Soft Paywall

Products that charged money used to charge at the entrance. Not being able to try something before buying it was normal, so the description and the reputation were the whole basis for a decision. That holds up while there are only a few alternatives to weigh.

Online products changed that. When ten similar things show up in the search results, asking for payment at the door means people leave for the next result without finishing the description. Pagedeck started out with the payment screen sitting directly after signup. Two hundred people signed up in a month and 60 of them paid. At 30 percent that looked like a fine number.

Opening a free portion turned the figures over. Three thousand people signed up the next month and 150 of them paid. The payment rate fell to 5 percent while the money coming in was two and a half times larger.

So instead of arguing the product into someone, the product gets some time to prove itself, and the question of payment comes after that.

A free portion stays open and the payment screen appears only at the moment someone crosses its edge. Three things get designed: what the edge counts, where the line sits, and what a blocked person keeps.

There are two ways to count the edge.

1. Feature gating divides access feature by feature according to plan. Writing and sharing documents work on the free plan, while permission settings and revision history open on a paid one. Good candidates are features a solo user never needs and a company working as a team cannot do without.
2. Usage limits cap how much can be used. Three documents, five teammates — written as numbers, and the unit chosen is one that grows as the customer grows. Every feature stays open and only volume is restricted, so people experience the whole product and hit the wall naturally as they grow into it.

Where the line sits decides whether any of this works. There is one rule: put it after the point where a user first feels the product is useful, the moment known as activation. At Pagedeck that moment was not writing a document alone; it was a colleague leaving the first comment on it. Block earlier and the user leaves with the memory of having briefly used a text editor. Block later and they weigh payment against losing a place their team is already working in.

That point gets found in usage records rather than in a meeting. Limit numbers come from the same place: line up how many documents free users created, find where half of them stop, find where accounts start being used by teams, and put the line between the two.

What a blocked person keeps is also a design decision. Take away what they already made and what comes back is resentment, not payment. Stopping the fourth document is fine; the first three have to stay readable and editable. The blocking screen says why the block happened and what a given price opens.

The first place is plan design. It gets settled together with pricing, the work of deciding what the product costs. What stays free determines what is left to sell.

In the product itself it surfaces in three places.

1. Locked features. Leave the button visible rather than hidden and show an explanation when it is pressed. Hidden, the feature is something nobody knows exists, so no reason to pay for it forms either. The visible lock is part of the pitch.
2. Limit warnings. Say something as the cap approaches. Showing the remaining count at two documents out of three turns the moment of blocking from an accident into something expected.
3. A path to checkout. From the blocked spot, payment should be one step away. Make people pass through two or more screens here and the number who cross drops noticeably.

Two numbers get watched together: the share of free users who convert, and the volume of people who use the free portion and leave. Watch only the first and narrowing the free portion always looks correct. The second one shows whether signups themselves shrank after the cut.

Some products are a poor fit. Content that is finished after one reading, or software enterprises adopt by signing a contract, would be giving away the entire product as its free portion. Those tend to block at the door instead, or open everything for a fixed window with a free trial that closes when the window ends.

"Let's go soft here. Asking people to pay before they've tried it means they won't even sign up."

A proposal to add a free portion so the product can show its value first.

"Three documents is too tight. People hit the wall before activation."

The line has been drawn earlier than the point where users feel the product is useful.

"Existing documents still have to open on the blocked screen."

A principle that crossing the line does not claw back what someone already made.

At Pagedeck, Mia and Grace set this line together. Mia pulls the differences between accounts that paid and accounts that left out of the usage records, and Grace checks the line against the revenue plan, so neither of them moves it alone. When the line moves it never moves far at once — it applies to new signups first. Cutting the limit out from under people already working inside it produces cancellations rather than payments.

The most common belief is that a narrower free portion produces more payments. Blocking earlier puts more people in front of the payment screen more often, so it reads that way. But payment does not follow from seeing a payment screen; it follows from needing the product. Block before activation and the user sees a price while still not needing anything, and simply leaves. Conversion going up and revenue going down in the month after a cut is a real outcome: the share of remaining people rose while the remaining people themselves got fewer.

The second is reading it as the lenient option and therefore worse for revenue. The word soft sounds like a concession, and the payment rate genuinely does come out lower than blocking at the door. The two rates have different denominators. Block at the door and only people who already intend to buy sign up; open a free portion and the browsers join the signup count too. Put 30 percent and 5 percent side by side and you pick the one that brought in two and a half times less money.