Freemium Trades Friction for Friction

You took the price off so more people would try it. That worked. Signups went up, the number looks good in a screenshot, and somewhere in the last few months you noticed that almost nobody is upgrading. The usual reading is that the paywall is in the wrong place. I think the problem happened much earlier than that, on the day you removed the price.

Conversion has levels, and the argument there was that stranger to user and user to customer are two separate events, with the second one at least as hard as the first. This post is about why that gap is so wide when the first event is free. The short version: friction and commitment are the same dial, and you cannot turn one down without turning the other down with it.

Friction and commitment are the same dial

Newsletters are the cleanest place to see this, because both versions exist and everybody has been on both sides of them.

A plain signup form with no confirmation step gets you more addresses. Anyone can type anything, nothing is verified, and the list grows faster. Add double opt-in and you lose a slice of those people at the confirmation email. The list is smaller. It is also made entirely of people who went to their inbox, found your message, and clicked a link to confirm they meant it.

Nobody who has run both is confused about which list is worth more per name. The step you removed was doing work. It was filtering. What you experienced as friction was, from the other side, a person demonstrating that they meant it.

Every step you remove to get more people also removes the evidence that any of them meant it.

The same trade shows up everywhere once you look for it. Job applications that take ten seconds produce a thousand applicants and no signal. A meeting nobody had to schedule is a meeting nobody prepared for. Free tickets have famously terrible attendance rates. There is nothing special about software here, and that is the point.

A free signup is a soft yes

Someone who signs up for free has given you a soft yes. They are trying it. They have risked nothing, spent nothing, and committed to nothing, and no one will ever ask them to justify the decision. They can forget it happened by Thursday and there is no cost to them in doing so.

Someone who pays has given you a hard yes. They spent money, which means the decision had to hold up, at minimum to themselves. In a company it means they may have to defend it to somebody with a budget.

Those are two different people. Frequently they are two different people inside the same body, thirty days apart.

The free user may never have looked at your product very hard. They may have poked at the onboarding, gotten interrupted, and closed the tab. They may not have tried very hard at the part that was difficult, because nothing was riding on it and there was always tomorrow. You have their email address. You do not have their attention, and you never charged them anything that would have bought it.

So the second conversion starts further back than you planned

The freemium theory is that people try it, see the value, and upgrade. Read it again and notice what it assumes. It assumes the trying happened.

When the trying did not happen, the job at level two is not what you budgeted for. You are not converting an engaged user into a paying customer. You have to convert a disengaged person into an engaged one first, and then convert that person into a paying one. Two conversions hiding inside the one you planned for, and you have to run both with somebody who has already demonstrated that they are not thinking about you much.

The mechanism that would have produced the engagement was the friction, and you removed it on purpose, at the start, for a good reason. That is the trade. More people at the top, less commitment per person, and a second conversion that begins much further back than the funnel diagram suggests.

The enthusiasm curve runs the wrong way

Freemium quietly assumes enthusiasm grows over time. Someone arrives mildly curious, uses the product, gets more convinced, and eventually cares enough to pay.

In my experience enthusiasm is highest on the day somebody signs up. That is the day they found you, read the page, recognized a problem they actually have, and did something about it. Thirty days later they have been at work, the problem has receded into the pile of things they tolerate, and your product is one of eleven tabs they have not opened.

So the model asks for money at the point where interest is lowest, having removed the thing that would have kept interest high. And the assumption that interest grows rests on an assumption that they use it, which rests on them having some reason to, which is the thing you removed in the first place. The reasoning runs in a circle and the circle is where the revenue was supposed to come from.

What a credit card at signup actually buys you

Put the card up front on a free trial and two things happen.

Fewer people start the trial. That cost is real and I am not going to talk you out of it. If you are pre-launch and trying to find out whether anyone wants this at all, the raw number of people who will engage matters more than what each one is worth, and a card wall will hide that answer from you.

The second thing is the one people miss. The conversion event stops being an event.

On day eleven of a fourteen day trial you send one email. I hope this is working out for you. In three days your card will be charged the amount we agreed when you signed up, and here is the link if you would rather not. Then you do nothing, the customer does nothing, and it converts. The default is yes.

Compare that with the version where no card was collected. Now the customer has to decide to pay, find their wallet, type sixteen digits, and do all of it on a day you did not choose, at a moment when their interest in you is lower than it has ever been. Everything your landing page and your demo and your onboarding accomplished has to be accomplished again, by the customer, alone, from memory. All of that work collapses into one moment and the moment is theirs, not yours.

