| name | ctr-snippet-batch-optimize |
| description | Batch-rewrites page titles, meta descriptions, and matching on-page headlines (including JSON-LD where the site duplicates the title) using tiers from Google Search Console data—prioritizing high impressions with zero or low CTR on page 1–2, then deeper positions. Encodes repeatable CTR patterns from production workflows: numbers and list counts in titles, bracket qualifiers, question-led descriptions, concrete examples, deduped title phrasing, and current-year refresh on time-sensitive pages. Use after a GSC audit or alongside gsc-ahrefs-browser-audit output when the user wants implementation, not just research. Trigger when the user mentions batch title tags, meta rewrites, GSC tiers, zero-CTR pages, snippet optimization, or SEOLayout-style title/description updates across many URLs. |
| license | MIT |
| compatibility | Requires access to the site codebase (React/Vue/Blade/etc.). No browser required unless you are still gathering GSC data. Respect the project’s existing SEO component conventions (e.g. layout props, head manager, CMS fields). |
| metadata | {"version":"1.0","source":"claude-code-hearth-sessions-657b0567"} |
CTR snippet batch optimize (from GSC tiers)
What this is
This skill implements what worked in practice: after you know which URLs underperform in Search Console, you batch-update titles, meta descriptions, and schema/article headlines so they match intent and earn clicks. It is the implementation complement to research skills such as gsc-ahrefs-browser-audit (which produces the tier lists).
Inputs you need
Either:
seo-work-brief.md at the site project root (from gsc-ahrefs-browser-audit) — use the Tier 1–3 tables and ignore the ## New guide backlog section (that section is for bulk-seo-guides-from-keywords), or- Any other finished report with URL paths, impressions, clicks, CTR, position, or
- Instructions to run or reuse a GSC Performance export / audit first, then proceed.
Tiering (use the same logic as the winning workflow)
| Tier | Signal | Primary lever |
|---|
| 1 | High impressions, ~0% CTR, position roughly 5–15 (page 1–2) | Title + meta rewrite (biggest ROI) |
| 2 | Page 1–2 but CTR clearly below typical band for that position | Title + meta; check weak intent match |
| 3 | High impressions but deep position (e.g. 25+) | Meta tuned for future lift; content depth and internal links matter more than the snippet alone |
Sort work by impressions within each tier so the largest gaps are addressed first.
Execution steps
- Normalize URLs — Strip domain to paths (e.g.
/guides/example) so you can search the repo.
- Locate files — Grep/route tables/CMS map: find the component or template that sets
<title>, meta description, and any schema headline that mirrors the title.
- Read before edit — Batch-read the relevant files (or the shared layout wrapper) so you do not fight per-page conventions.
- Apply patterns — Follow references/ctr-patterns.md on every targeted page unless a pattern would misrepresent the content.
- Keep schema and UI in sync — If the page exposes
headline / title in JSON-LD or hero H1 from the same string source, update once in the canonical place or update all mirrors consistently.
- Verify — Spot-check a few pages in the dev server; ensure no duplicate
<title> boilerplate (e.g. “Examples: Definition and Examples”).
Scope control
- If the user names a maximum page count, respect it (e.g. top 20 Tier 1 only).
- Do not invent traffic numbers; use only figures from the supplied audit or freshly pulled GSC data.
- Do not keyword-stuff or promise medical/legal/financial outcomes in snippets.
Related
