Revel Roots

what apps actually cost

How Trials Are Designed

A free trial sounds like an evaluation period. In app subscriptions it is largely not one, because around 90% of trial starts happen in the first session after install — before the product has been used enough to evaluate anything.

Six design choices produce that, and each has a clear commercial purpose. In the workplace-software market, this guide provides a contrasting example for examining how monitoring features, defaults, and disclosure are presented.

Reviewed August 9, 2026. Conversion figures from Adapty's State of in-app subscriptions 2026 — a subscription infrastructure provider reporting on its own market.

The six

One. The paywall at the end of onboarding. Not after first use — after the tour. Conversion is highest when enthusiasm is highest and information is lowest, and those are the same moment. For wider consumer-protection and interface-design context, Competition Bureau Canada provides an outside reference point.

Two. Payment details required upfront. A trial that does not convert automatically is worth a fraction of one that does. This is standard, disclosed, and it is the mechanism the rest depends on.

Three. The annual plan pre-selected. Presented as a saving against monthly, which it usually genuinely is. It also removes eleven future decision points.

Four. A trial length shorter than the evaluation. Three days for a product whose value appears in week two. Not always deliberate, and the effect is that the decision happens before the evidence.

Five. A reminder that arrives late or not at all. Platforms send one; many apps do not add their own. The reminder is the difference between a decision and a default.

Six. Feature gating during the trial. The trial includes enough to be pleasant and withholds the thing you actually came for, which appears behind the paid tier.

What is legitimate here

Worth saying plainly, because a page like this can read as an accusation.

Charging for software is legitimate. One-off purchases have largely disappeared and developers need recurring revenue to maintain products.

Requiring a card is reasonable — trials without one are abused, and the terms are disclosed.

Annual discounts are a genuine trade.

And a paywall after onboarding is where the person learns the product costs money, which is information they need.

The line is whether you could understand the outcome from the interface without effort. Most trial design sits on the legitimate side of it and still produces charges people did not intend, which is the more interesting problem.

The three defences

Cancel at subscription, not at trial end. Access runs to the end of the trial regardless in most implementations. You keep the trial, the conversion does not happen, and resubscribing is ten seconds if the product earns it. This is the whole answer and it requires no vigilance.

Use a virtual card with a limit. Merchant-locked cards from several providers; set the limit to zero after the trial period and the renewal declines cleanly.

Or diarise it yourself, two days before the end, not on the day. Weakest of the three because it depends on you at a future moment.

The question that settles it

Before starting a trial: would I pay the annual price today, for something I have not used?

That is functionally what a converting trial asks, and asking it consciously is the difference between a trial and a purchase with a delay on it.

If the answer is no, take the trial and cancel immediately. If the answer is yes, the trial was not doing much work anyway.

The short version