AnchorWorks ← The Briefing Room On the record
AnchorWorks
Executive Brief 004 · July 2, 2026

One Source of Truth

system introduction
Prepared by Chris Campbell AnchorWorks
The Briefing Room · 2026
  • Reconciled-ledger model§ Executive Decisions · Decision Log

    Adopted the reconciled-ledger model — the database is the ledger of record, never a hand-edited source of truth.

  • System-of-record per field§ Executive Decisions · Decision Log

    Established system-of-record per field — GoHighLevel owns pipeline; Stripe owns cash; the ledger owns totals.

  • One metric, many altitudes§ Executive Decisions · Decision Log

    Adopted one metric, many altitudes for ROAS/CAC; MER is the CEO's one number.

  • Two-axes rule§ Executive Decisions · Decision Log

    Adopted the two-axes rule — capture data day-one and universal; build dashboards in sequence.

See every standing order in the SOP Library →
01 · Executive Summary

The executive summary.

A full working session designed AnchorWorks' company-wide KPI and revenue-intelligence system — the KPI Command Center — and locked the technology, architecture, and doctrine that will run it. From first principles to a filed, version-controlled plan in its own repository, the measurement operating system now exists on paper, ready to build.

It means AnchorWorks will run on truth, not gut. One reconciled source of truth replaces the fifteen tools that quietly disagree; decision-first dashboards surface only the calls each leader must make; and the architecture is shaped like the $100-million version today, so it scales without a rebuild. Because every department's KPI is also an input to the financial model, the same system that measures the business also forecasts it.

Every number the company runs on — from a single ad's ROAS to blended enterprise value — now has one definition, one home, and one owner.

Current stage: Foundation — pre-launch, laying the instrumentation before the first client so the ship is pointed right before it leaves harbor.

02 · Executive Decisions · Decision Log

What was decided.

  • Stood up the KPI Command Center as its own private repository — anchorworkshq/KPI-Command-Center.

    locked

  • Locked the stack: Supabase (central database) + AI-generated dashboard built with Claude Code + MCP (live data) + GitHub.

    locked

  • Adopted the reconciled-ledger model — the database is the ledger of record, never a hand-edited source of truth.

    locked

  • Established system-of-record per field — GoHighLevel owns pipeline; Stripe owns cash; the ledger owns totals.

    locked

  • Adopted one metric, many altitudes for ROAS/CAC; MER is the CEO's one number.

    locked

  • Adopted the two-axes rule — capture data day-one and universal; build dashboards in sequence.

    locked

  • Defined the 7-department model + roll-up hierarchy + the two-meeting operating system.

    locked

  • Locked Ownership unit economics — CAC $1,000 · LTV ~$135K · LTGP:CAC ~71:1 · client-financed acquisition achieved.

    locked

  • Confirmed the offer ladder — $500 workshop → $4,800 / $7,800 → $15K / $30K partnership + 10%.

    locked

  • Set the dashboard design language — the MINIMIN aesthetic recolored to brand; gold is earned.

    locked

  • Established the operating model — founder owns all accounts and keys; Ethan supervises the build agent.

    locked

  • Retired Google Sheets as the production store; ruled out submodules — brand assets are vendored.

    locked
03 · Major Accomplishments

What was done.

01The Measurement Operating System

The problemKPIs were destined to scatter across fifteen tools with no single source of truth. At the prior agency, retrofitting tracking onto months of live data was a nightmare — dirty data that could not be reconstructed and a team that never adopted the numbers.
What was doneDecision: A four-layer architecture (collection → storage → reconciliation → presentation) where the database is the reconciled ledger, not a hand-edited source of truth — governed by system-of-record-per-field, one-metric-many-altitudes, two-axes capture, build-once-filter-many, and decision-first dashboards. Implementation: Filed as doctrine in the repo — 00-architecture/ (master architecture, source-of-truth map, governing principles, collection layer) and the Day-One Foundation punch list — and mirrored into the Obsidian concept map. Dependencies: The collection layer (pixels + webhooks) must be live from day one — pixels are forward-only.
The impactEvery number gets one definition, one home, and one owner. The system is shaped like the $100M version now, so it scales from pre-launch to enterprise without a rebuild.

