Document SEO draft publishing plan

README.md

@@ -7,6 +7,8 @@ Migration target for `blog.hichee.com`, moving from WordPress to Astro + Cloudfl
 - Repository remains **private** during migration.
 - WordPress drafts are migrated as unpublished content under `src/content/drafts/`.
 - We create one GitHub issue per unpublished draft (`draft-review` label) so nothing gets lost.
+- Drafts are not bulk-published for SEO; use `docs/seo-draft-publishing-plan.md`
+  to triage and rewrite selected candidates.
 - Media strategy defaults to object storage/CDN (Cloudflare R2), not repository storage, unless final measured media size is small enough.
 
 ## Local Setup

docs/seo-draft-publishing-plan.md

@@ -0,0 +1,182 @@
+# SEO Draft Publishing Plan
+
+Last reviewed: 2026-07-03 HST
+
+## Recommendation
+
+Do not bulk publish the migrated WordPress drafts for SEO.
+
+Selective publishing can help, but only after editorial review, media cleanup, and
+conversion planning. The draft collection should be treated as raw material for
+curated search pages, not as ready-to-index inventory.
+
+## Why Bulk Publishing Is Risky
+
+The local draft corpus currently contains 717 markdown files under
+`src/content/drafts/`.
+
+Snapshot from the local corpus:
+
+- 710 drafts, 6 future posts, and 1 private post.
+- 714 posts and 3 pages.
+- 312 drafts are under 250 words.
+- 374 drafts are under 500 words.
+- 186 drafts are 1,200 words or longer.
+- 691 drafts still contain absolute `https://blog.hichee.com/wp-content/...`
+  media references.
+- 61 drafts are missing a featured image field.
+- The most common categories include imported or low-control sources:
+  `rss`, `uncategorized`, `security`, `superhog`, `guess-where`,
+  `property-spotlight`, and `podcasts`.
+
+Observed quality issues:
+
+- Many drafts are thin placeholders, empty quote posts, or very short imported
+  snippets.
+- Several drafts have placeholder or broken titles such as `Private:` or
+  `Untitled post ...`.
+- Many long drafts appear to be RSS imports, syndicated partner content, or
+  podcast transcripts. Length alone does not make them good SEO pages.
+- Some older drafts include stale destinations, pricing, laws, or external
+  links that need fact checks before publishing.
+- Media references must be normalized before publication so published pages do
+  not depend on old WordPress absolute URLs.
+
+Google Search guidance favors helpful, reliable, people-first content and warns
+against scaled or low-value pages created primarily to attract search traffic.
+Use these as the editorial guardrails:
+
+- [Creating helpful, reliable, people-first content](https://developers.google.com/search/docs/fundamentals/creating-helpful-content)
+- [Google Search spam policies](https://developers.google.com/search/docs/essentials/spam-policies)
+- [Canonicalization and duplicate URL guidance](https://developers.google.com/search/docs/crawling-indexing/consolidate-duplicate-urls)
+
+## Positive Traffic Definition
+
+The goal is not more sessions in isolation. The goal is more qualified traffic
+that can plausibly become HiChee users, subscribers, partners, or brand-aware
+travel shoppers.
+
+Prefer pages that connect search intent to HiChee's value:
+
+- Saving money by booking direct.
+- Comparing Airbnb, Vrbo, hotel, and direct-booking prices.
+- Finding destination-specific vacation rental options.
+- Understanding guest trust, host safety, and short-term rental risk.
+- Planning trips where HiChee can help users compare stays.
+
+Avoid publishing pages that may increase low-quality traffic but do not advance
+the business:
+
+- Generic travel inspiration with no HiChee angle.
+- Syndicated RSS content without original commentary or canonical decisions.
+- Thin quote/image posts.
+- Old copied content from partner blogs unless rewritten and permission/canonical
+  status is clear.
+
+## Draft Triage Buckets
+
+Classify every draft before publication:
+
+- `publish-candidate`: strong search intent, original enough, useful to HiChee,
