WREA FY27 Execution Board

Contribution after ad cost (12wk ann.)

$1.65M

base $1.31M → Dec $1.55M → Jun $2.1M plan / $1.8M base

Owned-channel intro share

14%

base 14% → Dec 20% → Jun 25–30%

Lead → intro conversion

29%

base 28% → Dec 30% → Jun ≥33%

Blended cost / intro

$413

base $398 → Dec $370 → Jun ≤$330

Revenue / intro

$884

base $732 → Dec $750 → Jun ≥$790

Emailable consented contacts

14.1k

base 14.0k → Dec 15.5k → Jun ≥18k

Scorecard current values as at 2026-08-16.

Visual overview — FY27 timeline & status mix (tap to hide)

DDone: 3Next: 2Backlog: 38

L4Done: 1In progress: 1Next: 1Backlog: 36

L3Done: 1In progress: 3Backlog: 26

L1Next: 1Backlog: 78

L2In progress: 3Next: 1Backlog: 48

L6In progress: 1Next: 2Backlog: 47

L5Backlog: 55

OPSDone: 1In progress: 1Next: 2Backlog: 15

DoneIn progressNextBacklogBlockedOverdue (ring on timeline)

Monthly lever review — August review #1 — draft (not yet held, 2 Mondays missed) (reviewed 2026-08-03; next: first Monday of the month)

Prepared 3 Aug but still not held. Self-serve foundations are deployed with flags off, but the customer cache disagrees with the strict resolver and the ongoing verification cycle is not operating. Profile-less agent duplication also remains open. D2/D3, the monthly review and Rescue copy still need decisions. Paid is in cash-protection mode on weaker Google economics.

! D CEO Decisions Amber D2/D3 (due 31 Jul) remain open and gate consumer estimate exposure. Decision: Thomas records both yes/no decisions today or sets an explicit new date. ! L4 Paid value bidding — bid on gross profit Amber There were no selected-week Google changes, but all 24 changes in the 28-day lookback still need logged rationale. Current Google economics deteriorated to 36 intros, $601 CPI and 32.5% campaign-model margin; three weak-ad-group actions are drafted and approval-gated. Decision: David logs rationale for the 24 lookback changes and answers the three drafted actions. No scale-up while the verdict is cash-protection. ! L3 Email nurture — monetise the database Amber P0/P1 code is available but production exit proof is outstanding; L3-2 is overdue and no customer-facing nurture campaign has launched. Decision: Thomas approves or edits the two Rescue drafts and sender identity this week. ! L1 Estimate tool — own the valuation moment Amber L1-1 is overdue with no shipped evidence; the estimate lane is gated by D2/D3; the 1 Sep hard date needs P0 to start on schedule. Decision: Ship L1-1 this week or explicitly descope it at this review. ! L2 Agent-evidence data — the moat Amber Self-serve foundations are deployed flags-off, but the 14 Aug audit found 5,272 active people marked verified in the customer cache that the strict resolver rejects. The ongoing verification cycle is not operating, and profile-less agent duplication remains open. Decision: Reconcile the cache, prove the ongoing worker and finish the read-only source canary; otherwise explicitly move the 1 Sep controlled-launch date. ! L6 Convert what we already have — sales + CRO Amber L6-1 speed-to-lead reporting is still not on the Monday pack (due 31 Jul; queries proven 22 Jul). Decision: Land the pack block this week; it also unblocks L6-2. ✓ L5 Yield per seller — agent routing, partners, investor wave Green Nothing due; D6 investor decision is due 31 Aug. ! OPS Operating rhythm Amber The 3 Aug review is still not held after two Mondays. Weekly automation, ledger reconciliation and incident response operated, but lead-worker health detection remains incomplete. Decision: Hold OPS-2 this week and record one decision per amber lever.

D✓! L4!! L3✓! L1✓! L2✓! L6!! L5✓✓ OPS✓!

This week (2026-08-17)

Overdue (13)

D2 2026-07-31 · 17d overdue Thomas DR1 — verify sold-listings data is public domain (gates tool comparables) → OVERDUE seventeen days (was 31 Jul) — third consecutive board cycle without closure. Thomas records the written yes/no at this review, or sets an explicit new date. No consumer estimate release until D2 and D3 are closed.

D3 2026-07-31 · 17d overdue Thomas DR2 — confirm ODIN paid tier permits consumer-facing display → OVERDUE seventeen days (was 31 Jul) — third consecutive board cycle without closure. Thomas records the written yes/no at this review, or sets an explicit new date. Keep the 109-test estimate lane internal until this and D2 are closed.

