# How to Design a SaaS Free Trial

> Trial length, credit card or not, feature gating versus usage gating, and trial extensions, decided with conversion data rather than what a competitor does.

Source: https://saas-marketing.net/guides/free-trial-design/
Topic: SaaS Growth Marketing
Type: guide
Published: 2026-09-11
Last updated: 2026-09-11
Publisher: SaaS Marketing (saas-marketing.net)
License: CC BY 4.0. Quote or republish with attribution and a link to https://saas-marketing.net/guides/free-trial-design/

## Short answer

A SaaS free trial is four decisions: length, whether a credit card is required, what gets restricted, and what happens at expiry. Set length from your median time to first value rather than copying 14 days. Requiring a card raises trial to paid conversion to roughly 35 to 55 percent but cuts trial starts sharply, which is the right trade above roughly 10,000 dollars ACV. Below that, opt in trials at 8 to 22 percent conversion usually produce more revenue.

## Key takeaways

- Fourteen days is a convention nobody tested. Set trial length to roughly twice your median time to first value.
- Opt out trials convert at 35 to 55 percent of starts, opt in trials at 8 to 22 percent, but the bases are different.
- Credit card gating is the right call above roughly 10,000 dollars ACV and usually wrong below 2,000 dollars.
- Usage gating suits products with a natural volume metric. Feature gating suits products where the buyer is a different person to the user.
- Granting one extension on request is cheap. Auto-extending everyone destroys the deadline that drives conversion.
- Opt out trials carry a refund and support cost that never appears in the conversion rate you present to the board.

---

Almost every SaaS trial in the market is 14 days, no credit card, with a handful of features locked. Ask why and you get some version of "that's what everyone does". The four decisions inside a trial each move conversion by more than any landing page test you'll run this year, and three of them are usually made by copying a competitor whose economics are nothing like yours.

## Why 14 days is a default nobody tested

Fourteen days comes from nowhere defensible. It's an artifact of early SaaS copying early SaaS, and it persists because it's the length in every trial template.

The number you should use comes from your own data: median time to first value, doubled. First value means the activation event you can actually defend, not signup. For a project tool it might be a second team member accepting an invite. For an analytics product it's a data source connected and one report saved. For a support product it's the first ticket resolved in-app.

**Setting trial length from your own data**

Product complexity dominates. Datadog-style infrastructure tools need time for someone to get approval to install an agent. Figma-style tools deliver value in twenty minutes. Those should not have the same trial length, and the fact that both often ship 14 days is evidence nobody ran the calculation.

<Calculator id="trial-conversion" />

If you want to skip the spreadsheet, the [Free trial conversion calculator](/calculators/trial-conversion/) takes activation timing and conversion inputs and tells you where the cutoff costs you revenue.

## Opt in or opt out: the full conversion spread

This is the decision with the widest published range and the most misleading numbers in circulation.

Opt in trials, meaning no card at signup, convert somewhere between 8 and 22 percent of trial starts, with a median around 14 percent. Opt out trials, card collected up front with automatic billing at expiry, convert between 35 and 55 percent, median around 44 percent.

The published figures disagree because they measure different things. ChartMogul's analysis across roughly 200 products reported an average near 8.9 percent. First Page Sage has published 18.2 percent for opt in and 48.8 percent for opt out. These are not contradictory findings; they're different denominators and different product mixes. ChartMogul's figure spans a broad set of mostly self serve products where "trial start" includes anyone who created an account. First Page Sage's opt out figure counts people who entered payment details, which is a far more qualified population.

Stop comparing trial to paid conversion rates across models. Compare paid customers per 1,000 visitors to the pricing page. Opt in at 14 percent of 400 trial starts gives you 56 customers. Opt out at 44 percent of 120 trial starts gives you 53. Roughly a tie, and the opt in path also gave you 340 more people in your database to nurture.

