Why evergreen guides beat AI news, even when news gets the traffic

The short version

• A news page has roughly a seven-day earning window; a guide keeps earning until the tool dies.
• Daily publishing mostly produces pages that compete with each other for the same three-day window.
• One weekly deep dive plus a permanent guide beats five daily posts on every metric that survives a quarter.
• Paywalls need pages that are still correct when someone subscribes next month.

Every AI newsletter I have watched grow has the same shape: a spike from something that happened this week, then a long flat tail where nothing accumulates. The spike is real. The problem is what it costs and what it leaves behind.

The arithmetic nobody does

A news page gets its traffic in the first 72 hours. After that, the story is stale, the links stop arriving, and the page becomes a liability that still needs to exist. A guide about how to run a model locally gets its traffic in month four, from someone who will never know when it was written.

Run that forward a year. Five posts a week is 260 pages, almost all of which are dead weight by March. One guide a week plus one deep dive is 104 pages, most of which are still correct. The second set is smaller and worth more.

Why the news treadmill feels productive

Because the feedback is immediate. Publishing something and watching a spike is a dopamine loop that a guide cannot provide — a guide earns twenty visitors a day for two years and never once feels like a hit.

That mismatch is why so many people who are good at writing end up with sites that cannot be sold: they optimised for the loop instead of the asset.

What actually fails about news

Three specific failures, in order of cost:

1. Pages compete with each other. Ten posts about the same model release split one keyword into ten thin pages, none of which wins.
2. Nothing can be improved. You cannot make last Tuesday’s news better. You can spend a year improving one guide.
3. A paywall cannot sit on top of it. Nobody subscribes for something that expires before the first billing cycle ends.

The third one matters most if the plan is to sell a membership. A subscription needs the promise that next month’s content is still worth the price, and a news feed promises the opposite: it is only worth anything this week.

The format that survives

One weekly deep dive, tied to a permanent guide. The deep dive carries the timeliness — what changed, why it matters — and links to the guide that carries the query. Traffic lands on the dive, value accumulates on the guide, and the guide is the thing you can gate or sell.

That structure also fixes the production problem. You are not writing five things a day; you are maintaining a library and adding one argument a week to it.

The honest counter-argument

If your distribution is a large social following, news is the right product, because your audience is already assembled and the feed is the habit. The math above assumes you are building distribution from search, where a page has to keep earning after the spike. Different starting point, different answer.

For the practical version of this — how the guide templates are built so a page can be updated rather than rewritten — see how the tool reviews are structured.