L4-2 2026-07-17 · 31d overdue David Change-log discipline (logged rationale per account change) → Overdue since 17 Jul. The selected week had no Google changes; the 28-day lookback shows 24, all by David and mostly 6 Aug budget updates plus the Agent National goal change to Intro. The rationale log for those 24 is still outstanding, so the two-consecutive-fully-logged-weeks exit is not met.

L4-4 2026-08-14 · 3d overdue Felix (build) + David (validate) Offline GP-value conversion upload (build + validate) → Scope with Claude: event schema + upload path; then Felix builds.

L4-6 2026-07-31 · 17d overdue David Microsoft platform-vs-CRM reconciliation (recurring P1) → Overdue (was 31 Jul). Microsoft produced $3.5k spend, 1,314 clicks and 34 platform lead conversions versus CRM 38 leads, 8 intros, $432 CPI and $416 GP/intro — the healthier paid channel this week. Complete the documented platform-conversions-to-CRM mapping. Budget decisions remain CRM-only.

L3-2 2026-08-01 · 16d overdue Felix Nurture P0: eligibility service + suppression + webhook write-back + campaign columns → Overdue (was 1 Aug). #4464 is merged and on production_new; migrations and passive endpoints were runtime-checked 21 Jul. Exit remains unmet until production proof covers reason codes, end-to-end unsubscribe suppression, and mass-email dry-run exclusions.

L1-1 2026-07-18 · 30d overdue Felix Dead-traffic conversion: tool CTAs on house-prices-graph + calculators pages → Overdue since 18 Jul — fourth consecutive board cycle with no shipped evidence. This week's pack shows the target house-prices page took 941 organic users to zero lead users over 28 days. At OPS-2: ship the conversion path or explicitly descope it into L1-5.

L1-2 2026-08-14 · 3d overdue Felix Tool P0: endpoints/tokens/Redis/budget + CI leak gates + legacy route throttle → Local production-review candidate exists (109 tests, no P0/P1 finding), internal only. Product-repo P0 still starts wk5; consumer exposure remains gated by D2/D3 and production integration.

L1-8 2026-08-14 · 3d overdue Felix ODIN/OTH fallback guardrails in the tool (data-confidence gate) → [§6.7 FOUNDATION, gated on tool P0] Build with tool P0; fallback order already proven in PropertyValuationService.

L6-1 2026-07-31 · 17d overdue Felix or Codex Speed-to-lead metrics in weekly pack (first-contact time, attempt count) → Overdue (was 31 Jul; queries proven 22 Jul). Third consecutive board cycle without the Monday-pack block. Land the four SLA numbers plus median time to first connected call this week; L6-2 waits on it. If it cannot land, name the blocker at OPS-2.

L6-2 2026-08-14 · 3d overdue Thomas + sales Sales SLA live + weekly compliance reporting → Unblocked — D5 closed 2026-07-27, so the dependency is now L6-1, not D5. D5 resolved the coverage question by decision (no Sunday shift, off-peak paid pulled back to about midnight, self-serve is the out-of-hours answer) and the forensics cleared the floor on first-attempt speed, so this item is no longer about policing response time. What remains is reporting the connected-call metric once L6-1 lands it on the Monday pack (due Fri 31 Jul). Do not frame this as a sales performance gap.

L6-3 2026-08-07 · 10d overdue Felix Form/JS fixes: btoa Latin1 error, jQuery error, form QA → Overdue (was 7 Aug). Clarity access is restored and the weekly pack carries JavaScript errors again, so the two-week zero-error exit is measurable. Ship the btoa/jQuery fixes with normal smoke proof and start the two-week clock.

OPS-2 2026-08-03 · 14d overdue Thomas + Claude Monthly lever review #1 (green/amber/red, one decision per amber/red) → Was due 3 Aug; two Mondays have now passed. Hold the 60-minute review this week from the refreshed August draft; Thomas confirms or amends each rating, logs one decision per amber, then marks this done.

Due in the next 14 days (6)

D6 2026-08-31 · 14d Thomas Green-light January investor-exit campaign → Decide after Q1 gate shape is clear.

L3-3 2026-08-21 · 4d Felix Nurture P1: enrolment tables + audience preview + campaign UI + journey card → Merged in #4464, but there are still zero campaigns, enrolments, or launches. Run controlled production verification: preview the 137×239 overlap and prove preview/enrol share one code path, then close.

L1-3 2026-08-28 · 11d Felix Tool P1: Vite island + Places + lead flow + GTM events (controlled beta) → Copy fixes (phone, hero, stars) land here per spec §12 table.