02The Stack, Decided

The problemThe system needed to be simple enough to run today yet enterprise-shaped for scale. A spreadsheet as the hub hits hard ceilings (cell limits, write throttling, no locking); a hand-built dashboard is powerful but high-maintenance.
What was doneDecision: Supabase as the central database + an AI-generated dashboard built with Claude Code + MCP for live data + GitHub for version control. Ingestion via Supabase Edge Functions (webhooks) and pg_cron (scheduled pulls). Stripe is the cash source of truth. Implementation: Validated against a trusted AI co-build course (whose own default tool is Claude Code). Documented in 05-build/stack-options.md and 00-architecture/collection-layer.md; design tokens and reference filed. Dependencies: Founder-owned accounts — Supabase, Google AI Studio, and a deployment host.
The impactA real database with no spreadsheet ceiling, a beautiful on-brand custom dashboard, live data, and an AI-built UI that keeps it simple — the "$100M shape and keep-it-simple" tension resolved.

03The Numbers Behind the Machine

The problemBefore building the tracking, the business itself had to be proven fundable and sellable — the model, not just the dashboard.
What was doneDecision: Lock the Ownership unit economics — CAC $1,000 ($200/booked call × 20% close), LTV ~$135K, LTGP:CAC ~71:1, with client-financed acquisition achieved — and confirm the offer ladder as a Hormozi money model with a $100k/month graduation gate driving ascension. Implementation: A driver-based projection model where department KPIs are the forecast inputs; validated against the prior agency's real history (~50 clients/month, 250+/year). Filed in 04-economics/. Dependencies: Replace modeled assumptions with live actuals once the company launches.
The impactThe build-to-sell thesis is quantified and robust to a 2× downside. Early churn is identified as the single real risk, with ≤30-day time-to-first-lead as its insurance.

04Built to Hand Off

The problemA plan is worthless if the founder is the only one who can build it. Delegation without a clean boundary means lost context and endless revisions.
What was doneDecision: The system lives in its own repository as a clean handoff boundary — the founder owns every account and key; a partner (Ethan) supervises the Claude build agent. The dashboard follows the founder-chosen MINIMIN look recolored to brand. Implementation: A build-ready handoff kit — BUILD-HANDOFF.md (kickoff prompt + Definition of Done), a founder setup checklist, tokens.css, the vendored logo, design references, and a Google-Sheet KPI prototype to validate the math first. Dependencies: Ethan's GitHub username / access; the founder's Supabase and connector setup.
The impactDevelopment is delegated without losing control or context, and revision risk is minimized by exact, unambiguous specs.

05Why this increases enterprise value.

What was doneA measurement system is the difference between running the company on gut and running it on truth — and, done right, it is an enterprise-value asset in its own right. Enterprise value — A system a buyer trusts: Clean, auditable numbers, one source of truth, and low owner-dependence are exactly what lift the exit multiple. Less chaos — One number, one owner: No more fifteen tools disagreeing. The financial reconciler owns one ledger; leaders read it, they don't fudge it. Faster execution — Decisions in seconds: Decision-first dashboards surface only the two or three calls each leader must make — not a wall of vanity metrics. Compounding — Capture today, understand forever: Day-one capture means every future dashboard reads full, clean history the moment it is built. No retrofits.
04 · Metrics

By the numbers.

1
New repository stood up
35
Files filed in the repo
7
Departments defined
4
Architecture layers
71:1
LTGP:CAC (modeled)
$135K
Modeled LTV / client
100%
Planning & doctrine
100%
Stack & design locked
Queued
Dashboard build
05 · Current Company Stage

Where we are.

01
Foundation
02
Launch
03
Scale
04
Expansion
05
Acquisition

Pre-launch. The measurement system is fully designed, filed, and version-controlled; the build is sequenced and ready. AnchorWorks is in Foundation — deliberately instrumenting the business before the first client, so the ship is pointed right before it ever leaves the harbor.

06 · Open Decisions

Decisions still required.

  • Exit horizon and target enterprise value — sets the trajectory the financial projection must hit.
  • Offer-ladder specifics — the down-sell-to-partnership credit ($7k vs $8k), the $30K monthly fee, and expected ascension rates.
  • Reconcile the dashboard mock-up's "4-Pillar Revenue System · Retention" naming with canon — the "4 Engines: Traffic · Conversion · Sales · Nurture."
07 · Next Priorities

What comes next.

  1. 01
    Founder foundation setup: Create the Supabase project, stand up the founder-owned accounts, and collect credentials — the work that unblocks the build.
  2. 02
    Grant access & kick off Phase 0: Add Ethan via a GitHub Team; begin the Day-One Foundation capture layer.
  3. 03
    Validate the KPI model: Pressure-test every calculation in the Google-Sheet prototype before a line of dashboard is built.
  4. 04
    Stand up day-one capture: Pixels + the GoHighLevel data model live before traffic — retargeting audiences are forward-only.
  5. 05
    Build the dashboard spine: Supabase + the first live dashboard, then sequence outward — acquisition & sales, then churn defense.
Executive Brief · Nº 004 · July 2, 2026 · Canonical