Back to Blog

What is trial farming and how do you spot it?

Trial farming is one person restarting your free trial on new accounts. How it differs from credit farming, and the signs in numbers you already have.

$2,520 of this morning's $4,318.40 infrastructure invoice went to people who've never paid you, and you don't know that yet. It's the 1st and the total is $61 more than last month, so you nearly archive it and go back to the note you're writing to your investors. You run a reporting tool at $49 a month with a 14-day free trial, and the note says trial-to-paid is stuck at 5%, so next month you'll cut the price to $39.

Then you check support, and there's a message from 11:48 last night. "My trial ran out again. Can you copy my dashboards from my old account to the new one, like last time?"

Like last time. You search the inbox for the name in the signature and open the accounts behind it. There are ten in five months, on ten different email addresses, each one opened within a day of the last trial ending. None has ever paid you. By lunch you've typed what is trial farming into a search box, and the note is still unsent.

This company is made up, so here are numbers you can check. Say you find 90 people doing this. A 14-day trial fits into a month twice. A real person tries the product for a few days and costs you about $2 in compute. Someone farming uses it every day, at about $1 a day, so their trial costs you $14.

90 people x 2 trials a month = 180 trials
180 trials x $14 each        = $2,520 a month
$2,520 x 12 months           = $30,240 a year
Illustrative cost90 people, $14 a trialTrial 1$1,260Trial 2, month 1$2,520Trial 6, month 3$7,560Trial 12, month 6$15,120Trial 24, month 12$30,240
Illustrative cost90 people, $14 a trialTrial 1$1,260Trial 2, month 1$2,520Trial 6, month 3$7,560Trial 12, month 6$15,120Trial 24, month 12$30,240

So 58% of this morning's invoice went to people who've never paid you. Your 240 customers pay $49 each, and it takes the full payments of 52 of them to cover that $2,520. You burn $12,000 a month with $180,000 in the bank, which gives you 15 months. Without the farming you'd burn $9,480, and the same money lasts about 19. Those 90 people have cost you four months of runway.

Paying customers feel it too, because those trials run in the same queue. The reports your customers scheduled for 9 a.m. on the 26th arrived at 11:40. One of them wrote: "Finance needed this at 9. That's the third time this quarter. Please cancel us at the end of the month." That's $588 a year gone.

Now go back to the note. Of last month's 600 trials, 180 were those 90 people. Take them out and 30 new customers from 420 trials is 7.1%, not 5%. Your activation rate said 58%, because someone on their tenth trial never gets lost in setup. Without them it's 40%. High activation and weak conversion read like a pricing problem, which is why you were about to take $10 off the plan. Across 240 customers, that's $2,400 a month to fix a number that 90 people had bent.

Every customer you have came through that trial, so you can't switch it off. You can't leave it as it is either. And the note is still open, with a number in it you now know is wrong.

What is trial farming?

Trial farming is when one person keeps starting new free or discounted trials on new accounts, so a trial you meant to give once becomes free use with no end date. It's a form of multi-accounting, and it works on $1 introductory offers too.

Starts a free trialUses every day of itYou pay for the usageTrial ends, nothing paidOpens a new account
Starts a free trialUses every day of itYou pay for the usageTrial ends, nothing paidOpens a new account

Each trip around that loop looks like a new person trying your product, and that's why it lasts. The obvious fix is one trial per email address, and most products already have it. To that rule, a new address is a new customer.

The motive is ordinary: $588 a year saved, for a minute or two of signing up. Plenty of these people like your product, as why users keep restarting your free trial explains. And not every repeat is farming. Someone who comes back 18 months later because the product changed is a good sign, so decide what you'll do when a customer asks for a second trial before you need the answer.

How trial farming differs from credit farming

Credit farming is the same habit aimed at free credits instead of trial time.

Trial farmingCredit farming
What's takenWeeks of your product, trial after trialFree credits, often spent at once
Where you see itInside a normal-looking invoiceAt a provider limit, or on a graph
What it bendsTrial-to-paid, activation and your forecastMostly the bill and the signup count

Credit farming tends to hit a limit and wake someone up. Trial farming fits inside the budget you already approved, so nobody goes looking.

How to spot trial farming in numbers you already have

Pull the last six months from your billing, analytics and support tools, and compare each month with your own history.

