Almost every benchmark published this year compares trials to trials. The more useful finding buried in the same data is that above roughly $35,000 in contract value, the answer to "trial or demo" is neither — it's both, and the two models stop being alternatives.
There's a lot of 2026 data on this now, from several independent sources, and it disagrees with itself in ways worth understanding. Here's what holds, what doesn't, and the one comparison that actually settles the credit-card argument.
The comparison that settles the credit-card question
The debate is usually framed as conversion rate, which is the wrong denominator. Conversion rate is measured from signups, and the whole point of requiring a card is that it changes how many signups you get.
ChartMogul's 2026 report, built with ProductLed from a survey of 200 B2B software products, models it per visitor instead — and the result is unambiguous.
| Model | Signups | Paying customers |
|---|---|---|
| Standard free trial | 45 | 3.6 |
| Credit card required | 35 | 10.5 |
Fewer signups, roughly three times the customers, from the same traffic.
Why this argument never resolves internally A team measured on signups and a team measured on revenue will read the same experiment and reach opposite conclusions. Both will be correct about the number they were given.
Which is the real lesson. Track trial-to-paid separately from visitor-to-signup, report both, and treat the combination as the metric — because a model that wins on signups can lose on paying customers, and optimising either in isolation makes the other worse.
An independent read supports the direction. First Page Sage, tracking 86 SaaS clients back to early 2022, puts opt-in trials at 18.2% conversion and opt-out at 48.8% — close to three times higher, from a completely different dataset.
The median is close to useless
Before benchmarking yourself against anything, understand how wide the distribution is.
ChartMogul's median free-to-paid rate across all 200 products was 8%. But around a fifth of free-trial products convert below 2.5%, while roughly a quarter convert above 25%. A median sitting between a cluster at 2% and a cluster at 30% describes almost nobody.
Split by model, the numbers become usable:
- No credit card: 4–6% is good, 10–15% is great.
- Credit card required: 25–35% is good, 50–60% is great.
Note how far those sit from the 18.2% and 48.8% figures above. Both datasets are honest; they cover different populations, and one is an agency's own client base while the other is a broad survey. If you benchmark against a number without knowing which population produced it, you'll draw the wrong conclusion about your own performance — the same discipline that defending a budget to finance requires.
Contract value decides the model
The variable that explains most of the spread, and the one that answers the article's title.
2026 benchmarks put sub-$5,000 ACV trials at roughly 16–28% opt-in conversion, against 6–10% for enterprise deals above $75,000. That's a two-to-three-fold difference, and it isn't about product quality.
It's about who decides. A small purchase is a unilateral decision that one person can make after using the software, which is exactly what a trial produces. A large one requires committee approval, security review and budget process — and trial activation alone doesn't move any of that. The person who loved the trial still has to convince four colleagues who didn't use it, which is the buying reality described in why committees trust peers over ads.
Above roughly $35,000 ACV, a demo plus trial hybrid outperforms a standalone trial. The demo does the work a trial can't — framing the problem, addressing the committee's concerns, establishing a human who can be asked questions — and the trial does the work a demo can't, which is proving it actually works in their environment.
So the honest answer to "trial or demo" is that it's an ACV question with three bands: self-serve trial at the bottom, hybrid in the middle, and sales-led with a sandbox at the top. Choosing on philosophy rather than on deal size is how companies end up running a motion their price point can't support.
Apollo
Find and reach B2B prospects with verified contact data and outreach automation
Best for: B2B Lead Prospecting
The window closes on day seven
The timing finding that reframes most trial optimisation work.
Most B2B trial conversions happen when the trial expires. After day 14, conversion rates fall to roughly 1%.
Which means a 30-day trial is largely two weeks of activity followed by two weeks of almost nothing — and the instinct to extend a trial because people "need more time" is usually solving the wrong problem. If someone hasn't reached value in the first week, another fortnight rarely fixes it.
Payment UX now beats pricing model
The finding most likely to change what a team does this quarter, and it isn't about trials at all.