+  and feasible to clean up quickly.
+- `rewrite`: promising topic, but needs a stronger angle, updated facts,
+  rewritten structure, or consolidated duplicate material.
+- `merge`: useful fragments that should be folded into a stronger pillar page
+  instead of published as separate thin pages.
+- `syndicated-review`: RSS, partner, podcast, or externally sourced content that
+  needs permission, canonical, noindex, or substantial original analysis.
+- `archive`: not worth publishing because it is too thin, stale, duplicated, or
+  irrelevant.
+- `media-cleanup`: otherwise publishable, but blocked on image migration,
+  alt text, or broken embeds.
+
+## Publishing Checklist
+
+Before moving any draft from `src/content/drafts/` into published content:
+
+- Confirm the page serves a real search intent aligned with HiChee.
+- Replace `/wp-####/` placeholder paths with a clean, durable slug.
+- Update title, description, headings, and internal links.
+- Add original HiChee perspective, examples, and conversion path.
+- Remove duplicate RSS boilerplate and unrelated feed imports.
+- Verify permission, attribution, and canonical/noindex handling for syndicated
+  content.
+- Update facts, dates, legal claims, prices, platform policies, and destination
+  details.
+- Normalize media to root-relative or durable CDN-backed URLs; do not publish new
+  absolute `blog.hichee.com/wp-content` references.
+- Add meaningful image alt text where appropriate.
+- Link to relevant published HiChee blog posts and HiChee product pages.
+- Preview mobile and desktop rendering.
+- Build locally before merge.
+
+## First SEO Sprint
+
+Start with a small curated batch instead of a broad publication sweep.
+
+Recommended first sprint:
+
+1. Generate a draft audit report with word count, category, title quality, media
+   risk, and duplicate/syndication indicators.
+2. Select 25 candidates manually.
+3. Choose 10 high-intent pages for the first publishing batch.
+4. Rewrite and publish those 10 pages with clean slugs and conversion CTAs.
+5. Submit or inspect indexing in Google Search Console.
+6. Monitor impressions, queries, clicks, click-through rate, indexed status, and
+   downstream conversion signals.
+7. Use the results to decide the next batch.
+
+Good first-batch themes:
+
+- Direct booking and saving money on Airbnb-style stays.
+- Airbnb alternatives and booking-site fee comparisons.
+- Destination guides where HiChee can connect readers to stay comparisons.
+- High-quality vacation rental listicles that can be made original and current.
+- Guest safety, host trust, and short-term rental risk explainers.
+
+Avoid first-batch themes:
+
+- Empty travel quote posts.
+- Raw RSS reposts.
+- Raw podcast transcripts without a summary, original analysis, and clear
+  canonical strategy.
+- Generic old travel articles with no HiChee-specific value.
+
+## Operating Model
+
+Use GitHub issues with the existing `draft-review` label as the intake queue.
+Add more granular labels as drafts are reviewed:
+
+- `seo-candidate`
+- `rewrite`
+- `merge`
+- `syndicated-review`
+- `archive`
+- `media-cleanup`
+
+Publishing should happen through normal pull requests. Each PR should include:
+
+- The selected draft issue link.
+- The intended search intent.
+- The target query cluster, if known.
+- The old draft file path.
+- The new published path.
+- Media cleanup notes.
+- Local build verification.
+
+## Success Metrics
+
+Measure after publication in Google Search Console and analytics:
+
+- Indexed pages and indexing errors.
+- Search impressions by page and query.
+- Click-through rate.
+- Average position trend.
+- Organic sessions that reach HiChee product or conversion pages.
+- Engagement quality compared with existing blog traffic.
+- Pages that attract irrelevant traffic and should be consolidated, rewritten, or
+  removed.
+
+Do not continue publishing drafts merely because unpublished inventory exists.
+Continue only when the prior batch creates qualified traffic or useful search
+visibility.