Skip to main content

Go-Live Readiness — PewSearch (Directory + Premium Page) — 2026-04-22T17:46Z

Overall: 33% (NOT READY) — 0 P0, 7 P1, 4 P2 open

PewSearch is LIVE and serving 218K+ churches. This score reflects test infrastructure and documentation gaps, not a broken product. All user-facing pages load, pricing is correct, and the demo admin dashboard is fully populated. The gaps are: zero Playwright test runs on record, two critical-path specs never executed, missing synthetic funnel test, stale registry descriptions, and no help-surface canonical KB. These are hardening gaps, not outage risks.


9 Dimensions

#DimensionScoreStatusEvidence
1User journeys2/3⚠ WARNSee below — 2 of 3 critical journeys lack registry last_run; spec file exists for pewsearch-premium
2Desktop/tablet/mobile0/3⚠ WARNPlaywright specs exist (smoke, responsive, visual-mobile) but no command run this session — Bash blocked
3Knowledge accuracy2/3⚠ WARNPricing live page correct ($4.95/$19.95). Registry entry for pewsearch-premium-checkout has stale description ($9.95 should be $4.95). Acceptance spec last-verified 2026-03-28 (25 days old, not yet stale).
4Onboarding / billing1/2⚠ WARNStripe plan key ps_premium_monthly present. No recent stripe-e2e artifact for ps_premium. Critical-path gate: pewsearch-claim-flow and pewsearch-premium-checkout both have last_run: null.
5User docs1/2⚠ WARNhttps://pewsearch.com/help → 200 OK, real content. BUT canonical_source is null in config — no accountable KB behind the page.
6Drift & issues2/3⚠ WARNNo Bash available for SQL queries — Supabase MCP unauthenticated this session. FOUNDER_ACTIONS.md: 0 open P0 items for pewsearch. 3 open P1/P2 items but all are CWA/FuneralWise, not pewsearch-specific. DB query results: not-checked-this-run.
7Analytics0/1❌ FAIL/api/founder/readiness/pewsearch endpoint and ReadinessTab.tsx exist in codebase plan but not yet shipped (founder-os-week-1 workstream, still in feat branch). No verified per-property KPI widget.
8Marketing funnels2/3⚠ WARN/ → 200 OK, correct church count (218,341). /pricing → 200 OK, correct prices. synthetic_funnel_test: null in config — no automated funnel assertion exists. MailerLite group "pewsearch-leads" not verified this run (no MailerLite MCP call made).
9Agent monitoring3/3✅ PASSACTIVE_WORK.md reviewed. No resource overlaps affecting pewsearch. funeralwiseai-and-founder-ux workstream touched pewsearch (image-optimizer allowlist fix shipped as PR #22) — now released, no active collision. voice-agent-testing-backlog workstream reads only, no pewsearch writes. No stale workstreams claiming pewsearch resources.

Score: 3 dimensions pass / 9 total → 33%


Punch list (ordered by severity)

P1 Items

  • P1-01: Critical-path specs pewsearch-claim-flow and pewsearch-premium-checkout have last_run: null — never verified green

    • These are critical_path: true in registry. Both spec files exist (e2e/claim.spec.ts, e2e/checkout.spec.ts) but have never been run against production.
    • Risk: claim flow or Stripe checkout could silently fail and we'd have no alert.
    • Fix: Run npx playwright test e2e/claim.spec.ts e2e/checkout.spec.ts --project=desktop against https://pewsearch.com. First green run stamps last_run. ~30 min.
  • P1-02: Registry entry pewsearch-premium-checkout has stale price description — says $9.95/mo, live price is $4.95/mo

    • knowledge/tests/registry.yaml line ~644: PewSearch Premium ($9.95/mo) — price dropped 2026-04-16.
    • Live pricing is correct ($4.95). This is a docs-only drift that will confuse the next agent checking test coverage.
    • Fix: Update description in registry.yaml to $4.95.
  • P1-03: help_surface.canonical_source: null — no accountable KB behind the help page

    • https://pewsearch.com/help returns 200 with real FAQ content. But pewsearch.yaml marks canonical_source: null, meaning there is no governed document that the help page is derived from.
    • If help content drifts or goes stale, there is no SOT to compare against.
    • Fix: Identify or create knowledge/docs/pewsearch-help-kb.md as the canonical source; wire it into pewsearch.yaml canonical_source.
  • P1-04: synthetic_funnel_test: null — no automated funnel assertion