A 2026 analysis of 920 SaaS checkout flows found opt-out trial signup rates collapsing to as low as 1.1% where payment forms carried more than four fields — and recovering to around 3.8% with streamlined one-click wallet options.
That's a threefold difference driven entirely by form design, on the same pricing model. Which means a large number of teams who "tested requiring a card and it killed our signups" tested a nine-field payment form and concluded something about pricing strategy.
Before abandoning the model that produces three times the customers, count the fields. The principles are the same ones in form and lead-capture design and checkout flow optimisation — and if the signup flow itself is the constraint, that's a website and build problem rather than a growth-model one.
Worth holding alongside a related figure: only around 2.5% of users sign up for opt-out trials from organic sources. Card-required trials work considerably better on high-intent traffic than on cold discovery, which is an argument for running both models against different sources rather than picking one globally.
Industry moves it more than most teams expect
Useful context before you conclude your product is underperforming.
First Page Sage's data shows CRM tools converting trials to paid at around 29%, education at 24.8%, IoT at 25.2%, cybersecurity at 21.9%, healthcare and medtech at 21.5%, and enterprise software at 18.6%. Enterprise and fintech sit lowest at roughly 18–19%.
The pattern behind it is consistency with the ACV finding: categories with longer approval chains and heavier security review convert worse, regardless of how good the trial is. If you're in one of those categories and benchmarking against a CRM company's numbers, you're measuring your sales cycle rather than your product.
One number that's moving, and worth watching
An honest caveat rather than a finding.
At least one 2026 reading shows opt-out trial conversion falling from around 48.8% to 31.4% year on year. That may reflect changing dataset composition rather than a real decline, and it hasn't been corroborated across the other sources here.
But it's plausible for a reason worth noting: as card-required trials became more common, they stopped being a strong intent filter. When few companies asked for a card, the people who gave one were unusually committed. As more ask, the population handing over a card looks more like the general market — and conversion drifts toward the average. If that's what's happening, the advantage narrows over time rather than disappearing, and the per-visitor comparison still favours it.
What this means for the funnel around it
Two consequences that sit outside the trial itself.
The in-trial sequence matters more than the trial length. If the decision is made by day seven, your email and in-app sequence has one week to work, and a generic drip won't do it. The structure that performs is activation-led: get them to the value moment, then demonstrate the value, then ask — rather than a countdown of days remaining. That's a lifecycle design problem covered in the core automated flows, and it's what most of the 14-day generic sequences in the market are not doing.
Expired trials are a segment, not a loss. Most companies write off non-converters entirely. A structured re-engagement programme against expired trials recovers a meaningful share, and those people are far better qualified than any cold list — related to the argument in turning interest into pipeline and to what intent data can tell you about which of them came back.
What a demo is actually for
Worth stating, because "demo" gets used to mean two very different things and only one of them is worth the ACV band it sits in.
The weak version is a product tour: a salesperson walking through features while the buyer watches. That is a video, and it should be one — self-serve, on demand, no calendar involved. Making a prospect book a call to watch someone click through screens is friction charged for no return, and it's the version buyers avoid.
The strong version is a diagnostic. The seller spends most of the call understanding the buyer's situation and shows only the parts of the product that answer it. What that produces is not a demonstration but a shared understanding of the problem — and crucially, a person the committee can send questions to over the following six weeks.
That second thing is what a trial structurally cannot provide, and it's why the hybrid works above $35K. The trial proves the software functions in their environment; the human handles the four people who never logged in, the security questionnaire, and the objection that surfaces in week five. Building the material to support that is a content problem as much as a sales one, which is the case made in building a content engine around sales objections.
Which also means the demo-versus-trial framing hides a third option that suits a lot of mid-market products: a recorded diagnostic-style walkthrough available without a form, followed by a trial, with a human available on request rather than required. It removes the calendar friction while keeping the human in reach — and it fits the way committees now research, described in how buyer journeys changed in 2026 and in how AI is changing demand generation.
