The Free Trial That Bills You Anyway

A free trial is not a gift. It is a payment method warming up. Every trial in the table below asks for a card, and every one of them will charge that card automatically if you do nothing — the only real variable is how long the grace period lasts, how much the first charge is, and whether cancelling early cuts you off or lets you ride out the month you were promised. The headline this year is that the grace period is shrinking, and the trial itself is quietly disappearing.

Below are eight common subscriptions, straight from their pricing pages and trial terms, checked on September 17, 2026. The numbers that matter are not how long you get free — they are what happens on the first day you stop being free.

Service	Free trial (US)	Card required?	Auto-converts to	Cancel during the trial
Spotify Premium	1 month (promos vary, e.g. “3 months for $0”)	Yes	$12.99/month	Lose Premium immediately
YouTube Premium	1 month (first-time members)	Yes	Full monthly rate	Access runs to end of trial
Amazon Prime	30 days (6 months for Student)	Yes	$14.99/month or $139/year	Keep benefits to end of trial
Apple TV+	7 days (3 months with a device purchase)	Yes	$14.99/month	Lose access immediately
Netflix	None in the US and UK; 7–30 days in select test markets	Yes (where offered)	Full monthly rate	Not confirmed
Hulu	No standalone trial since 2026-06-15; 3 days for Live TV	Yes	$89.99/month (Live TV)	Not confirmed
Canva Pro	30 days (limited-time offer)	Yes	≈$15–18/month	Not confirmed
Adobe Creative Cloud	7 days, plus a 14-day refund window	Yes	Bills at day 7	Full refund within 14 days

Checked on 2026-09-17. Trial terms and prices change.

Sources: Spotify Premium (US pricing and trial terms), Spotify Paid Subscription Terms §4, YouTube Premium trials and promotions, Amazon Prime Terms, Apple TV+ offer terms, Netflix free trial help page, Hulu free trials help, TechRadar on Hulu ending its trial, Canva pricing, Adobe Creative Cloud free trial.

The first thing the table shows is that the long trial is dead. Netflix killed its 30-day trial in 2020 and still does not offer one in the US or the UK, instead running limited seven-to-30-day trials in a handful of test markets. Hulu was the last big holdout and ended its 30-day standalone trial on June 15, 2026 — after sixteen years — leaving only a three-day trial for its Live TV bundle. Apple TV+ gives new subscribers seven days. Adobe gives seven. Seven is becoming the industry default, and the direction of travel is only ever downward.

The reason the trial keeps shrinking is that it was never really free. A trial costs the vendor money — servers, content, support — and only pays off when a fraction of trial-takers forget to cancel. As acquisition got easier, the generous trial stopped earning its keep, so services shrank it to the shortest window that still clears the bar of “I could try this before paying.” The trial you are offered now is calibrated precisely to the minimum that feels like an honest test, not to your convenience.

The second thing to notice is that a card is almost always required upfront. Spotify says it plainly: you enter a payment method to confirm your country and so it “can take payments if you want to keep Premium after the offer ends.” Canva, Amazon, Apple, Adobe and YouTube all do the same. That requirement is the entire mechanism of the trial — it is not there to verify your age, it is there so that the conversion to paid requires no further action from you at all. The subscription is already built; you are just deciding whether to intervene before the first charge.

Then comes the split that actually matters on cancellation day: when you cancel during the trial, do you keep the free time you were promised, or does the service cut you off instantly? The table shows two camps. Amazon is the generous one — cancel your Prime trial on day one and you keep shipping, video and everything else for the full 30 days, so there is no downside to cancelling the moment you sign up. Spotify and Apple TV+ are the opposite: cancel a zero-priced trial there and Premium or TV+ drops to the free tier immediately, forfeiting the days you had left. If you want to test something fully, this is the difference between a trial you can safely start and a trial you should only start when you are ready to decide.

Adobe is the odd shape in the table, and worth pausing on because it reads as more generous than it is. The seven-day trial is real, but the day it ends you are charged in full for the first month — the “14-day money-back guarantee” is a refund after the fact, not an extension of the free period. In practice that gives you up to about three weeks of reversible access, but the money leaves your account at day seven and only comes back if you remember to ask. It is a trial with a built-in refund form, and it is the same structure as the annual plan you regret later: the charge happens first, and the exit is an exercise you have to complete.

Apple TV+ hides its best trial behind a purchase. The three-month offer only exists if you buy a new iPhone, iPad, Mac or Apple TV, which makes it less a free trial than a discount buried in a thousand-dollar device. The standard seven-day trial is what most people actually get, and it is among the shortest on the table. Spotting this pattern — the generous trial that is really a hardware perk — matters because the “free” framing hides a very ordinary seven-day window underneath.

There is a third layer most people never see: the trial is rationed. YouTube Premium allows one trial per payment method and one per twelve-month period, and it can only be taken by first-time members. Amazon allows roughly one Prime trial per account per year. Spotify gates its trial to people who have never had Premium before. The word they use is “eligible,” but the effect is that you cannot reset the clock by deleting an account or swapping a card — the deck is stacked so the free period is a one-time event you cannot farm.

This rationing is what turns a forgotten trial into a permanent problem. Because the free period cannot be repeated, the trial you sign up for and forget is not a small mistake you can shrug off and try again — it is the one shot, spent, converting silently into a $12 or $15 monthly charge you never consciously agreed to keep. The trial’s real product is not the free month; it is the subscription that survives it.

The practical move follows directly from the Amazon lesson, and it is the single most reliable rule in this whole topic: if the service lets you keep the free period after cancelling, cancel on day one and decide later with no clock running. Where the service cuts you off immediately, the safer play is to set the trial only when you have a specific, time-boxed use for it — a weekend to edit a video, a week to finish a series — rather than as an open-ended “let’s see.” Either way, the decision about keeping it belongs after the trial, and it is better made with the per-use arithmetic than with the vague feeling that you might circle back to it.

There is a timing trap underneath all of this that looks like it is about the trial but is really about renewal math. Cancelling feels like the end of the obligation, but for the services that keep access running to the end of the period, cancelling the day after a conversion still means the first paid cycle has already been bought. The pause button is a trap for the same reason — it feels like stopping while the meter keeps running underneath. The moment that matters is not the moment you decide to cancel, it is where that moment falls relative to the billing date.

The other place trials leak into real money is the annual handoff. A trial frequently ends by offering a discounted first year, which is a soft-landing into a full-year commitment rather than a clean month-to-month renewal. The honest read of that offer is the difference between what annual billing actually saves you and what it costs you to leave early — and it is easy to take the discount on the enthusiasm of a trial that just ended well, without noticing you have just agreed to a year.

None of this means trials are a scam to be avoided on principle. A trial is a perfectly good way to test something you were already planning to buy, on a real project, within the window you are given. The failure mode is not taking a trial; it is taking one as a passive gesture. The trial, charitably read, is a schedule you are agreeing to — and the only reliable way to keep a schedule that ends in a charge you did not intend is to act on it the day it starts, before the clock has had a chance to make you the forgetful one.

The free trial is shrinking because it was always less a gift than a conversion funnel, and the shorter it gets, the less runway you have to notice that. Seven days does not feel like enough time, which is itself the point — it pressures you to let the trial run its course into a billing cycle rather than cut it short. The antidote is not discipline or willpower, it is one mechanical action taken once: find out whether cancelling kills the trial, and if it does not, cancel now. Everything else in the table is just the vendor’s way of hoping you will not.