Skip to content

Cost Estimates — Phase 1

This is a Phase 2 addendum to the Phase 1 Green cost document. It itemizes everything we should account for as we transition into Phase 2 (paid motion + archetype scaling), without assigning dollar values yet.

Phase 2 Definition (what changes vs Phase 1)

Phase 2 begins when: - We run a paid employer motion (contracts + billing + procurement realities) - We operate archetype-focused GTM (repeatable per-archetype loops) - We shift from “founder-heavy ops” to real staffing, real payroll, and measurable customer success

A. One-Time / Step-Function Costs (Phase 2 “enablement”)

A1) Payroll + employment setup (Day 1 required)

  • Payroll provider selection + setup
  • State registrations / tax accounts (as needed)
  • Contractor → employee conversions (if any)
  • HR policy baseline (handbook-lite)
  • Offer letters + standard employment agreements
  • Background checks (optional; depends on role sensitivity)

A2) Billing + revenue operations (Day 1 required for paid)

  • Payments + invoicing setup
  • Product catalog / plans / metering definitions
  • Proration + upgrades/downgrades rules
  • Dunning (failed payments) workflow
  • Refund policy and handling rules
  • Basic revenue reporting (MRR/ARR, churn definitions)

A3) Contracting + procurement baseline (Day 1 required)

  • MSA + Order Form templates
  • Data Processing Addendum (DPA) and privacy addendum
  • Security addendum / standard security position doc
  • Vendor onboarding packet (W-9, insurance certs if needed)
  • Redline workflow + internal approval rules

A4) Security baseline uplift (Day 1–Day 2 likely)

  • Centralized secrets management and rotation policy
  • Access control hygiene (least privilege, role separation)
  • Audit logging posture (who did what, when)
  • Incident response playbook (even if small)
  • Vulnerability scanning + dependency monitoring
  • Pen test planning (often Day 2, but customers ask early)

A5) Privacy + data retention posture (Day 1 required)

  • Documented retention policy for:
  • prompts, transcripts, claims, embeddings, match briefs
  • Data deletion workflow (manual is OK at first, but defined)
  • Data export workflow (manual is OK at first)
  • PII boundaries documented and enforceable
  • Customer requests process (privacy inbox, response SLA)

A6) Trust & safety escalation readiness (Day 1–Day 2 likely)

  • Abuse case taxonomy + severity levels
  • Moderator/admin tooling (even lightweight)
  • Evidence capture and audit trail for enforcement actions
  • Bidirectional reporting workflows:
  • candidates report companies
  • companies report candidates
  • Appeal workflow rules (what is appealable, what isn’t)

A7) Customer support + ops “realness” (Day 1 required)

  • Support intake channel(s) + triage playbook
  • Standard response SLAs and escalation path
  • Account onboarding checklist
  • Account offboarding checklist
  • Internal runbook for “incident / outage” communications

A8) Sales stack enablement (Day 1–Day 2 likely)

  • CRM configured properly (stages, fields, attribution)
  • Outreach sequencing tool (optional)
  • Sales collateral baseline:
  • one-pager
  • deck
  • product walkthrough
  • security overview
  • Standard “pricing + packaging” FAQ doc
  • Pipeline reporting (lead → meeting → pilot → paid)

A9) Insurance + corporate admin (Day 2 likely, but often requested)

  • General liability
  • E&O / cyber coverage
  • Finance ops scaling:
  • bookkeeping cadence
  • tax prep planning
  • cap table tool (if raising)

B. Recurring Costs (Phase 2 run-rate)

B1) Headcount (recurring)

You’ll likely carry some mix of: - Customer ops / support (baseline coverage) - Sales (founder-led initially, then hire) - Partnerships / BD (optional) - Marketing (content + lifecycle + paid ops) - Trust & safety / risk (could be fractional early) - Engineering (product + platform + reliability) - Security (could be fractional/consulting early)

B2) Compute & platform (recurring; scales with usage)

  • LLM inference (prompt depth grows with paid expectations)
  • Evals and monitoring (quality + safety regression testing)
  • Observability (logs/metrics/traces) at higher volume
  • Background jobs + orchestration (pipelines, follow-up loops)
  • Storage growth (transcripts, evidence, brief history)

B3) GTM (recurring)

