[product] BorderPulse UX and findability roadmap

[sbc1-code]

Purpose

Make BorderPulse easier to understand, use, discover, and return to—while keeping the product’s evidence and data boundaries clear.

Completed in this pass — 2026-08-06

• Unified GitHub Pages deployment ownership so scheduled data fetches no longer compete with the deploy workflow.
• Removed the duplicate Pages retry path that amplified deployment_queued failures.
• Promoted Best time / Mejor hora into the primary bilingual navigation.
• Verified the local production build and prerender output.
• Closed incident #55 after documenting the Pages queue root cause.

Next product work

• Verify the next scheduled fetch and Pages deploy run end to end.
• Audit the dashboard’s first-visit path on mobile and desktop: choose crossing, understand direction, see freshness, and reach a useful detail page.
• Strengthen internal linking between live crossing pages, best-time pages, comparison pages, and bilingual blog guides.
• Improve page-level search snippets and structured metadata around high-intent crossing queries.
• Capture a small UX baseline (findability, route coverage, mobile clarity, repeat-use actions) before the next implementation pass.

Definition of better

A first-time visitor can identify the right crossing and direction in seconds; a returning commuter can get back to a saved/favorite crossing quickly; and a search visitor can move from a specific answer into trustworthy live data and related guidance.

[sbc1-code]

A durable project record is now available in Google Drive: https://docs.google.com/document/d/1hAPLQy1yK0HFqqvtT_ayyqEvDfui1-TQcRbSRVIl-i8/edit

It records the shipped foundation, the Pages deployment incident and resolution, the current UX/findability pass, working links, and the next implementation queue. Future updates should append to this record and the issue rather than relying on chat history alone.

[sbc1-code]

Boarded in Engineering: Status = Todo, Pillar = Product. This is now the canonical queue item for the next UX/findability pass.

[sbc1-code]

Evidence-backed execution queue — 2026-08-06

The UX/SEO/analytics audit is complete enough to sequence implementation. Work in this order:

P0 — restore trust before growth

1. [P0][ops] Restore automatic deployment after scheduled data refreshes #57 — restore automatic deploys after scheduled data refreshes.
2. [P0][bug] Fix the /best-time/ hub crash and add canonical-route smoke tests #61 — fix the live /best-time/ crash and add canonical-route smoke tests.

P1 — make answers accurate and discoverable

3. [P1][data] Never recommend crossing times outside official operating hours #65 — exclude closed hours from every best-time recommendation.
4. [P1][trust] Make freshness, missing-data, and summary labels accurate #58 — make freshness, missing-data, and summary labels honest.
5. [P1][SEO] Reconcile sitemap URLs, aliases, and Search Console indexing #59 — reconcile the sitemap, aliases, and Search Console indexing.
6. [P1][SEO] Improve organic CTR on proven page-one crossing queries #64 — improve CTR on proven page-one queries/pages.
7. [P1][analytics] Instrument the core border-crossing decision funnel in Umami #60 — instrument the core decision funnel in Umami.

P2 — improve the product system

8. [P2][UX] Simplify the first-visit and mobile crossing-selection journey #62 — simplify the first-visit/mobile crossing-selection journey.
9. [P2][maintenance] Upgrade vulnerable dependencies and add a CI security gate #63 — upgrade vulnerable dependencies and add a CI security gate.

Baseline saved in the child issues

• Search Console, last 28 days: 1,836 clicks / 102,223 impressions / 1.8% CTR / position 8.4.
• Indexing: 118 indexed, with 43 discovered/not indexed and 17 legacy 404 examples awaiting evidence-based reconciliation.
• Experience: 71 good mobile Core Web Vitals URLs, zero poor/needs-improvement; HTTPS and breadcrumb/Dataset enhancement reports are clean.
• Data QA: 172 published claims matched the aggregate data.
• Product analytics: Umami pageviews exist, but core decision actions are not instrumented.
• Dependency audit: 7 vulnerable packages (3 high, 3 moderate, 1 low).
• Expansion gate: do not mass-produce more best-time/day pages yet; repair reliability and lift the existing cohort first.

All nine child issues are on the Engineering board as Todo, with Ops / Engineering / Product / Content pillars set by work type.

[sbc1-code]

P0 implementation update — 2026-08-06

• [P0][bug] Fix the /best-time/ hub crash and add canonical-route smoke tests #61 closed: the live /best-time/ hub is repaired; metadata/schema now reflect 44 crossings; local, CI, and live-domain browser sweeps each passed 390 navigations.
• [P0][ops] Restore automatic deployment after scheduled data refreshes #57 functionally repaired: a real bot refresh created 83814c4, exactly one Pages workflow deployed that same SHA, and production now serves its exact fetched_at timestamp. Keep [P0][ops] Restore automatic deployment after scheduled data refreshes #57 open only for the original two naturally scheduled-cycle verification gate.

Once those scheduled receipts arrive, the next implementation item remains #65: exclude closed operating hours from Best-time recommendations.

[sbc1-code]

Commercial-product state reconciliation shipped in a6f1a7c0 on agent/operator-control-plane. The current product remains free while reliability, analytics readback, and corridor workflow discovery establish whether a paid job exists. Generic widgets, public CBP data, and basic consumer alerts are not an active product plan; a manual business pilot now requires three independent repetitions of the same buyer-owned workflow plus one written paid commitment. Updated the operator guide, roadmap, durable decisions, product copy, and an explicit truth-regression test. Local build/prerender, non-browser verification, and the 400-navigation Chrome smoke pass completed; remote release verification is pending. No production deployment, provider, billing, or customer change occurred.