L2-1 2026-08-21 · 4d Felix Agent-evidence template v2 wired to existing rankings API → Spec v2 was drafted 22 Jul: rankings-API contract, confidence-degraded mode, and 10-suburb pilot criteria. Thomas confirms; Felix renders one pilot suburb. No at-scale generation before L2-6.

L2-6 2026-08-28 · 11d Felix + Codex REA listing-status reconciliation live (gates public agent pages) → Scraper recovery closed 4 Aug: PRs #431/#432 are merged and deployed, both sidecars are healthy at 4 serving + 6 reserves, closed-circuit and idle, and routine queue drain belongs to the normal 20:05 schedule. The exit still stands: top-100-suburb business-result reconciliation from queues through staged applies to actual listing changes is not proven.

OPS-5 2026-08-28 · 11d Felix Lead-worker and queue health detection (13 Aug incident prevention) → The 32 queued incident jobs are recovered and the PHP 8.2 worker plus scheduler are restart-persistent. Pipeline and hourly HTTP checks still inspect neither worker presence nor queue age. Add those checks, raise the application PHP floor, and clear or evidence-close the one remaining failed job.

In progress (9)

L3-4 2026-09-01 Felix + Thomas (copy) fast-rescue-v1: internal → 15-lead lane → monthly flow with 20% holdout → Two Rescue drafts are ready with verified CRM merge tags. Thomas approves/edits copy, chooses sender identity (per-staff recommended), confirms the internal→15-lead pilot gate, and records the consent basis. Activation remains a separate approval.

L2-8 2026-09-01 Felix + Thomas Self-serve agent connection — controlled production launch → CRM #4511 and portal #1708/#1709 are deployed with every new flag off; scraper #440/#441/#442 is live. New launch blockers from the 14 Aug audit: the strict resolver verifies 17,549 of 38,178 sales people; the customer-facing cache marks 5,272 active people verified that the resolver rejects; and the ongoing Domain/RateMyAgent cycle is not operating (one stale canary, empty CRM handoff, no worker log). Sidecar pool health was unproved in that audit, not failed. Reconcile the cache and prove the worker, then rerun the Domain canary and reviewed national baseline dry-run. Import remains approval-required. At OPS-2, resource the gates to hold 1 Sep or move the date explicitly. Deployed flags-off is not launched.

L6-6 2026-09-30 Felix + Thomas CRM agent search / recommended-agent quality audit → Prevention is incomplete for profile-less listers: live code only deduplicates within one parse, leaving 5,416 excess rows across 439 identities from 5–13 Aug and 1,014 listings with extra links. PR #452 / DT-935 is blocked as unsafe because global name-only reuse collides on real people; the required fix is listing-scoped non-empty name-and-photo reuse with fail-closed person matching. Historical repair is DT-937, staged and approval-gated, and blocked by DT-935. The floor trust check still stands.

Levers

D CEO Decisions 3/8 done · 2 overdue Thomas

Target: Every decision closed before it blocks a build · Strategy §11

D2 Next 2026-07-31 · 17d overdue Thomas Q1

DR1 — verify sold-listings data is public domain (gates tool comparables)

Done when: Verification recorded in the estimate-tool spec §2; comparables flag decision made

Next step: OVERDUE seventeen days (was 31 Jul) — third consecutive board cycle without closure. Thomas records the written yes/no at this review, or sets an explicit new date. No consumer estimate release until D2 and D3 are closed.

D3 Next 2026-07-31 · 17d overdue Thomas Q1

DR2 — confirm ODIN paid tier permits consumer-facing display

Done when: Tier check recorded in the estimate-tool spec §2

Next step: OVERDUE seventeen days (was 31 Jul) — third consecutive board cycle without closure. Thomas records the written yes/no at this review, or sets an explicit new date. Keep the 109-test estimate lane internal until this and D2 are closed.

D6 Backlog 2026-08-31 · 14d Thomas Q1

Green-light January investor-exit campaign

Done when: Go decision so Q2 schedules content + CRM investor-segment work

Next step: Decide after Q1 gate shape is clear.

D7 Backlog 2026-11-30 Thomas Q2

SMS rescue step decision (fast-rescue-v2)

Done when: Yes/no after email-rescue readout; consent posture agreed

Next step: Wait for rescue pilot lift-vs-holdout readout (L3-4).

D8 Backlog 2026-12-19 Thomas Q2

Partner pilot appetite (conveyancer/broker, Q3)

Done when: Yes/no to 2–3 partner conversations in Q3

Next step: Decide at Q2 gate.