  • Paid acquisition budgets (split by archetype + channel)
  • Content production (founder + contractor + internal)
  • Community/partnership sponsorships (optional)
  • Sales tools (CRM seats, automation, enrichment)

B4) Customer success + retention (recurring)

  • Onboarding support per account
  • QBR cadence (for higher-touch accounts)
  • Renewal management + expansion motion
  • Training materials and enablement updates
  • Account health scoring and interventions
  • Ongoing contract redlines
  • Periodic policy updates
  • Compliance program work (e.g., SOC2 path)
  • Data protection requests handling

B6) Trust & safety operations (recurring)

  • Moderation time (manual reviews)
  • Abuse monitoring and investigations
  • Enforcement actions and appeals

C. Per-Archetype Cost Model (repeatable loop)

Phase 2 is not “one transition.” It’s a repeated loop as you add archetypes. Each archetype incurs:

C1) One-time per archetype (setup)

  • Archetype definition:
  • scope, seniority bands, geo/timezones, comp bands (optional)
  • Claim schema tuning:
  • facet prompts tuned to the role
  • follow-up triggers tuned to uncertainty patterns
  • Quality calibration pack:
  • small labeled set of good/bad match examples
  • rubric for “Match Brief quality” for this archetype
  • Distribution assets:
  • landing page variant
  • “example agents” (candidate + role)
  • 2–4 new creatives + new ad angles
  • Channel map:
  • target subreddits, communities, newsletters, slacks
  • employer list for outreach
  • Sales enablement:
  • objection handling specific to that archetype
  • ROI narrative specific to that archetype

C2) Recurring per archetype (run)

  • Paid spend allocation (prospecting + retargeting)
  • Creative refresh cadence (fatigue is faster in narrow audiences)
  • Content cadence specific to archetype pain points
  • Supply/demand balancing work:
  • if candidate inventory is low, spend/time shifts there
  • if employer demand is low, spend/time shifts there
  • Ongoing calibration:
  • monitor match outcomes
  • update prompts/triggers and weighting
  • Trust & safety patterns:
  • archetype-specific gaming attempts and checks
  • Customer success playbook:
  • “how to interpret briefs” for that archetype
  • interview question templates, handshake guidance

D. “Hidden” Phase 2 Costs to explicitly account for

These are the ones that sneak up:

D1) Payment processing fees

  • Processor take rate
  • Chargebacks / disputes handling
  • Refund processing overhead

D2) Integrations and export demands

  • ATS integration requests (Greenhouse/Lever etc.)
  • Calendar / scheduling integration requests
  • Data export format expectations (CSV, JSON, audit logs)

D3) Procurement friction cost

  • Security questionnaires (time sink)
  • Vendor onboarding portals
  • Legal redline cycles (slow burn)

D4) Operational resilience

  • On-call expectations
  • Incident communications
  • Backups + restore testing cadence

E. Phase 2 KPIs that influence spend

This is not a metrics doc, but these are the spend-driving measures:

E1) Per-archetype acquisition efficiency

  • Candidate CPA (start + activated)
  • Employer CPL (lead + qualified lead + meeting booked)
  • Activation rates and completion depth

E2) Matching throughput & quality

  • Match-ready inventory thresholds by archetype
  • Coverage ratio by archetype
  • Brief yield rate (BYR)
  • Mutual affirm rate (MAR)

E3) Trust & safety load

  • Report rate per 1,000 sessions
  • Manual review minutes per incident
  • Enforcement rate and appeal rate

F. Phase 2 Checklist (to turn into a real spreadsheet later)

When we add numbers, we’ll want to model:

1) Fixed monthly headcount plan (roles + start month) 2) Payroll burden/benefits assumptions (now applicable) 3) Paid GTM budgets split by: - channel (Reddit / LinkedIn / others) - objective (candidate vs employer) - archetype 4) Variable compute: - prompts per engaged visitor - calls per prompt by workflow - token budgets (input/output) - follow-up loop parameters 5) Contracting/procurement load: - average redline hours per deal - security questionnaire hours per deal 6) Customer success load: - onboarding hours per account - support tickets per account per month 7) Trust & safety load: - incident rate and review time 8) One-time enablement costs (SOC2 path, pen test, etc.) 9) Per-archetype setup costs (content + calibration + enablement) 10) Per-archetype recurring costs (spend + ops + support)