[P1][SEO] Improve organic CTR on proven page-one crossing queries #64

Current next action - 2026-10-22

• Status: Published and functionally verified. PR #146 merged/source ef9d5447d74e1d9a19eefa4ce946ff62ea4e1116, reviewed head 12c3767109bb458e1ef1e8d6161528db143b2d71. Automatic Vercel run 39870435982 SUCCESS. Deployment dpl_8epUqPVCJrxrjotZNjjhMe9sbRdF READY at exact deployment, canonical borderpulse.com, apex/www aliases verified. Public manifest source matched before/after native served QA beginning 2026-10-22T23:18:45Z; both EN and ES generated guide pages return HTTP200, self-canonical, one sitemap entry each. Four-file correction only; original dates, author, slugs and translation pair remain.
• Current GSC evidence supplied by the coordinating session, recorded October 22: September 23–October 20, 2026 versus August 26–September 22, 2026: 636/814 clicks; 50,019/69,167 impressions; displayed CTR 1.3%/1.2%; average position 9.1/9.5; mobile 580/636 current clicks. These show fewer clicks/impressions, not a global CTR decline. The exact property identifier, original extraction timestamp, raw export and query-to-page join were not included in this handoff; do not invent them or treat this recording time as the extraction time. No current reliable on-site funnel/drop-off read was obtained.
• Three opportunities diagnosed:
  • SENTRI renewal: query 508 impressions, 1 click, average position 8.5. Renewal intent calls for the official application route and clear next steps. Existing mobile EN/ES pages buried the portal link below the first viewport and incorrectly made conditional approval the basis of continued benefits, including telling pending renewals after expiration that access was suspended. The official DHS TTP FAQ, rendered publicly and checked October 22, instead ties continued benefits to submitting before expiration; it also covers possible interview waiver and card activation. The query-to-guide landing-page join remains unverified.
  • Santa Teresa port of entry: 457 impressions, 0 clicks, position 11.6. The existing crossing page already shows northbound reports, report clocks, hours and a walk-or-drive link; local EN/ES mobile navigation is checked. General port intent may include location/hours, but the query alone does not establish that intent or explain non-clicks. City of El Paso names Santa Teresa and pedestrian/passenger capability but lists midnight closing, while the current CBP-derived page shows 6 am–10 pm. Reconcile source scope/current hours before adding a fixed hours answer. No change to this page or its report values.
  • Donna page: 30 versus 50 clicks, displayed CTR 1.4%/1.8%; query, position and device mix for that page are not supplied. The existing page shows reports/hours and links to nearby Progreso comparison; EN/ES mobile link/Back works. The City of Donna identifies Donna–Rio Bravo, while CBP's grouping produces the existing “Progreso - Donna” label. Possible naming/intention friction is a hypothesis, not a proven cause. Preserve its canonical and distinguish it from Progreso International Bridge. A toll listing alone does not establish current northbound pedestrian service.
• Selected change: Correct the existing SENTRI guide pair around verified renewal steps; put the official TTP action on the first screen; distinguish timely pending renewal, late renewal and conditional approval; remove unsupported fingerprint/reschedule/mailing/processing promises. Preserve useful current-waits navigation, translation, sharing and Back. No new routes, dependency changes, ad changes or broad metadata rewrite.
• Real checks: Latest-main combined tree (base ff546d878db16f49f1ecee72c5ec126719921c0b, retaining CBP/FX/anomaly updates) passed fresh build and full npm test: 133 unit tests, 402 manifest browser navigations. Actual merge tree 781b8b8300bc91ddfb40ca67ea434aad8655d31d exactly equals that tested integration. Both earlier exact-head push CI and PR CI SUCCESS. Production workflow repeated build/release checks successfully with 133 unit tests, 402 navigations and the SENTRI fixtures. Actual served EN/ES native QA PASS at 1440×900, 390×844 and 320×568: official action on first screen, keyboard activation/new tab/null opener, pending/late cautions, document language, canonical/publication/revision metadata, current-waits/Back and translation, copy/WhatsApp target construction. Six production captures, zero captured application errors; served entry and blog-runtime hashes retained, blog runtime byte-for-byte matches the tested integration. Portal/clipboard actions use controlled fixtures; no real TTP login/application/payment/eligibility or message delivery is claimed. FAQ source re-read covered the four exact official answers; individual status/lane access is deferred to TTP/CBP.
• Budgets/risk: Production entry 61.88 KiB/unchanged 62, CSS 19.11/20; narrow margin remains. Two pre-existing moderate npm package entries represent one advisory, GHSA-rj75-hqrm-r3gf: devDependency @tailwindcss/typography@0.5.19 → postcss-selector-parser@6.0.10, fixed parser 7.1.6. Package/lock blobs are unchanged before/after PR146. Inspected use is the trusted-source CSS build plugin, not a discovered browser/request-path parser; npm audit --omit=dev reports zero entries. Full audit still reports 2 moderate, 0 high/critical. No unrelated dependency or audit-exception change and no security certification. All 16 pre-existing checkout/worktree HEAD/status/modified-file hashes remain preserved, including original 18 paths and reviewed branch. GH007 was resolved with a verified public noreply identity on the unpublished commit only, no privacy/account setting changes.
• Next review/publication gates: Release completed through one successful merge and its single automatic canonical Vercel publication, without manual dispatch or duplicate deploy. Manifest prepared 23:14:42.040Z, CBP/history latest 23:07:32.031Z, 42 crossings/2,539 history snapshots, FX 23:07:32.447Z; these are retrieval/history times, not proof of each lane report's freshness. Broader issue64 stays OPEN for measurement. Bounded rollback is the inverse of the four-file merge relative to its first parent (or revert of the reviewed change), preserving all later collector/dependency/history work; checked and not applied. Own headless browsers closed; shared Chrome unused; Current Focus write not retried. Legacy automatic Pages run 39870436086 remains pending with no jobs started; canonical Vercel publication and served verification succeeded independently, and no hosting/access setting was changed.
• Measurement test: Hypothesis: an accurate, direct renewal answer better serves renewal intent. After any approved release, compare canonical renewal-guide query-cohort CTR over 28 days against the preceding 28, with average position, query mix, device mix and impressions as guardrails. First obtain the exact property and query-to-page export; until then, the CTR test is unmeasured. Separately distinguish guide → official TTP task completion from guide → wait reports. A successful outbound renewal action need not increase on-site wait clicks; current outbound completion/drop-off is not established or newly instrumented. No causal or conversion-lift claim. Keep source correctness as a guardrail; later presentation tests must not restore unsupported eligibility claims.
• Historical context only: Previously reported September totals were 827 clicks/69.2k impressions/about 1.2% CTR versus August 1.07k/93.2k. Those historical rounded totals did not demonstrate a CTR decline. The original July–August baseline and issue scope below remain unchanged. GSC Wizard's expired paid trial is not retried or upgraded; use an existing authorized measurement path. #59 continues to own inventory/indexing reconciliation.

