dare.co.uk 404 redirects report — 2026-06-21

Daily top-50 404 paths from Cloudflare Analytics, last 24h. Operationalises feedback_404_redirect_threshold_rule — at ≥50 hits/day, examine and decide whether to redirect editorial 404s to a deserving archive page.

TL;DR

• 1173 total 404 hits in the last 24h across 50 distinct paths.
• 0 path(s) over the 50/day threshold needing a redirect decision.
• 31 bot-probe paths — leave 404 (per rule, security signal).
• 0 section-root paths — build index pages, not redirect.

Legacy editorial — decide redirect target

(19 path(s) · 452 total hits)

Bot-probe paths — leave 404

(31 path(s) · 721 total hits)

How to apply the decisions

For each row with DECIDE:

1. Inspect the path — is it a real legacy URL (old article slug, external referrer link) or accidentally-popular bot noise?
2. If editorial: pick a deserving target from the existing archive-page roots: - /methods-of-business-design/ - /culture-means-thriving-teams/ - /fields-notes-from-business-design/ - /cinema/ - /photography/ - /observations/
3. Add a 301 to ~/Code/dare-co-uk/_redirects: /<path> /<target> 301
4. Commit + push; Cloudflare Pages redeploys in ~30s.

Bot-probe paths stay 404 — redirecting them rewards scanners. Section roots get index-page builds, not redirects, because the underlying content is stunning and deserves its own home.

Generated 2026-06-21 via dare_404_redirects_report.py. Companion to feedback_404_redirect_threshold_rule.

Automated recommendation — maturity ladder

Site housekeeping wants to become a scheduled script + intelligence engine. Criteria are set in the rule (feedback_404_redirect_threshold_rule); this report applies them. Maturity ladder for how the recommendations become action:

Criteria the report already delineates (the substrate v2/v3 build on):

• Bot-probe paths → leave 404 (security signal; never reward scanners with 200)
• Section-root paths → build index page (stunning content earns its own home, not a redirect)
• Favicon paths → already-redirected verification (deploy-lag sanity check)
• Legacy editorial paths above threshold → DECIDE redirect target (the human-in-the-loop slot)
• Legacy editorial paths below threshold → monitor (wait for signal before spending decision budget)

v2 adds: a confidence score per editorial row (based on path-shape regex + historical hit pattern + the article corpus’ nearest-neighbour match for a sensible target), and a draft PR against _redirects per weekly cohort. v3 lets the engine merge the high-confidence cohort autonomously, with Dan reviewing the medium/low cohorts.

Each level inherits the same criteria — only the level of automation changes. The rule stays the source of truth; the engine becomes a faster applicator of it.