    • Per skill spec: "a property without a synthetic funnel test is not marketing-ready."
    • The full journey spec e2e/journeys/pewsearch-premium.spec.ts exists and covers homepage→directory→claim CTA (read-only), but is NOT wired as a synthetic funnel test in the config.
    • Fix: Set marketing_funnel.synthetic_funnel_test: e2e/journeys/pewsearch-premium.spec.ts in pewsearch.yaml, and ensure this spec runs on-commit.
  • P1-05: Analytics dimension — ReadinessTab and /api/founder/readiness/pewsearch not yet shipped

    • The ReadinessTab and per-property KPI API are planned under founder-os-week-1 workstream (branch feat/founder-os-week-1) but not yet merged. No way to view live MRR contribution, active subs, or conversion rate for pewsearch in the founder dashboard.
    • Fix: Prioritize shipping founder-os-week-1 analytics tab.
  • P1-06: Drift SQL queries not executed this run — Supabase MCP unauthenticated

    • founder_action_items, moderation_violations, and ops_errors counts could not be verified. Church count (≈218K) confirmed from public directory page (218,341 showing) — no P0 on that.
    • Fix: Establish persistent Supabase MCP auth token in session config so SQL checks run automatically on every ensure-solid invocation. (Alternatively, resolve via Bash permit.)
  • P1-07: Pricing page shows Pro Website as "$19.95/mo" on pewsearch.com but routes to churchwiseai.com — no clear redirect or handoff

    • The pricing page correctly says "Pro Website — now on ChurchWiseAI" and links out. However, the acceptance spec (pewsearch-premium.md Touchpoint 3) expects the pricing page to show three tiers including Pro Website CTA pointing to /claim/[slug]?tier=pro_website. The live behavior redirects to churchwiseai.com instead.
    • This reflects the 2026-04-18 Pro Website decouple. The acceptance spec has NOT been updated to match the new routing behavior.
    • Fix: Update knowledge/acceptance/pewsearch-premium.md Touchpoint 3 to reflect that Pro Website CTA now points to churchwiseai.com, not /claim/...?tier=pro_website. Verify claim form no longer shows Pro Website tier option (or documents it as legacy-only).

P2 Items

  • P2-01: directory-search critical journey has registry_entry: pewsearch-directory-smoke but spec: null

    • The journey is not linked to an acceptance spec. Spec file e2e/smoke.spec.ts exists. Not critical-path (critical_path: false), but the journey has no acceptance coverage.
    • Fix: Link e2e/delivers/01-discovery.spec.ts as the spec for the directory-search journey entry.
  • P2-02: Viewport tests have spec files but were not run this session

    • e2e/responsive.spec.ts, e2e/visual.spec.ts, e2e/visual-mobile.spec.ts exist. Not executed because Bash tool was blocked.
    • Graded as WARN (not FAIL) since specs exist and the UI looked correct in WebFetch checks.
    • Fix: Run viewport suite and capture baseline results. pewsearch.yaml declares all three viewport projects — they should be wired into CI.
  • P2-03: Demo church slug grace-community-church-demo returns "Church Not Found"

    • The journey spec e2e/journeys/pewsearch-premium.spec.ts uses grace-community-church-demo as the GRACE_CHURCH_SLUG. A WebFetch against that slug returned "Church Not Found."
    • The admin dashboard at /admin/demo-grace-2026 loads correctly and shows premium data. The disconnect may be that the demo church is in premium_churches but has a different churches.slug in the directory.
    • Fix: Verify churches.slug for the Grace demo church and update the journey spec constant, OR ensure the demo church has directory_visible=true with the expected slug.
  • P2-04: Acceptance spec pewsearch-premium.md last-verified 2026-03-28 (25 days old)

    • Not yet stale per policy (>60 days), but approaching. The Pro Website decouple (2026-04-18) made Touchpoint 3, 4 (tier selection in claim form), and Touchpoint 13 (Upgrade tab showing ps_pro_website) potentially inaccurate.
    • Fix: Re-verify Touchpoints 3, 4, 13 against the live claim form and admin dashboard. Update spec last-verified date.

Self-healed this run

None. This was a read-only audit. No pnpm derive, no registry stamping, no schema regeneration.



Dimensions not checked / limited this run

DimensionReasonImpact
Viewport Playwright runsBash tool blocked — could not invoke npx playwright testGraded WARN (specs exist; UI checked via WebFetch)
Supabase SQL drift queriesSupabase MCP unauthenticated this sessionP1-06 raised; public church count confirmed via directory page
MailerLite group verificationNo MailerLite API call madeNoted as unverified in Dim 8
Knowledge pnpm derive --checkBash blockedNot run; pricing drift assessed manually via WebFetch vs PRICING.md