How to decide
- Start with your ACV, not your philosophy. Under $5K, self-serve trial. Around $35K and above, demo plus trial. In between, test both — and if you sell into named accounts, the motion should match your ABM programme rather than run beside it.
- Report per-visitor, not per-signup. Signups and customers move in opposite directions when you change the card requirement.
- Count your payment form fields before concluding anything about the card requirement.
- Measure activation rate as your leading indicator, and compress time to value in the first seven days.
- Stop extending the trial. After day 14 you're paying infrastructure costs for a 1% conversion window.
- Segment the model by traffic source, since card-required works better on high-intent visitors than on cold discovery.
- Benchmark against your ACV band and industry, not against a headline median that describes almost nobody.
And be honest about which model your offer and pricing structure can actually support. A product-led motion attached to enterprise pricing produces a lot of signups and very few customers, which looks like a marketing problem and is really a pricing one. The economics only resolve when you check them against lifetime value rather than against conversion rate.
The short version
Per 1,000 visitors, a standard free trial produces around 45 signups and 3.6 customers; requiring a card produces 35 signups and 10.5 — so the argument only resolves when you measure per visitor rather than per signup. The 8% median is close to meaningless given a fifth of products sit below 2.5% and a quarter above 25%; benchmark against your model instead. Contract value decides the motion: sub-$5K trials convert at 16–28% opt-in against 6–10% above $75K, and above roughly $35K ACV a demo-plus-trial hybrid beats either alone. The decision window closes early — after day 14 conversion falls to about 1% — so compress time to value rather than extending the trial. And before you conclude a credit-card requirement kills signups, count your payment form fields, because form design moved that number from 1.1% to 3.8% on its own.
The question was never trials versus demos. It's what your contract value can support, and whether anyone reached the value moment in week one.
Plenty of trial signups, not enough conversions?
We rebuild the in-trial sequence around activation instead of a countdown — where the week-one decision is actually made.
Fix my trial sequence →Frequently asked questions
Should you require a credit card for a free trial?
The only honest comparison is per visitor rather than per signup. ChartMogul's 2026 model across 200 B2B software products found that per 1,000 visitors, a standard free trial produced roughly 45 signups and 3.6 paying customers, while a card-required trial produced fewer signups at 35 but 10.5 paying customers. Requiring a card reduces signups and increases customers, so a team optimising signup volume will reach the opposite conclusion to one optimising revenue.
What is a good free trial conversion rate in 2026?
It depends entirely on the trial model, which is why single-number benchmarks mislead. ChartMogul's 2026 survey puts good at 4-6% and great at 10-15% for trials without a credit card, rising to 25-35% good and 50-60% great where a card is required. The median across all products was 8%, but the spread makes that figure close to useless: around a fifth of products convert below 2.5% while roughly a quarter convert above 25%.
Do free trials work for enterprise SaaS?
Less well on their own, and the gap widens with contract value. Benchmarks for 2026 put sub-$5,000 ACV trials at roughly 16-28% opt-in conversion, against 6-10% for enterprise deals above $75,000. The reason is structural rather than about product quality: a small purchase is a unilateral decision that a trial can settle, while a large one needs committee approval that trial activation alone does not move. Above roughly $35,000 ACV, a demo plus trial hybrid outperforms a standalone trial.
How long should a SaaS free trial be?
Shorter than most companies run, because the conversion window closes early. Most B2B conversions occur when the trial expires, and after day 14 conversion rates fall to around 1%. A 30-day trial is therefore largely two weeks of activity followed by two weeks of very little. The higher-return intervention is not extending the trial but compressing time to value inside the first seven days, which is where the decision is effectively made.
What most affects credit-card trial signup rates?
Payment form design, more than the decision to ask for a card at all. A 2026 analysis of 920 SaaS checkout flows found opt-out trial signup rates falling as low as 1.1% where payment forms carried more than four fields, and recovering to around 3.8% with one-click wallet payment options. That makes checkout user experience, rather than pricing model, the primary driver of opt-out signup performance — and it means many teams reject the model when they have really only tested a bad form.
