Freemium

When software was sold before it was used, buyers decided first and found out later. A brochure, a demo, a signature on an annual contract, and only then did the product arrive — which is when it became clear whether it fit the work. Approval meetings had the same shape every time. Someone brought a proposal saying the tool looked good, the person holding the budget asked what was good about it, and the answer was whatever the sales rep had said, because nobody in the room had used it.

The structure was heavy on the selling side too. Winning one account for a tool at $8 a month — calls, material, sometimes a visit — costs more than that account pays in a year. Cheap tools either didn't get built or got built expensive, for the few customers large enough to justify the effort.

Distributing software over the internet changed the arithmetic. Adding one more user stopped costing a box and a shipment and started costing a corner of a server. Once that was true, a model became possible where you leave the door open and collect from a fraction of the people who walk in. Freemium is the name for that arithmetic. The word joins "free" and "premium," and it describes a product that carries both stretches inside it.

You draw one line between what people can use for free and what they have to pay to open, and you draw that line across features rather than across time. The free stretch has no expiry date, so someone who stays inside the line keeps using the product for free indefinitely.

Three decisions define the line.

1. What the free stretch includes. The usual instinct is to leave everything a single person needs wide open. Pagedeck puts no cap on how many documents you create and caps invitations at three teammates.
2. Where the charging starts. That point is the paywall — the place where progress stops until someone pays. At Pagedeck it sits on the invitation screen, at the moment you add a fourth teammate.
3. What share of free users will pay. This number checks the other two.

The third one is where the arithmetic bites. Two or three paying customers out of every hundred free users is an ordinary result. Say Pagedeck has 10,000 free users and 3 percent of them convert. That's 300 paying users at $10 a month, or $3,000 in monthly revenue. The remaining 9,700 still need servers and storage, and at $0.25 per free user per month that's $2,425 going back out. What's left is $575. Let conversion slip from 3 percent to 1 percent and revenue falls to $1,000 while the cost of carrying everyone barely moves, which turns the month into a loss.

So where the line sits is close to the whole game. Keep the free stretch too narrow and you collect people who poked at the product and left. Keep it too wide and nobody has a reason to pay. The usual placement is at the boundary between working alone and working with others: put the features that only start to matter as a team grows on the paid side, and using the product well carries people toward the line.

It fits products that one person tries alone and then recommends to the desk next to them — collaborative document tools, file storage, developer tools. In these products the next customer arrives because of the product rather than a sales rep, and that pattern, where using the product drives both spread and purchase, is called product-led growth. A free user shares a document, the person who receives it sees the product for the first time, and that person starts out free as well.

The places it doesn't fit are just as clear.

1. Products with a real cost per action. Anything that transcodes video or runs an AI model can burn a few dollars a day on a single free user, which pushes it toward usage-based pricing, where you're billed for what you consume.
2. Products with a small universe of possible buyers. If the market holds 200 companies, there's no reservoir of free users to collect from. The model works by gathering a hundred people to keep two or three.
3. Products people use once. A service built around moving house never gets the second visit where the free version starts to feel limiting.

The shape of the spending changes as well. Money that would go to advertising elsewhere goes into the servers carrying free users. When signups spike, costs rise first and revenue follows a quarter or two later, which is why the paid stretch usually runs on a subscription — a fixed amount on a fixed date is what makes the money spent up front recoverable.

"What if the free plan capped invitations at three and the fourth teammate triggered the upgrade?"

A proposal to put the paywall on headcount rather than on a feature.

"Going freemium raises the server bill before it raises revenue. We need 3 percent conversion to carry it."

Free users cost money before any of them pay, so the conversion assumption has to be made explicit up front.

"This one stays free. If people hit a wall here, they never get as far as showing it to their team."

Features that sit on the path of spread don't go behind the paywall.

Pagedeck revisits the line every quarter. Last quarter Ravi brought numbers showing that signups following an externally shared document ran four times higher than any other path, so external sharing stayed free. Grace brought the server cost trend and proposed moving the line instead: drop the 50-document cap and start charging at the fourth teammate.

Mia's summary was one sentence. What you do alone stays open, and what you do together is what we charge for.

The most common mistake is reading free users as customers who cost nothing. They contribute nothing to revenue, so that's how the spreadsheet presents them. In reality they consume servers and support time while also acting as the channel that delivers the next paying customer. Look only at the cost side and the decisions that follow all point the same way — stop answering questions from free accounts, quietly shrink what the free plan includes. The server bill drops that month, the people who used to recommend the product leave first, and two or three quarters later paid signups dry up with them.

The second mistake is assuming a more generous free plan always produces more paying customers. More people inside the door sounds like more people eventually crossing the line. But conversion depends on where the line sits, not on how many people are standing behind it. Widen the free plan until it covers everything a small team actually needs and nobody has a reason to cross. Teams that do this watch signups climb 40 percent while conversion falls from 3 percent to 1.5 percent, leaving revenue flat and the server bill higher.