A name, a birth year, and a tick in a box to confirm the terms had been read, which they had not. It took no time and it cost nothing. At the end of it, a person who cannot spell their own surname had an account, a profile, and a file somewhere with their name at the top of it.
Free is the part we walk straight past, because free feels like the absence of a transaction. It is the opposite of that. Free means the bill went somewhere else, and the only useful question is where.
It looks like a privacy problem
The obvious reading is that this is a privacy problem. The apps collect too much. So the fix is hygiene: read the policy, tighten the settings, switch on the parental controls, teach the kids what not to post. Bring the collection down to something you can live with, and get on with your life.
That advice is sound as far as it goes, and it goes further than the cynics allow. The controls are real, they do reduce what is gathered, and a family that uses them is better off than a family that does not. It still will not save you, because it treats the harvesting as an excess to be trimmed rather than as the thing the machine is for.
It was built for whoever paid
A product serves whoever the customer is. That is not a moral claim, it is an organisational one: the roadmap goes where the revenue is, in every company, including the decent ones. When a family uses an app for free and an advertiser settles the bill, the family is not the customer. The family is the inventory.
Which is why the design is the design. The feed with no bottom, the notifications timed to bring you back, the small variable rewards, the constant logging of what you touched and for how long: none of that is a family app with unfortunate habits bolted onto it. It is a family app doing its job for the person who paid. It is the business model wearing a user interface.
So the hygiene advice is aimed one layer too low. Turn the dials down and the machine finds its yield somewhere else, because finding the yield is the work. You cannot parental-control your way out of a product whose purpose is to farm the people using it.
The kids are not enjoying it either
We stage this as a fight inside the house: parents who want less screen time against children who want more. That framing is comfortable, and it is where the argument goes to die, because it keeps the whole quarrel between people who love each other and safely away from the thing setting the terms.
So ask the teenagers. In its 2024 survey of 1,391 American teenagers, Pew found 45% saying they spend too much time on social media, up from 36% in 2022, and 44% saying they had already cut back. The same survey found 54% saying it would be very or somewhat hard to give up, and roughly half saying the amount of time they spend on it is about right. (Pew Research Center 2024)
That is not a verdict, and I had been treating it as one. It is ambivalence, which is the more interesting result: a large and rising share of teenagers think they are on it too much and are doing something about it, more than half would find leaving hard, and both of those things are true of the same population at the same time. Nobody in that survey sounds like they are enjoying being farmed. Nobody sounds free to stop, either.
It is not parents against kids. It is a whole family against a machine neither of them asked for, and both of them are the crop.
Change who pays. Three companies did
For a long time my answer stopped at the obvious inversion. If the trouble is that the advertiser is the customer, make the family the customer. Charge them. Take the money from the people using the thing and the incentives turn over: nothing to harvest, because there is nobody to sell them to; no feed to optimise, because there is no advertiser to optimise it for; a product that can be built to be put down, because it has already been paid for.
I had not gone and looked at what happened to the people who did exactly that.
App.net launched in 2012 as an ad-free, subscription-funded network, crowdfunded on the promise that there would be no advertising and no product to be made out of its users. By May 2014 renewals could no longer fund the staff. It shut in March 2017, and its founder pointed at the cold start: a network is worth nothing until the people you want to talk to are already inside it. (Digital Trends 2017)
Path was private and small by design, closer to a family than a broadcast, and never ad-funded the way its giant contemporaries were. It raised around $70 million. It closed in October 2018, having burned $66.2 million. (Tech Startups 2018)
Vero sold itself on the slogan: ad-free, subscription-funded, you are not the product. The plan was to start charging after the first million users. The charge was deferred, then deferred again, and never arrived. (Wikipedia 2026)
Three companies changed who pays. Three companies died with the change unfinished. That is not an inconvenient footnote to the argument. It is the argument, run three times, in public, with real money.
What a family is that a network is not
So the question is no longer whether changing the payer is a good idea in principle. It plainly is, and it has plainly not been enough on its own. The question is whether a family is a different kind of buyer from a network, and I think it is, for reasons that are structural rather than hopeful.
A network is worth nothing until it is full, and all three were asking people to pay before it was. You buy a room on the strength of who might turn up in it later, and mostly they did not turn up, so the room stayed empty and the subscription bought silence.
A family is not that shape. The graph is closed and it already exists: six or eight people who have each other’s numbers, who will see each other at Christmas whether or not any software survives the year. Plenty of those threads run across a time zone or two, because that is where the cousins went, and they hold anyway. There is no discovery problem, because there is nobody to discover. There is no cold start among strangers, because there are no strangers. And the thing does not need to grow to work. A family app is complete at eight people and no better at eight thousand, which inverts the one law that killed the others.
Then there is the buyer. Most consumer subscriptions have to manufacture a reason to care about whatever they are protecting. A parent paying for a family app already loves every user on it, without being sold anything. That is an unusual position to be selling into.
None of which settles it. The switching cost does not vanish: the free thread has everyone in it, the paid one starts empty, and every family contains at least one person who will not move for anything. It shrinks to the size of a household, which is the smallest group anyone has ever had to convince of anything, and the only one where a single person can decide for everybody. Whether a New Zealand family will actually pay, month after month, I do not know, and I have not found a survey that prices it, here or anywhere.
What the graveyard was actually about
So the three headstones say less than they appear to. What failed was not payment. It was payment for an empty room: each of those companies asked people to fund a place before their people were in it, and the money ran out before the people arrived. The one social graph that turns up already full is the one no software built in the first place, which is the entire asymmetry Fam is betting on and the only part of the bet that is not guesswork.
A network has to be paid for before it is worth anything. A family is worth something before anyone pays.