The game you end up playing

There is a second cost to freemium and it shows up in your roadmap rather than your dashboard.

Once revenue depends on people hitting a wall, you start designing walls. Features that cost you nothing to serve get moved behind the paywall, not because serving them is expensive but because you know people will eventually need them. The genuinely useful configuration is always one tier up from wherever the user is standing. You are building a funnel, and from the inside it looks like packaging.

Users can feel it. It feels like a trap because it is one, and the moment somebody notices, the relationship turns adversarial. You are hiding the ball and they are trying to find the ball, and neither of you is building anything while that is going on. Your support inbox fills with people asking why an obvious thing is missing, and you have to answer in a way that is technically true.

That is the slow slide I wrote about in enshittification, with one difference. Those companies drifted into it over a decade under pressure. A badly designed free tier builds it into the product on day one, before there is any pressure at all.

When the free tier is telling you something

Sometimes the answer is that the thing should be free, and the free tier is how you found out.

If thousands of people sign up and essentially nobody pays, you have learned something specific, which is more than most experiments give you. It is usually one of two things.

Either what you built is good and not necessary. People will happily use it and will not reorganize a budget for it, which is the version I wrote about in why your SaaS has zero users. Or the valuable part is the part you are giving away and the part you are charging for is decoration, which is a scoping problem rather than a pricing one.

Both are worth knowing and neither gets fixed by moving the paywall. The free-user pile is a reading, and the reading is about the product. Treat a year of it as expensive research you already paid for, then go and change the product or change who it is for.

Where freemium genuinely works

I am not arguing against product-led growth. It built several of the largest software companies in the world and the pattern is real. It is worth being precise about the conditions, because that is what separates a strategy from a hope.

The free tier is the distribution. Each free user brings paying users with them. Collaboration tools, anything with a network effect, anything where one person invites four colleagues and one of those colleagues has signing authority. The free account is doing marketing work, so it earns its cost.

The ceiling is real and people reach it while they still remember signing up. Storage, seats, volume, message counts. This works when the limit tracks the value someone is actually receiving and arrives early. It fails when the limit is artificial, because an artificial limit is the wall-building problem above wearing a hat.

The paid version is a different product for a different buyer. Free for individuals, paid for organizations that need control, audit trails, or somebody to call. The individual was never your customer. They were your recommendation.

You can afford the wait. A long gap between signup and first payment is normal in credit-based and usage-based products, and reading it as a broken funnel is a mistake I have watched people make with their own healthy business. It only becomes a problem if the runway is shorter than the gap.

What those have in common is that the free user is doing something for you beyond existing. If your free tier’s only job is to demonstrate value, you are running a demo and paying for the hosting.

None of this means charge everybody from the first second. It means pricing is not a decision you make after the product works. It is part of the product, it determines who shows up, and it determines how much of themselves they bring when they do.

Frequently Asked Questions

Why don't free users upgrade to paid?

Usually because the free signup never produced much engagement to convert. A free account costs nothing and commits the person to nothing, so many of them never use the product hard enough to learn whether they need it. That leaves you converting a disengaged person rather than a convinced one, at a point where their interest is lower than it was the day they signed up. The paywall's position is rarely the actual problem.

Should I require a credit card for a free trial?

If you have a working product and you are trying to find paying customers, usually yes. Collecting the card turns the conversion into a default rather than a decision, so the customer converts by doing nothing instead of having to re-sell themselves at their least enthusiastic moment. The cost is fewer trials started. The exception is early validation, where you need to know whether anybody wants this at all and a card wall hides that answer.

How do I know if my free tier is too generous?

Look at whether your paid features solve a problem people currently tolerate, or a problem they are actively working around. If the free tier already handles the part that hurts, the paid tier is selling convenience to people who have already stopped hurting. The other tell is your own roadmap: if you are moving features behind the paywall because users will eventually need them rather than because they cost you anything, you are building a wall rather than a product.

What free-to-paid conversion rate should I expect?

Your own rate, broken down by traffic source, is the only one worth planning against, because published benchmarks average across products and channels that behave nothing alike. A site-wide average hides the fact that one source converts many times better than another, which is the only thing you can act on. Conversion Has Levels covers how to instrument that and why the rate by source is the number that matters.

Is freemium a good model for a small SaaS?

Only when the free user does something for you beyond existing. That means bringing other users with them, hitting a real capacity limit early, or being a different person from the buyer you are actually selling to. If the free tier exists purely to demonstrate value, it is a demo you are paying to host, and a time-limited trial will teach you the same thing faster and cheaper.