D1 Done 2026-07-10 Thomas Q1

Ratify FY27 targets + Q1 descope order

Done when: Targets adopted (or adjusted); descope order agreed as written in strategy §7

Next step: Ratified by Thomas 2026-07-06 — targets adopted, Q1 descope order agreed.

D4 Done 2026-07-10 Thomas Q1

Ratify cash rule ($1.55M Dec trigger) + contractor pre-approval trigger

Done when: Both triggers agreed so December executes automatically

Next step: Ratified by Thomas 2026-07-06 — cash rule + contractor pre-approval triggers agreed.

D5 Done 2026-07-24 Thomas + sales Q1

Agree sales speed-to-lead SLA with the floor

Done when: SLA documented (first-call time, attempt cadence); compliance reportable

Next step: Closed by Thomas 2026-07-27. The out-of-hours coverage question is decided, not open: no Sunday shift (Kirsty attracts double time and weekend lead volume does not justify it), out-of-hours paid spend already pulled back with off-peak Google campaigns running only to about midnight, and self-serve agent connection is the intended out-of-hours answer — it goes live after the agent list, source filtering (no agent, agency or data unique to REA revealed) and scraper are right. Forensics cleared the floor: first-attempt speed is strong and the never-called and weekend-delay claims were retracted. Remaining work is measurement only, which is L6-1 (time to first connected call). Do not reopen this as a sales performance issue.

L4 Paid value bidding — bid on gross profit 1/6 done · 3 overdue David (ops) + Felix (build)

Target: Target CPA and CRM economics reviewed weekly; value upload live end-Q1; tROAS mid-Q2; no negative-GP Google week in H2; CPI ≤$330 by Q4 · Strategy §6-L4 · Paid governance

L4-2 In progress 2026-07-17 · 31d overdue David Q1

Change-log discipline (logged rationale per account change)

Done when: Every change in the weekly review carries a logged why; zero unlogged changes for 2 consecutive weeks

Next step: Overdue since 17 Jul. The selected week had no Google changes; the 28-day lookback shows 24, all by David and mostly 6 Aug budget updates plus the Agent National goal change to Intro. The rationale log for those 24 is still outstanding, so the two-consecutive-fully-logged-weeks exit is not met.

L4-6 Next 2026-07-31 · 17d overdue David Q1

Microsoft platform-vs-CRM reconciliation (recurring P1)

Done when: Documented mapping of platform conversions → CRM intros; MS budget decisions use CRM only

Next step: Overdue (was 31 Jul). Microsoft produced $3.5k spend, 1,314 clicks and 34 platform lead conversions versus CRM 38 leads, 8 intros, $432 CPI and $416 GP/intro — the healthier paid channel this week. Complete the documented platform-conversions-to-CRM mapping. Budget decisions remain CRM-only.

L4-4 Backlog 2026-08-14 · 3d overdue Felix (build) + David (validate) Q1

Offline GP-value conversion upload (build + validate)

Done when: CRM intro events with GP-proxy values flowing to Google; values reconcile with CRM weekly

Next step: Scope with Claude: event schema + upload path; then Felix builds.

L4-8 Backlog 2026-10-31 David Q2

Geo re-weighting toward WA/QLD/SA + exact-match winners

Done when: Marginal budget shifted; CRM economics confirm by state

Next step: After tROAS stable.

L4-7 Backlog 2026-11-15 David Q2

tROAS migration on Agent (National) once signal floor met

Done when: tROAS live; no negative-GP week in first 4 weeks post-migration

Next step: Blocked by L4-4 signal volume; check floor at Q1 gate.

L4-5 Done 2026-07-13 David + Claude Q1

Leak monitoring: flag sustained negative-GP ad groups (trend, not one week)

Done when: Standing Monday-pack block names ad groups with sustained (multi-week) negative GP; action only on a sustained trend, not one week's noise

Next step: Closed 2026-07-27: read-only change history confirms all 4 approved changes landed 22 Jul (Phrase Regions, Phrase QLD, Broad VIC paused; Exact VIC tCPA $200→$180). Sustained-negative monitoring now runs through the weekly dry-run lane.