  • Trial starts are climbing and new customers aren't, with no launch or campaign to explain it.
  • Activation and conversion are moving in opposite directions. Someone who's done this before finishes setup every time and still doesn't pay.
  • Your heaviest users aren't customers. Sort last month's compute cost by account and count how many of the top 20 are on a trial.
  • Support keeps hearing about earlier trials. Search your inbox for "old account" and "trial again". It's the cheapest thing on this list.
  • Trial cost per new customer keeps rising. Work it out every month:
Trial cost per new customer=total cost of all trialsnew paying customers\text{Trial cost per new customer} = \frac{\text{total cost of all trials}}{\text{new paying customers}}

Here's the illustrative month from above, counted both ways:

Last monthAs reportedReal people only
Trials started600420
Trial-to-paid5.0% (30 ÷ 600)7.1% (30 ÷ 420)
Activation58% (348 ÷ 600)40% (168 ÷ 420)
Cost of trials$3,360$840
Trial cost per new customer$112$28

Farming is 30% of those trials and 75% of what they cost. None of these signs tells you who's doing it, since a press mention lifts trial starts and a serious buyer uses every day of a trial. They tell you whether to dig further. When you do, find repeat free trial users in your database and measure your free trial abuse rate.

Why the usual fixes for trial farming keep breaking

Someone on their tenth trial has a new address and a reason to want your product. Most of the basics were built to stop something else.

Checks on the email address

Email verification catches typos and made-up addresses. It proves someone can open an inbox today, and a second inbox is free. Keep it, but expect to lose some real people when the code lands in spam. A unique-email rule stops the same address signing up twice, wrongly blocks almost nobody and says nothing about a different address.

Disposable-domain blocks catch the well-known throwaway services. They miss new domains and every ordinary mailbox, they turn away people using privacy forwarding, and someone has to keep the list current forever.

Gmail ignores dots, and mail with a plus and a word before the @ reaches the same inbox, so several strings can be one mailbox. Counting those as one catches the most casual repeat, and a second real mailbox isn't affected at all. A rule like that needs upkeep of its own.

Making a new account harder

A CAPTCHA filters out simple scripts, but this is mostly a person with a browser, and a person can solve it. It's hardest on people using screen readers.

Phone numbers and cards put off casual repeaters. Anyone with a second number or card carries on. People without a reliable mobile number drop out, along with students and anyone who wants to try before paying, and every text is a charge.

One trial per IP address catches repeats from one connection until that person tries from their phone. It also blocks the second person in an office or household, who for a business product is the team you wanted to win. VPN blocks catch the services on your list, turn away staff on company VPNs and give you another list to update.

Limiting what a trial can do

Rate limits stop bursts. Someone using your product normally for 14 days doesn't burst, and the real user in a busy hour hits the limit first.

Usage caps are worth keeping. A cap on trial accounts, in a queue apart from paying customers, means those 9 a.m. reports arrive at 9. It doesn't reduce the number of trials, and a real prospect can hit it just as the product starts to make sense.

Shorter trials don't survive the arithmetic. Cut 14 days to 7 and each farmed trial costs $7, but four now fit in a month: 90 people × 4 trials × $7 = $2,520. It's the same bill, and every honest person has half the time to see the value.

Manual review

A person can notice things a rule won't, as that inbox search did. At an illustrative two minutes per trial, 600 trials is 20 hours every month, and when you close a real person's trial by mistake, they rarely write back.

If you only do two things this month, give trial accounts their own capped budget and queue, and add that table to your monthly numbers. Neither one stops the next farmed trial from starting.

Evaluate each trial claim before it starts

A new address satisfies a unique-email rule and a new connection satisfies an IP limit. AbuseZero starts where those rules run out, with the question they can't ask: does this trial claim look like a repeat? Your server makes one API call and gets back a clear allow or deny, under rules you control. 140,000+ signals go through multi-layer AI risk analysis, which detects linked claims and flags abuse that single checks miss. They're read together, never one at a time. There's also a proprietary algorithm, with AI that checks the claim again. The email rules you'd have to keep updating stop being your job, because AbuseZero checks mailbox aliases, disposable domains and mail routing. 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. You don't have to build your own abuse prevention system.

Eligibility can be once per user, or once per rolling number of days if you're happy to see people come back. Observe mode shows you what would be denied before anyone is blocked, and the quickstart protects your first benefit.

Frequently asked questions

Is trial farming the same as free trial abuse?

It's one kind of it. Free trial abuse also covers things like one company starting trial after trial.

How much trial farming is normal?

Nobody has a reliable figure for your product. Measure yours, put a cost on it and decide how much you can accept.

How do I stop trial farming?

Cap what trial accounts can spend, then evaluate each trial claim before it starts. AbuseZero returns allow or deny. For the email side, see how to stop users signing up for a free trial again with a new email.

Give your next trial to someone new

Keep reading

Copying your prompt…

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