adds friction, but one person can still pay $1 again and again. See $1 trial abuse.

Can I ask for a card and still limit repeat trials?

Yes. Your app still handles payment, and an AbuseZero evaluation decides whether this claim should receive the trial.

Start with your own numbers

Keep reading

Copying your prompt…

Copied the to your clipboard. Give it to your coding agent & it will handle the rest.