Free trial vs freemium: which to choose

You have a SaaS, a few signups and one or two paying customers, and you cannot decide between a free trial and freemium (a free plan people can keep forever). Pick wrong and you either pay to host free users who never pay, or you blame the model when the real problem is that too few people tried it. Here is how to settle free trial vs freemium today with 4 numbers, set it up in an hour, and know when to change it.

Do these three things

About 2 hours today, then about 2½ hours a day and 20 minutes each Sunday. At 100 messages a day you reach 30 ended trials in about 6 weeks.

Show the three actions in fullHide the full list

1. 1Today: answer 4 questions about what a free user costs you, how fast they get a result and whether they show your product to others. Pick a 14-day free trial with no card unless all 3 tests point to a free plan. Put it on your pricing page and signup button.
2. 2Every day: send 100 personal messages that invite people who fit to start the trial. Write to every new trial user within 24 hours, and ask every user to pay 2 days before the trial ends.
3. 3Every Sunday: count trials started, first results and trials paid. After 30 ended trials, keep, lengthen, add a card or add a free plan, using the rule in step 4.

Get help doing these

ChatGPT or Claude reads this guide and works through it with you.

Do this with ChatGPT(opens in a new tab)Do this with Claude(opens in a new tab)

1. Pick a free trial or a free plan with four numbers

Start with a 14-day free trial and no credit card. Choose a free plan instead only if three things are all true: a free user costs you under $1 a month, a new user gets the first result alone in under 10 minutes, and each use shows your product to someone new.

Why a trial is the default. A trial ends. In 14 days each person says yes or no with money, so you learn fast. A free user can stay free for months, so you wait much longer to find out if anyone will pay. And a free plan costs you hosting, email and support for every person on it.

Why the choice matters less than you think. The newest benchmark is ChartMogul's February 2026 report on 200 software companies. For every 1,000 visitors, freemium products got 90 signups and 5 paying customers. Free trials with no card got 45 signups and 3.6 paying customers. Trials that asked for a card got 35 signups and 10.5 paying customers. Do the sums for your site. At 300 visits a month, freemium gives about 1.5 paying customers a month and a no-card trial about 1. That is less than one customer apart. What changes your month is how many people try the product and how many of them get a result, which is steps 3 and 4. If nobody pays whatever you offer, the model is not the problem: start with why no one is buying.

Those companies are much bigger than you, with $1–10M a year in revenue. So use the numbers to compare the models, not to predict your own month.

Now fill in your four numbers.

What one free user costs you a month (hosting + email + AI use in a heavy month + tools): $[ ]
Minutes until a new user gets the first result, on their own: [ ]
Does each use show your product to someone new (a shared link, an invite, a "made with" mark)? [yes / no]
Signups a month now: [ ]
Free plan only if: 1 is under $1, 2 is under 10 minutes, and 3 is yes.
Otherwise: a 14-day free trial, no card.
Make it 30 days only if the first result needs something that happens once a week or less, like a weekly report or a monthly invoice run.Our choice: [free trial, 14 days, no card / free plan with one limit] because [the number that decided it].

For number 1, use the sums from what one customer costs you. Paste last month's bills into ChatGPT or Claude and ask it to divide by the number of people who used the product; check the total yourself. If your product calls an AI model, count a heavy user's month. Every free action runs a paid model call, and that is why founders now say free plans are risky for AI products. For number 2, sign up as a stranger with a new email and time yourself to the first real result. Number 4 tells you how long any change will take to judge. If nobody outside your friends has used the product yet, run a short beta with set-up calls first, then open the trial.

• Tool: Google Sheets, your bills, and ChatGPT or Claude to add them up.
• Time: 30 minutes.
• You will have: one line: what you chose, and the number that decided it.
• It worked if: all four lines hold real numbers, not guesses, and the choice follows the rule.
• Common mistake: choosing freemium because a famous product grew with it. Those products had millions of visitors, so a small share paying was still a lot of customers. With a few hundred visitors, the same share is close to zero.

2. Set up a 14-day free trial with no credit card

Give every new account 14 days of the full product without asking for a card, and say the same thing about it everywhere: the pricing page, the homepage button and the signup page.

Pick where the trial lives. There are two simple ways.

