How to Outline an Article Before Writing

The outline is drafting's highest-leverage hour: structure decided while it's still cheap to change, coverage verified against what actually ranks, and the blank-page problem dissolved before it starts (a good outline makes the draft transcription; a missing one makes it exploration at typing speed). Here's the outlining method for search content — SERP-derived, reader-ordered, and sized to the piece — plus the failure patterns that produce well-written articles about the wrong things.

The method, in five passes

1. Harvest the coverage floor: the top results' H2/H3s side by side (manually or via the research-panel tools) — what every winner covers is the floor; appearing without it means competing incomplete, per the coverage rule.
2. Harvest the question layer: People-Also-Ask and the brief's subquestions — each real question is a section candidate and a snippet ticket; the FAQ block collects what doesn't earn full sections.
3. Add the angle's sections: the brief's differentiator translated into structure — the comparison table nobody else has, the data section, the honest-limitations block: if the angle isn't visible in the outline's headings, it won't survive drafting.
4. Order as the reader's journey, not the research's: what/why → how → edge cases → questions (the pattern per format — how-tos, comparisons and listicles each carrying their own canonical order); research notes arrive grouped by source, readers need them grouped by need — the re-sort is the outline's actual intellectual work.
5. Write the headings as promises: each H2 a specific, scannable claim (the formula craft at section scale — "The five fixes, in impact order" beats "Solutions") — because headings are what skimmers read, TOCs display, and snippet systems extract.

Sizing and testing the outline

Depth follows the piece: a 900-word answer wants four H2s; a pillar wants the full tree with H3s — the outline's section count times realistic section length is the length estimate, which is how outlines catch scope problems the brief's range only implies. The one-minute test before drafting: read the headings alone, in order — do they tell the whole story, keep the promise of the title, and show the angle? An outline passing that test produces a draft that edits lightly; one failing it produces the restructure-after-drafting that costs triple. And the outline travels: approved outlines are the checkpoint that makes freelance workflows cheap to correct — five minutes on headings versus an hour on a drifted draft.

Frequently asked questions

Doesn't outlining from the SERP just produce the same article as everyone's?

The floor does — which is why pass three exists: the SERP supplies completeness, the angle supplies difference, and the failure mode is skipping either (incomplete-but-original loses to the coverage bar; complete-but-identical loses the angle test). The outline is precisely where both get enforced at once.

How detailed should outline bullets be under each heading?

One to three cues per section — the claim, the evidence to use, the link to place — enough that drafting is filling, not deciding. Full sentences under every heading is drafting in disguise; bare headings alone re-import the blank-page problem one level down.

Can AI generate my outlines?

It compresses passes one and two genuinely (the coverage-and-questions harvest — the where-they-help column); passes three through five are the judgment layer — angle, order, promises — where generated outlines converge on the SERP's average, which is the article that ranks last among complete ones. Tool for the floor, human for the difference — the silo's standing division, all the way down to the rankings themselves: craft plus authority (our outline's simple).