Applied guidance for the article brief and article standards. The reasoning behind how search and answer engines pick sources is in SEO and GEO concepts.
Carry information nobody else has
An article that restates what already ranks has no reason to be cited and no reason to be read. Before briefing, name the thing this piece knows that the current results don’t: your own data, a mechanism nobody explains properly, a position competitors hedge on, or a worked example at real scale.
If nothing survives that question, the honest answer is not to write it.
Write each section to answer its own heading
Retrieval systems pull passages, not pages. A section that only makes sense after the two above it is a section that cannot be quoted. Each heading is a question; the paragraph under it answers that question completely enough to stand alone.
This also happens to be how a skimming reader uses the page, so the two goals don’t conflict.
Answer in the first paragraph, then earn the rest
The reader arrived with a question. An article that spends four paragraphs establishing context before answering loses both the reader and the citation. Answer, then explain, then qualify.
Cite the source next to the claim
A statistic with no traceable source is the one that gets repeated after it stopped being true. Record where each claim came from as you draft, not afterwards from memory. This is the same discipline as the reasons-to-believe table in positioning, for the same reason.
Cap the library
A capped set of maintained articles outperforms a large set of thin ones, because maintenance is the constraint, not production. The article library overview sets the hero and supporting tiers and the cap on each.
Every new article is a maintenance commitment, not a one-off cost. Adding a twenty-first hero article means either dropping one or maintaining twenty-one badly.
Set the review date at publish
Content decay does not announce itself. Traffic drifts down, the question shifts, someone publishes something more current, and nothing in the article changes to signal any of it. The review date set at publish is the only thing that catches it.
The article maintenance process covers the cycle; the article update checklist covers the pass itself.
Link both ways
An article added without inbound links from existing articles is an orphan, and orphans neither rank nor get found internally. When you publish, update the related articles to point at it, not just the other way round.
The failures
Last modified on August 10, 2026