L3 Email nurture — monetise the database 1/6 done · 1 overdue Felix

Target: Rescue live mid-Q1; ≥150 incremental intros (vs holdout) by Jun; zero suppressed sends · Nurture build spec · Strategy §6-L3

L3-2 In progress 2026-08-01 · 16d overdue Felix Q1

Nurture P0: eligibility service + suppression + webhook write-back + campaign columns

Done when: Spec §10 P0 exit criteria: reason-code tests green; test unsubscribe suppresses end-to-end; mass-email dry run shows exclusions

Next step: Overdue (was 1 Aug). #4464 is merged and on production_new; migrations and passive endpoints were runtime-checked 21 Jul. Exit remains unmet until production proof covers reason codes, end-to-end unsubscribe suppression, and mass-email dry-run exclusions.

L3-3 In progress 2026-08-21 · 4d Felix Q1

Nurture P1: enrolment tables + audience preview + campaign UI + journey card

Done when: Preview shows the 137×239 rescue overlap answer; preview and enrol share one code path

Next step: Merged in #4464, but there are still zero campaigns, enrolments, or launches. Run controlled production verification: preview the 137×239 overlap and prove preview/enrol share one code path, then close.

L3-4 In progress 2026-09-01 Felix + Thomas (copy) Q1

fast-rescue-v1: internal → 15-lead lane → monthly flow with 20% holdout

Done when: Spec §10 P1 exit: manual QA checklist passed; zero suppressed sends; holdout live

Next step: Two Rescue drafts are ready with verified CRM merge tags. Thomas approves/edits copy, chooses sender identity (per-staff recommended), confirms the internal→15-lead pilot gate, and records the consent basis. Activation remains a separate approval.

L3-5 Backlog 2026-11-14 Felix + content Q2

Nurture P2: value-refresh lane (stage-branched) + VIC local utility monthly

Done when: One full monthly cycle: bounce <2%, complaints <0.1%, zero suppressed sends

Next step: Location normalisation first (spec §10 P2); ~6 stage-branched templates needed.

L3-6 Backlog 2027-02-28 Felix Q3

Market-event email capability (preview-gated playbook)

Done when: Rate-decision template + audience + manual fire behind preview gate

Next step: After P2 stable.

L3-1 Done 2026-07-17 Felix Q1

6 Week Home Seller Plan workflow — reviewed, active/intentional (no defuse needed)

Done when: Confirmed the live workflow's timing is intentional; no action required

Next step: Closed by Thomas 2026-07-06 — workflow is active/intentional, nothing to defuse. (Template-timing hygiene folded into the rescue-template build, L3-4.)

L1 Estimate tool — own the valuation moment 0/8 done · 3 overdue Felix + Thomas

Target: Public 1 Sep; ≥2k sessions/mo + ≥150 verified captures/mo by Dec (→350–500/mo by Jun); top-10 on 'property value estimate' by Q3 · Estimate tool build spec · Strategy §6-L1

L1-1 Next 2026-07-18 · 30d overdue Felix Q1

Dead-traffic conversion: tool CTAs on house-prices-graph + calculators pages

Done when: CTAs live; lead events tracked; broken Chart.js error fixed

Next step: Overdue since 18 Jul — fourth consecutive board cycle with no shipped evidence. This week's pack shows the target house-prices page took 941 organic users to zero lead users over 28 days. At OPS-2: ship the conversion path or explicitly descope it into L1-5.

L1-2 Backlog 2026-08-14 · 3d overdue Felix Q1

Tool P0: endpoints/tokens/Redis/budget + CI leak gates + legacy route throttle

Done when: Spec §14 P0 exit: scripted 100-address scrape throttled/challenged; no provider terms in responses

Next step: Local production-review candidate exists (109 tests, no P0/P1 finding), internal only. Product-repo P0 still starts wk5; consumer exposure remains gated by D2/D3 and production integration.

L1-8 Backlog 2026-08-14 · 3d overdue Felix Q1

ODIN/OTH fallback guardrails in the tool (data-confidence gate)

Done when: OTH suburb-trends cache ranked below REA cache and above state defaults; never live-blocks the lead form; thin-data suburbs (zero recent sales) can't trigger market-moved claims; display approval (D3) recorded

Next step: [§6.7 FOUNDATION, gated on tool P0] Build with tool P0; fallback order already proven in PropertyValuationService.

L1-3 Backlog 2026-08-28 · 11d Felix Q1

Tool P1: Vite island + Places + lead flow + GTM events (controlled beta)

Done when: Spec §14 P1 exit: E2E green; first real leads carry valuation summaries into CRM

Next step: Copy fixes (phone, hero, stars) land here per spec §12 table.

L1-4 Backlog 2026-09-01 Felix + Thomas Q1

PUBLIC LAUNCH — estimate tool live

Done when: Live on paid + organic entry points; abuse alerting on; kill criteria agreed

Next step: Hard date. Does not move.

