90-day roadmap

Weeks 1–2: PostgreSQL ingestion

  • schedule feed download and direct COPY ingestion
  • create unlogged staging and canonical tables
  • normalize and deduplicate domains
  • create service and geography taxonomies
  • record checksums, parser versions, counts, and failures
  • expose daily funnel counts

Weeks 3–4: lifecycle watchlist

  • implement cheap deterministic filters
  • integrate RDAP with a controlled WHOIS fallback
  • store state changes as append-only observations
  • implement registry-specific drop-window rules
  • close renewed, transferred, and auction-only candidates
  • validate predictions against actual drops

Weeks 5–6: full intelligence

  • trigger paid enrichment only for confirmed pendingDelete candidates
  • add archive, safety, backlink, and anchor checks
  • predict keyword clusters
  • request city, county, metro, and state keyword metrics
  • calculate modeled geo demand, lead value, and confidence
  • publish the first drop-day ranked report

Weeks 7–8: scoring and execution

  • version scoring weights and risk rules
  • add structured AI and adversarial reviews
  • build the approval and drop-queue screen
  • integrate at least two registrar APIs
  • implement locks, idempotency, reconciliation, and attempt logs
  • run dry tests without purchase calls

Weeks 9–10: WordPress factory

  • finalize the block theme and pattern library
  • define the site-brief schema
  • automate staging builds and content injection
  • configure measurement and call tracking
  • automate technical and visual QA

Weeks 11–12: controlled operation

  • enable human-approved timed registration attempts
  • launch a small cohort
  • measure capture rate, indexing, rankings, calls, and leads
  • inspect missed and rejected candidates
  • adjust lifecycle rules and scoring weights
  • document operator procedures and incident recovery

By the end of week four, the system should produce a credible lifecycle watchlist. By the end of week eight, it should produce and execute a human-approved ranked queue for each drop window.