• In your own app (simplest if you have it already). Give each new account a trial end date 14 days after signup, and show "Trial ends on [date]" inside the product. At the end, send a payment link. Create one in Stripe: Payment Links, click "+New", pick your product, click "Create link".
• In Stripe. Create a payment link for your subscription product and select "Include a free trial", set 14 days, then select "Let customers start trial without payment method". Then open Settings, Subscriptions and emails, and under "Manage free trial messaging" turn on the reminder and choose "Link to a Stripe-hosted page", so the email has a link to add a card. If you skip the reminder, Stripe pauses the trial when it ends. If you use Paddle, its trials without a card are still in early access and need code, so the first way is easier.

Write one trial line and use it everywhere. Put it on your pricing page, under your homepage headline and the line with your value proposition, and on your landing pages. With a few hundred visits a month, an A/B test of two trial lines will not give you an answer, so use one line and judge by trials. If visitors come but do not sign up, that is a website problem, not a trial problem.

Button: Start your 14-day free trialUnder the button: No card needed. Full product for 14 days. Then $[price] a [team / user] a month, or cancel with one click.

If you ask for a card at signup later, the rules change. Show the price, the date of the first charge and how to cancel right beside the button. Turn on Stripe's trial reminder, which it sends 7 days before the trial ends, because card networks require a reminder. Since July 2025, California's law covers free trials too: people who sign up online must be able to cancel online, and a trial longer than 31 days needs a reminder 3 to 21 days before it ends. The pricing page guide has the rest of the billing terms.

If step 1 chose a free plan, set it up with three parts:

• One limit, where the value is felt. Pick one: projects, team members, or uses a month. Set it so a user gets real value and then hits the limit within their first 2 weeks.
• A cap on AI use. Give free users a set number of AI actions a month. In ChartMogul's 2026 report, a company that gave free users a limited number of AI credits each month saw more people use the product each month.
• An upgrade message at the limit. Say what they hit and what the paid plan adds. The onboarding emails guide has the email for this moment.

Free: [one thing, e.g. 1 project] and [number] AI [actions] a month, for as long as you like.
Paid, $[price] a month: [what grows, e.g. 10 projects] and [number] AI [actions] a month.

• Tool: your app or Stripe Payment Links, and your site.
• Time: 1 hour.
• You will have: a working trial (or free plan), and one line about it on three pages.
• It worked if: you sign up with a new email, get into the product without a card, and see the end date. The three pages say the same thing, and the signup page opens fast on your phone; if it does not, fix the speed.
• Common mistake: "Free forever" on one page and "14-day trial" on another. People do not know what they are signing up for, and some leave. The second mistake is an "unlimited" free plan when every use costs you money. The third is a free plan, a trial and a lifetime deal all at once.

3. Bring in trials every day and talk to every trial user

Your trial only teaches you something if people start it. Every day, send 100 personal messages that invite people who fit to try it, write to each new trial user within 24 hours, and ask each one to pay 2 days before the trial ends.

Why messages, not waiting. At 300 visits a month, the benchmark above gives about 13 trials a month from your site. At that pace, 30 ended trials take more than 2 months. One-to-one messages bring about 1 new user per 100 messages, as the first 100 users guide found. So 100 a day adds about 30 trials a month, and you can judge your trial in about 6 weeks.

The 100 messages. Use your list of people who fit your buyer sentence. Send them by email from a second, warmed-up domain and on LinkedIn. LinkedIn limits connection requests each week and a new domain must start slowly, so spread the 100 across email, LinkedIn messages and communities such as Reddit, where each group sets its own rules on self-promotion. AI drafts each message from your list; you add one true line about the person. That takes about a minute each. The full method is in how to write a cold email, and follow up up to 3 times. If fewer than 1 in 100 messages brings a trial, rewrite the first line before you send the next 100.

Within 24 hours of each new trial. Write to the person yourself and offer a 15-minute set-up call, where you get them to the first result before the call ends. It is a short demo on their own work, not a tour. ChartMogul's 2026 report lists one-to-one onboarding calls among the changes companies made to get more users to pay. If they do not take the call, send one step to the first result. The onboarding emails guide has the full set.

Subject: your [product] trialHi [first name],Thanks for trying [product]. [One true line: what you saw them do, or why they signed up.]The fastest way to see if it helps is to [the first result]. Want me to set it up with you? It takes 15 minutes: [booking link].If you'd rather do it alone, start here: [one link to the first step].[Your name]

2 days before each trial ends. Ask them to pay, with the price and a payment link, and ask for one line if the answer is no. Write down every answer: paid, no (and why), or no reply. Getting more of these to say yes is a whole guide of its own.