My position: credit card required trials buy you a better-looking conversion rate and a worse top of funnel. That's the right trade above roughly 10,000 dollars ACV, where every trial gets sales attention and filtering for intent saves real money. Below 2,000 dollars ACV in a competitive category, it's usually wrong, because a prospect comparing four tools will start trials on the three that don't ask for a card.

The middle band, 2,000 to 10,000 dollars, is genuinely ambiguous and worth testing rather than assuming.

Automatic conversion generates refund requests. Plan for 5 to 15 percent of opt out conversions to ask for their money back within 60 days, plus support tickets, plus a steady trickle of G2 reviews complaining about surprise charges. If you run opt out, send a real reminder three days before billing and make cancellation one click. Companies that hide the cancel flow win one quarter and lose the category reputation.

## What to gate: features, usage or time

Three restriction models, and the right one depends on who uses the product versus who buys it.

Usage gating gives full functionality with a ceiling: rows, seats, runs, events, messages. Zapier's task-based tiers are the canonical example. It converts well in self serve because the user experiences the complete product and hits a wall caused by their own adoption, which is a much better conversion trigger than a locked button.

Feature gating hides capabilities. It works when the buyer and the user are different people, and the gated features are the ones the buyer cares about: SSO, audit logs, permissions, admin controls, SLAs. The end user gets the full working product; the security review triggers the upgrade. This is why almost every SaaS company gates SAML rather than gating core workflow.

Time gating restricts nothing but the clock. Cleanest experience, sharpest deadline, and it requires the most confidence that your product proves itself quickly.

The gating mistake that costs most: locking the feature that demonstrates the value. If your differentiator is the automated report and the trial can't generate one, the trial proves nothing. Gate the scale, not the proof. This connects directly to how the paid tiers are built, which we cover in the growth-side view of [Free trial vs freemium](/comparisons/free-trial-vs-freemium/).

The reverse trial deserves a mention here: full premium access for a window, then downgrade to a permanent free plan rather than a wall. It keeps the user in your database indefinitely and gives you a second conversion opportunity later. The mechanics and when it beats a standard trial are in [Reverse Trial](/glossary/reverse-trial/), and the strategic choice between trial and freemium models is covered in [Free trial vs freemium](/comparisons/free-trial-vs-freemium/).

## Extensions and reactivation: what to grant and what to refuse

A user who asks for an extension has intent and a blocker. Refusing costs you a customer to protect a policy.

Grant one extension, seven days, on request, with a reason captured. Those reasons are the best qualitative data in your funnel: "our IT hasn't approved the integration yet" and "I was on holiday" are different problems and the first one is a product roadmap signal.

Put the extension behind an in-app button on the expiry warning screen, limited to one use, requiring a single-select reason. You will get a higher grant rate than an email-to-support flow, you will get structured data, and you will not need a human in the loop. Cap it at one so the deadline still means something.

Automatic extension for everyone is the version to avoid. The expiry date is doing a large share of your conversion work, and the days immediately before it produce a disproportionate share of upgrades. Remove the date and you flatten that spike without gaining anything.

Reactivation is separate and underused. Users who let a trial lapse without converting are a better audience than cold prospects, and a three touch sequence at 30, 60 and 90 days that leads with a shipped feature rather than a discount recovers a meaningful slice. Lead with "we built the thing you asked about" where you can tie it to their actual blocker.

## The trial expiry sequence that does the work

Most of the conversion lift available in a trial program lives in the emails, not the product. Five touches, branched on behaviour.

**Trial email sequence**

Branching on behaviour matters more than cadence. A user who has connected a data source and a user who logged in once need different messages, and sending both the same day-three email wastes your best window. The measured differences between behavioural and time-based sequences are collected in [Trial Conversion and Email: The Data](/research/trial-email-conversion-study/).