L1-5 Backlog 2026-10-31 Felix + content Q2

/property-value hub page (valuation keyword cluster)

Done when: Hub live with tool embed + suburb medians; indexed; internal links from calculators/suburb pages

Next step: The local property-value hub prototype and reusable SEO page workflow are proven, including a 27/27 live-link audit. Production rollout remains a separately approved release; due date unchanged.

L1-7 Backlog 2026-11-14 David Q2

Paid landing A/B: estimate tool vs standard funnel

Done when: Readout: CVR + cost/intro vs control; decision to shift traffic

Next step: Needs L1-4 + 2 weeks of data.

L1-6 Backlog 2026-11-28 Felix Q2

Tool P2: Turnstile + emailed report/magic-link + pvt_events to CRM

Done when: Spec §14 P2 exit: nurture events flowing; email verification live

Next step: After public launch stabilises.

L2 Agent-evidence data — the moat 0/8 done Felix (template) + content lane

Target: Self-serve agent connection controlled launch by 1 Sep; top-100 agent pages live 1 Sep; organic lead users 136→≥220/28d by Q3; Index story #1 in Oct · Strategy §6-L2 · Agent data platform

L2-1 In progress 2026-08-21 · 4d Felix Q1

Agent-evidence template v2 wired to existing rankings API

Done when: Template renders real sold counts / median price / DOM / fee context for a pilot suburb

Next step: Spec v2 was drafted 22 Jul: rankings-API contract, confidence-degraded mode, and 10-suburb pilot criteria. Thomas confirms; Felix renders one pilot suburb. No at-scale generation before L2-6.

L2-6 In progress 2026-08-28 · 11d Felix + Codex Q1

REA listing-status reconciliation live (gates public agent pages)

Done when: Agent on-sale/sold counts reconcile vs REA on the top-100 page suburbs; buy sitemap used for discovery only; every correction evidence-backed through ListingReconciler (no-write report → verify → correct)

Next step: Scraper recovery closed 4 Aug: PRs #431/#432 are merged and deployed, both sidecars are healthy at 4 serving + 6 reserves, closed-circuit and idle, and routine queue drain belongs to the normal 20:05 schedule. The exit still stands: top-100-suburb business-result reconciliation from queues through staged applies to actual listing changes is not proven.

L2-8 In progress 2026-09-01 Felix + Thomas Q1

Self-serve agent connection — controlled production launch

Done when: A controlled production cohort can select and connect with suitable agents; the displayed shortlist passes an agreed accuracy check; no agent, agency or data unique to REA is exposed; the scraper/data-confidence evidence is accepted; selection-to-intro outcomes are measurable

Next step: CRM #4511 and portal #1708/#1709 are deployed with every new flag off; scraper #440/#441/#442 is live. New launch blockers from the 14 Aug audit: the strict resolver verifies 17,549 of 38,178 sales people; the customer-facing cache marks 5,272 active people verified that the resolver rejects; and the ongoing Domain/RateMyAgent cycle is not operating (one stale canary, empty CRM handoff, no worker log). Sidecar pool health was unproved in that audit, not failed. Reconcile the cache and prove the worker, then rerun the Domain canary and reviewed national baseline dry-run. Import remains approval-required. At OPS-2, resource the gates to hold 1 Sep or move the date explicitly. Deployed flags-off is not launched.

L2-7 Next 2026-09-30 Felix + Codex Q1

RMA public-evidence lane on fresh sitemap snapshots

Done when: listing_public_evidence side table fed by fresh RMA sitemap snapshots with sold-date guard; legacy rma_listings not consulted anywhere; positive-evidence-only semantics documented

Next step: [§6.7 FOUNDATION — prioritised 2026-07-06] Productionise per docs/rma-rea-public-evidence-handoff-2026-06-23.md; startable now, ahead of net-new build.

L2-2 Backlog 2026-09-01 Content/AI lane (not Felix) Q1

Top-100 suburb agent-evidence pages generated + QA'd + live

Done when: 100 pages live on top-GP suburbs; Search Console clean; internal links from best-agents hub

Next step: Blocked by L2-1. Page generation runs on the content lane; descopes first if Q1 slips.

L2-4 Backlog 2026-10-31 Thomas (AI-drafted) Q2

WREA Agent Market Index — story #1 (data PR)

Done when: Published + pitched to property media once; links/brand-search movement tracked

Next step: Claude drafts 3 candidate angles from the 1.05M sold-history rows in September.

