Spearleaf · Position Zero Playbook v11 · 2026-06-16 Start here Changelog
Manual

On-page discipline

Headings as a research outline, internal + external linking, schema hierarchy — the other half of why 10 links work.

On this page

"I only care about referring-domain count when first publishing. After that it's on-page + schema, then 10 at a time." This is the on-page half. The Publish quickstart is the checklist; this is the why.

Heading-tag discipline

Every heading ties to the primary keyword or is a subheading of it. Google scans trillions of pages with minimal-viable-crawl heuristics — every heading is a signal about what the page is about. "About the Author," "Read Next," "100 Millisecond Verdict" are noise that confuses the read. Web devs add H2s for visual consistency; Google cares about the semantic outline, not the font.

Write headings like a research-paper outline: H1 main topic, H2 subtopics, H3 sub-subtopics, H4 conclusion.

For an existing page, run the heading-audit prompt sequence in Content buckets → identify wasted headings → rewrite or demote.

Bonus: hand an AI an outline (vs "write me an article about X") and the output is dramatically better. The outline IS the structure.

Q&A paragraph format

Pair the outline above with a three-sentence answer paragraph under each question heading (direct answer, source, example) — the one durable lever for AI citations. That full treatment now lives in Get cited by AI.

External linking

Stop generic external linking — testing showed it does nothing.

Internal linking is a publish step

"If you don't link to your new content, why should Google index you?" On every publish:

  1. Publish the page.
  2. Link OUT from the new page to relevant existing pages.
  3. Link IN — update older content to link to the new page (the priority).
  4. Then everything else.

90 days live with no internal/external links is a deindex signal; enough deindexed pages drops your site score and you stop ranking for anything. For HTML/Eleventy sites, bake "link old → new" into the publish script — don't rely on a monthly pass.

Schema hierarchy (basics)

Single-location:

Test in Rich Results — should validate as ONE connected graph. Multi-location detail in Strategy; the deep entity/About/Mention work is in Entities + schema.

Triage — when to apply all this

Workflow: GSC → filter 11–25 → heading audit → republish → RD100 or cloud at the URL (both only if hard).