**Day 0 to 3** Window where most trial abandonment happens, making the first two emails the most influential assets in the funnel

## Testing trial changes without fooling yourself

Trial design tests are slow and easy to misread. A 14 versus 21 day test cannot be read until the longer cohort has fully expired plus a lag for late conversions, which means six to eight weeks minimum.

Three rules. Change one variable at a time, because a test that alters length and gating tells you nothing about either. Calculate sample size before starting, since detecting a change from 14 percent to 17 percent needs far more traffic than teams assume; our [A/B Test Sample Size Calculator](/calculators/ab-test-sample-size/) will tell you whether the test is even feasible at your volume. And measure revenue per visitor, not trial conversion rate, or you'll optimise yourself into a smaller business with a prettier metric.

Segment the read too. A trial change that helps small teams can hurt enterprise, and the blended number shows nothing. Compare against the segmented figures in [Free Trial and Freemium Conversion Benchmarks](/research/trial-conversion-benchmarks/) rather than against a single industry average, which is the mistake that produces most of the bad trial decisions in this category.

The honest tradeoff to close on: every trial design choice trades volume against quality, and there is no configuration that wins both. Opt out gives you a conversion rate the board likes and a smaller funnel. Long trials give more users time to activate and let more of them forget you. Pick the trade that matches your ACV and your sales capacity, write down why, and revisit it when either changes.

## What to do next

Measure median time to first value this week, because every other decision on this page depends on it. Then check your ACV against the 10,000 dollar line to settle the credit card question, pick usage gating if you have a natural volume metric and feature gating if your buyer is an admin, and rebuild the expiry sequence with behavioural branches before touching anything else. The sequencing work is covered in more depth in [Improving Trial to Paid Conversion](/playbooks/trial-to-paid-conversion/), and the wider funnel context sits in [SaaS Growth Marketing](/saas-growth/).

## Frequently asked questions

### How long should a SaaS free trial be?

Set it to roughly double your median time to first value, measured from signup to the activation event you can defend. For most B2B tools that lands between 7 and 21 days. A 30 day trial on a product that delivers value in an afternoon just gives users 29 days to forget you exist. Longer trials only help products requiring data connection, admin approval or team rollout.

### Should a SaaS free trial require a credit card?

Require it when your ACV is above roughly 10,000 dollars and sales touches every deal anyway, because the card filters for intent and your team's time is expensive. Do not require it below roughly 2,000 dollars ACV or in a crowded category, where the drop in trial starts outweighs the higher conversion rate. Measure paid customers per thousand visitors, not conversion rate.

### What is the difference between opt in and opt out free trials?

An opt in trial does not collect payment details and ends unless the user actively buys. An opt out trial collects a card at signup and converts to paid automatically unless the user cancels. Opt out produces much higher trial to paid rates because the population entering the trial is far more qualified, not because the trial experience is better.

### What is the difference between feature gating and usage gating?

Feature gating hides specific capabilities behind the paywall, so trial users see a locked button. Usage gating gives full functionality with a cap on volume, seats, records or runs. Usage gating generally converts better in self serve products because the user experiences the whole product and hits a ceiling caused by their own success.

### Should you extend a free trial when someone asks?

Yes, once, for a stated reason, ideally handled by a human or a one click in-app option. A user asking for more time is telling you they have intent and a blocker. What you should not do is extend automatically for everyone, because the expiry deadline is doing most of the conversion work and removing it flattens your conversion curve.

### What is a good free trial conversion rate for SaaS?

It depends entirely on whether a card is required. ChartMogul's analysis across roughly 200 products put the overall average near 8.9 percent for opt in style trials. First Page Sage has published figures around 18.2 percent for opt in and 48.8 percent for opt out. The spread reflects different definitions of the denominator, not different product quality.

### What should the trial expiry email sequence look like?

Five touches: a welcome that points at one activation action, a day three nudge based on whether that action happened, a mid trial message showing what they have produced, a three day warning naming the specific data or work they will lose access to, and an expiry day message with a one click upgrade. Behavioural branching beats cadence every time.