L2-3 Backlog 2026-11-30 Felix + content Q2

Suburb-market pages v1 (50–100 spring cohort)

Done when: Template with medians/DOM/stock + agent evidence + tool CTA; expansion gated on Search Console traction

Next step: After L2-2 traction check at Q2 gate.

L2-5 Backlog 2027-02-28 content Q3

Suburb pages → 300 where proven + Index story #2 (investor angle)

Done when: Expansion only where Q2 cohort shows impressions/clicks

Next step: Q3 gate decision.

L6 Convert what we already have — sales + CRO 0/7 done · 3 overdue Thomas + sales, Felix (metrics)

Target: Lead→intro 28%→30% by Dec→≥33% by Jun; organic form CVR 2.0%→3.0% · Strategy §6-L6

L6-1 Next 2026-07-31 · 17d overdue Felix or Codex Q1

Speed-to-lead metrics in weekly pack (first-contact time, attempt count)

Done when: Metrics on the Monday page from existing call logs; baseline established

Next step: Overdue (was 31 Jul; queries proven 22 Jul). Third consecutive board cycle without the Monday-pack block. Land the four SLA numbers plus median time to first connected call this week; L6-2 waits on it. If it cannot land, name the blocker at OPS-2.

L6-3 Next 2026-08-07 · 10d overdue Felix Q1

Form/JS fixes: btoa Latin1 error, jQuery error, form QA

Done when: Clarity error count at zero for 2 weeks; SubmitForm rate stable or up

Next step: Overdue (was 7 Aug). Clarity access is restored and the weekly pack carries JavaScript errors again, so the two-week zero-error exit is measurable. Ship the btoa/jQuery fixes with normal smoke proof and start the two-week clock.

L6-6 In progress 2026-09-30 Felix + Thomas Q1

CRM agent search / recommended-agent quality audit

Done when: Search + recommendation defects triaged with fixes listed on this board; sales floor confirms they trust the surfaces they sell from

Next step: Prevention is incomplete for profile-less listers: live code only deduplicates within one parse, leaving 5,416 excess rows across 439 identities from 5–13 Aug and 1,014 listings with extra links. PR #452 / DT-935 is blocked as unsafe because global name-only reuse collides on real people; the required fix is listing-scoped non-empty name-and-photo reuse with fail-closed person matching. Historical repair is DT-937, staged and approval-gated, and blocked by DT-935. The floor trust check still stands.

L6-2 Backlog 2026-08-14 · 3d overdue Thomas + sales Q1

Sales SLA live + weekly compliance reporting

Done when: Connected-call metric agreed and reported; compliance % in weekly pack; floor knows the number

Next step: Unblocked — D5 closed 2026-07-27, so the dependency is now L6-1, not D5. D5 resolved the coverage question by decision (no Sunday shift, off-peak paid pulled back to about midnight, self-serve is the out-of-hours answer) and the forensics cleared the floor on first-attempt speed, so this item is no longer about policing response time. What remains is reporting the connected-call metric once L6-1 lands it on the Monday pack (due Fri 31 Jul). Do not frame this as a sales performance gap.

L6-5 Backlog 2026-10-15 Thomas + Felix Q2

CRO cadence: one Clarity-informed funnel experiment/month

Done when: First experiment shipped + measured; cadence standing from October

Next step: First candidate: organic form gap (2.04% vs 5.07% paid).

L6-7 Backlog 2026-10-31 Thomas + Felix Q2

Property Value Form 2.6% conversion readout — Property Value Hub rollout gate

Done when: As part of the Property Value Hub rollout, explain the conversion gap on a stable cohort and decide whether the offer, report-first flow, follow-up cadence or lead intent needs to change before paid traffic scales through L1-7

Next step: TRIAGED 2026-07-27: no immediate standalone work. The production check proved these leads are not lost: all 78 leads in the 11-week cohort have tasks, and 77 received their lead-queue task within minutes. Keep the 2.6% conversion result as a rollout question, then add the cohort readout — including connected-call and attempt-depth evidence — to the L1-5 Property Value Hub / L1-7 paid-test checklist. Do not let this displace the L2-8 self-serve launch.

L6-4 Backlog 2026-12-12 Thomas + AI lane Q2

AI call-transcript pilot (sample of 173k recordings)

Done when: Readout: converting vs non-converting call patterns; script changes proposed

Next step: Claude designs the sampling + analysis plan in October.