28-day baseline (Search Console, Jul 8–Aug 4, 2026)

• 1,836 clicks
• 102,223 impressions
• 1.8% CTR
• Average position 8.4

High-leverage landing pages already near page one include:

• /crossing/nogales-deconcini/: 231 clicks / 8,725 impressions / 2.6% CTR / position 7.5
• /crossing/hidalgo-pharr-anzalduas-international-bridge/: 222 / 5,803 / 3.8% / 6.8
• /crossing/san-ysidro/: 76 / 8,122 / 0.9% / 10.4
• Homepage: 105 / 8,590 / 1.2% / 9.7
• Medication guide: 143 / 9,483 / 1.5% / 6.9

Best-time pages have only early impressions/clicks, so bulk page expansion is premature.

Scope

• Build a query-to-page map for the top opportunity pages.
• Rewrite title, description, on-page answer, and internal-link anchors to match real query language while preserving accuracy.
• Prioritize Nogales, Anzalduas/Donna, San Ysidro, homepage, and the high-impression medication/SENTRI guides.
• Resolve historical trailing-slash split signals and measure the canonical versions only.
• Compare snippets against current competitors; avoid unverifiable superlatives or “live” claims when data is stale.
• Do not generate hundreds of new day/route pages until the existing cohort earns meaningful impressions.

Definition of done

• Pre-change query/page baseline is saved.
• Each target page has one clear intent, differentiated snippet, useful first-screen answer, and relevant next link.
• Medical/legal travel copy remains source-verified.
• Recheck CTR and position after 28 days; keep/revert based on evidence.

Parent roadmap: #56