Subject: your [product] trial ends on [day]Hi [first name],[One true line about what they did in the product.]Your trial ends on [day]. To keep going, it's $[price] a [team / user] a month: [payment link].If it's not worth that to you, could you reply with one line on why?[Your name]

• Tool: your lead sheet, Gmail or LinkedIn, and ChatGPT or Claude to draft. A CRM can wait.
• Time: about 2 hours for the 100 messages; about 2 minutes for each trial message; 15 minutes for each set-up call.
• You will have: a row for every trial: start date, first result (yes or no), end date, and paid, no (with the reason) or no reply.
• It worked if: every new trial got the day-1 message within 24 hours, and at least half of the people who took a set-up call reached the first result.
• Common mistake: waiting for your site to bring trials, then judging the model on 6 of them. Six trials tell you nothing.

4. Count every Sunday and change the trial after 30 ended trials

Every Sunday, count trials started, trials that reached the first result, trials that ended and trials that paid. After every 30 ended trials, use the rule below to keep your trial, make it longer, ask for a card, or add a free plan.

Take the counts from your app or from Stripe's subscriptions list, and the reasons from the replies to your trial-end messages. If you also want to know which page or message each trial came from, set up conversion tracking for the trial start. When someone says no without a reason, ask them for one line; the reasons decide rules 2 and 5.

Week | Trials started | Reached the first result | Trials ended | Paid | Top reason for "no", in their words

After 30 ended trials, check these in order and act on the first that is true:

1. Fewer than half reached the first result (under 15 of 30): keep your trial as it is. People are not getting to the point where the product helps, and no model fixes that. Work on your activation rate first.
2. 3 or more "no" answers say they ran out of time: make the trial 30 days for the next 30 trials.
3. Half or more reached the first result, but fewer than 1 in 10 paid: keep the model. Look at the trial-end ask and your price, and check the people fit your buyer.
4. 1 in 10 or more paid: your trial works. Ask for a card at signup on the next 30 trials, with the billing terms from step 2. Keep the card if those 30 bring at least as many paying customers a week as the 30 before. Fewer signups but more payers also changes what each customer costs you to win. In the benchmark, card trials brought about 3 times as many paying customers per visitor, but your own numbers decide.
5. 5 or more "no" answers say it is too small to pay for every month, and a free user costs you under $1: keep the trial, but when it ends without payment, move the person to a small free plan with one limit (step 2) instead of locking them out. This is called a reverse trial.

Change one thing at a time, and judge the change after the next 30 ended trials. Once a month, put these numbers next to your churn in your monthly marketing review. More paying customers only add MRR if they stay, which is reducing churn. Once they stay, raising your price is the next lever, not a bigger free offer.

• Tool: Google Sheets.
• Time: 20 minutes each Sunday; 1 hour when you change something.
• You will have: one row a week, and a written decision with its reason after every 30 ended trials.
• It worked if: each change is judged on the next 30 ended trials, the same way, and the share of trials that pay does not fall after it.
• Common mistake: switching to freemium, or adding a card, after 5 trials, or to fix people who never used the product. That changes who signs up, not whether the product helps them.

Your free trial or free plan checklist for the next 6 weeks

0 of 10 done

Work through this list with an AI assistant

It drafts the work and keeps track of what is done.

Do this with ChatGPT(opens in a new tab)Do this with Claude(opens in a new tab)

Step 1 · Pick a free trial or a free plan

• Fill in the four numbers on your trial or free plan sheet.
• Pick a 14-day free trial with no card, or a free plan if all three tests pass.

Step 2 · Set up a 14-day free trial with no credit card

• Set up the trial in your app or with a Stripe payment link.
• Put one trial line on your pricing page, homepage button and signup page.
• If you chose a free plan, set one limit, an AI cap and the upgrade message.

Step 3 · Bring in trials and talk to every trial user

• Every daySend 100 personal messages that invite people who fit to start the trial.
• Every daySend the day-1 message to every new trial user.
• Every daySend the trial-end message to every trial that ends in 2 days.

Step 4 · Count every Sunday and change the trial

• Every SundayFill in this week's trial row.
• After 30 ended trialsCheck the five rules in order and make one change.

Copy listDownloadReset