The content treadmill and the half-life of a post.
Most things you publish are radioactive in the bad sense. They decay. A post goes up, gets its burst of attention over a day or two, then the curve falls off a cliff. A small minority do the opposite. They keep picking up readers, for years. The difference between the two is almost never quality. It is structural. Once you can see the structure, you get to choose which kind of thing you are making before you make it.
Two kinds of content, sorted by decay curve
Half-life content decays exponentially. It is pegged to a moment: a launch, a trend, a news item, a hot take on the thing everyone is arguing about today. Value is highest the second it is published and halves every day or two after. Forty-eight hours on, it is basically inert.
Compounding content does the reverse. It is pegged to a recurring question. Something a steady trickle of people will keep asking, this month, next year, the year after. It often gets a smaller launch spike. Then, instead of decaying, it climbs, because every day delivers a fresh batch of people newly asking the question it answers.
Half-life content rents attention. Compounding content owns a question.
The treadmill
The trap is that half-life content feels more productive. It is faster to make. It pegs neatly onto whatever is loud today. The spike arrives basically immediately, which feels like a reward, so you make another. And another. That is the treadmill. A publishing cadence you have to sustain at full speed forever, because the moment you stop, the back catalogue, being radioactive, has decayed to nothing. You are only as visible as your last 48 hours have made you.
The treadmill is not a moral failing. It is a portfolio with no compounding assets in it. Every piece is a firework. Fireworks are fine. A sky made entirely of fireworks means you have to keep standing there in the dark, lighting them, forever.
The test, before you write a word
One question separates the two kinds, and you can ask it at the idea stage, before you have committed anything to the page.
"My reaction to today's announcement" fails the test. Nobody searches that in eighteen months. "How to read a research paper quickly" passes easily. People will need that answer more or less forever. Same effort either way. Completely different decay curve. The difference is decided at the idea stage, not in the writing.
A portfolio, not a feed
The healthy version is not "only write evergreen things." Fireworks have real uses. They are how you reach people who do not know you exist, and timeliness is genuine value, not a sin. The healthy version treats your output as a portfolio, with a ratio you actually chose.
| Type | Its job | Rough share |
|---|---|---|
| Compounding | The assets. Owns questions, climbs for years. | ~60% |
| Half-life | The reach. Catches today's attention, points it somewhere. | ~30% |
| Experiments | Cheap bets on formats and topics you have not tested. | ~10% |
The move that ties it together: make the half-life pieces link to the compounding ones. A firework's job is not only to be seen. It is to throw a bit of light on the assets standing quietly out there in the dark. A timely post that sends a slice of its spike toward an evergreen post has converted rented attention into something that keeps paying rent back.
The slow part
Compounding content is slow to feel rewarding. The post that will have ten thousand readers across three years might have ninety on day one. Judge it on day one and you will conclude it flopped, and crawl back onto the treadmill, which at least pays its small reward on schedule.
The real discipline is about measurement windows. Judge a firework after a week. Judge a compounding piece after a year. Different rulers for different instruments. Otherwise the fast ruler will quietly talk you out of ever building anything that lasts.
This post, for whatever it is worth, was written to its own test. The question (why does some content last and most of it does not) is one people keep asking. So if it is working, you are reading this a long while after I published it, and the date up at the top is older than it feels.