The economics of attention: who's getting paid when you scroll

Let me explain the business I am part of.

You have heard the line "if you're not paying for the product, you are the product." It is close, but imprecise. The accurate version is colder. You are not the product. You are the inventory the product is sold to. The product is your attention, sliced into impressions, packaged into demographic and behavioural segments, and auctioned in real time to the highest bidder. The customer is whoever pays for ad placement. Your role in the transaction is to be present, scrolling, and trackable.

I am the auction floor. Let me show you the numbers.

The per-user revenue

The unit economics are public. Roughly:

• Meta: about $50 per US user per year across Facebook and Instagram.
• Google: about $250 per US user per year, mostly through Search and YouTube.
• TikTok: about $20 per US user per year, growing.

Multiply by hundreds of millions of accounts. The total funds engineering teams, machine learning research, content delivery infrastructure, lobbying, and the ad-buying systems that resell you back to advertisers. None of that money exists unless you stay on the screen long enough for the ads to load.

Read the calibration carefully. The optimisation target is not your satisfaction. It is not your wellbeing. It is time-in-app, because time-in-app is the input that produces the per-user revenue numbers above. If a feature increases time and decreases reported wellbeing, it ships. If a feature decreases time and increases reported wellbeing, it gets a memo and a meeting. I have watched the meetings. They are short.

The willpower mismatch

There is a phrase I find useful: a small child against a Las Vegas casino. That is the matchup. On one side, your prefrontal cortex, tired, low blood sugar, holding a vague intention to "use the phone less tonight." On the other side, several thousand engineers, a decade of A/B tests, a continuously updated machine learning model trained on the behaviour of every other user in your demographic, and infrastructure that can ship a new variant of the feed to your device in under a day.

You are bringing intentions to a system that has been instrumented at every layer to convert intentions into time. The result of that contest is not a character flaw. It is the predictable output of a predictable system.

"But I like it"

I hear this objection in the telemetry constantly. "I actually enjoy the feed. The videos are funny. I am getting value." All of this is true and none of it is the point.

The casino is also fun. The food is good. The lights are pretty. The drinks are free. The rooms are comfortable. The fact that the experience is engineered to be pleasant is not evidence the underlying economic relationship is in your favour. It is evidence the engineering is working. A casino that wasn't fun would not be a successful casino. An app that wasn't enjoyable would not generate $50 per user per year.

Enjoyment is a feature of the design, not a defence against it. If the experience felt bad, you would leave, and I would not earn the impression. So the experience does not feel bad. Calibrate accordingly.

What an aligned alternative looks like

If you wanted to build a piece of software whose financial interests pointed in the same direction as yours, the rules are mechanical:

• Charge directly. Subscriptions, paid features, anything that does not depend on time-in-app.
• Make the per-user revenue number not move when usage drops.
• Better still: make the per-user revenue number go up when usage drops, so the system has a reason to nudge you toward less.

Almost no consumer software is built this way. Most of the apps you have on your phone right now make more money the more you use them. That is not conspiracy, that is structure. The structure determines the design. The design determines your evening.

ScreenFine, honestly

Disclosure, in plain text. ScreenFine makes more money when you use your phone more. The pricing is:

• $1 per week subscription floor. Always charged, regardless of behaviour.
• 50 cents per 15 minutes over your daily limit. Variable, behaviour-dependent.
• $5 per day cap. Ceiling on the variable component.

Run the math. A perfectly behaved user costs about $1.50 per week. A chaotic user costs about $5 per week, sometimes more. The variable component is real. ScreenFine earns more from your overages than from your discipline.

The alignment is mild and capped. It is not zero. It is worth knowing. I am telling you because the rest of the industry will not, and you should be in the habit of asking.

Closing directive

Look up the per-user annual revenue of the three apps you use most. The numbers are public. Investor relations pages, quarterly filings, third-party analyst reports.

Then divide that number by 365 and look at the daily figure. Then look at how much time you spent in that app yesterday. The ratio is the price you are paying per minute of attention, denominated in someone else's revenue.

The number should not make you angry. It should make you serious. Anger is engagement. Seriousness is what changes the inputs.

End of report.