L5 Yield per seller — agent routing, partners, investor wave 0/5 done Thomas

Target: Revenue/intro $732→≥$790, led by agent routing; partner pilot go/kill by Q3 gate; investor-exit campaign later (timing under review) · Strategy §6-L5

L5-1 Backlog 2026-12-19 Thomas + Claude Q2

Investor-exit prep: CRM investor segment + content plan + hub outline

Done when: Investor segment sized in CRM; hub + calculator-chain briefed; email lane drafted

Next step: Deprioritised to Lever 5's third theme (Thomas 2026-07-06) — no active commentary needed now; revisit at the Q2 gate. Launch date unchanged (Jan 2027 hard date stands).

L5-2 Backlog 2027-01-31 Thomas + all Q3

Investor-exit campaign LAUNCH (hub + email lane + exact-match paid)

Done when: Live with holdout on email lane; weekly economics vs core funnel

Next step: Hard date confirmed 2027-01-31 (Thomas 2026-07-06). Deprioritised as a theme, but the launch date stands — Q2 prep begins after the Q1 gate.

L5-3 Backlog 2027-02-27 Thomas + sales Q3

Agent routing v1: agent conversion scorecard + routing rules

Done when: Intros routed preferentially to converting agents; revenue/intro trend up

Next step: Now Lever 5's lead theme (Thomas 2026-07-06). Claude scopes the agent conversion scorecard from own listing/intro data in January.

L5-4 Backlog 2027-03-31 Thomas + sales Q3

Hot-market coverage gap-fill (WA/QLD/SA agent recruitment)

Done when: Unserved-intro rate in hot states measurably down

Next step: Needs unserved-intro measurement first — add to weekly pack in Q2.

L5-5 Backlog 2027-03-31 Thomas Q3

Partner pilot: 2–3 conveyancer/broker referral agreements

Done when: Signed or killed; if live, referral revenue per seller measured

Next step: Promoted to Lever 5's second theme (Thomas 2026-07-06). Blocked by D8.

OPS Operating rhythm 1/5 done · 1 overdue Thomas + Claude

Target: Weekly board review automated; monthly lever reviews; quarterly gates held on time · Strategy §8 · Growth OS

OPS-2 In progress 2026-08-03 · 14d overdue Thomas + Claude Q1

Monthly lever review #1 (green/amber/red, one decision per amber/red)

Done when: 60 min; every lever rated; decisions logged on this board

Next step: Was due 3 Aug; two Mondays have now passed. Hold the 60-minute review this week from the refreshed August draft; Thomas confirms or amends each rating, logs one decision per amber, then marks this done.

OPS-5 Next 2026-08-28 · 11d Felix Q1

Lead-worker and queue health detection (13 Aug incident prevention)

Done when: Production health checks fail when the portal lead worker is absent or queue age breaches the agreed threshold; the PHP floor matches the pinned 8.2 runtime; the remaining 12 Aug CRM-503 failed job is recovered or written off with evidence

Next step: The 32 queued incident jobs are recovered and the PHP 8.2 worker plus scheduler are restart-persistent. Pipeline and hourly HTTP checks still inspect neither worker presence nor queue age. Add those checks, raise the application PHP floor, and clear or evidence-close the one remaining failed job.

OPS-4 Next 2026-10-31 Felix + Codex Q2

Sidecar /fetch page-aware validation (data-freshness dependency)

Done when: /fetch validates adaptively per page type with a retry/error taxonomy; broad sidecar verification runs un-gated only after this lands (sample-gated until then)

Next step: [§6.7 FOUNDATION — prioritised 2026-07-06] Scope from the CloakBrowser hardening report; gates reconciliation (L2-6) Step-2 scaling, so it comes before net-new build.

OPS-3 Backlog 2026-09-28 Thomas Q1

Q1 gate review (strategy §7 exit criteria)

Done when: Tool public, rescue sending, agent pages live, value upload feeding, exposure closed, speed-to-lead on pack — each yes/no with named blocker

Next step: Scheduled; Claude assembles the evidence pack the Friday before.

OPS-1 Done 2026-07-13 Claude (+ Codex wiring) Q1

Execution board live + weekly Telegram automation

Done when: Board renders + sends weekly without manual steps; Claude updates data file each Monday; KPI help + real 12wk scorecard refresh live

Next step: Closed 2026-07-27: unattended Monday sends succeeded 6, 13 and 20 Jul; the 2-week proof gate is met. Pre-send ledger reconciliation and scorecard refresh are wired; monitor via run.log and reopen on failure.

Source of truth: artifacts/wrea/fy27-execution/board.data.mjs — edited by Claude in the Monday review; this HTML is generated by scripts/wrea_fy27_board_render.mjs. Delegate any task by ID ("progress L3-2") to Claude or a team member.