Build

Value Architecture β€” Driver Tree

6 min read

🌳 Value Architecture β€” Driver Tree

What this is

The Mission & Policies charter says what we maximise: net worth that compounds every cycle. This note decomposes that single outcome down into the smallest numbers an agent can actually move, and names the lever for each one. When an agent asks β€œwhat should I do to grow the business?”, the answer is: find the weakest node below, pull its lever. This is the map from a number on the dashboard to an action in the world.

How to read it: each box is a measurable node. A parent equals its children combined (=, Γ—, βˆ’). Move a leaf node and the effect ripples up to net worth. The constraint is whichever leaf, if improved, moves the top number most for the least effort β€” find it, fix it, re-evaluate.

The tree

flowchart TD
    NW["πŸ“ˆ NET WORTH (to date)<br/>= Profit βˆ’ Draws + Assets<br/>cash commingled β€” see note"]

    NW --> NP["πŸ’° Net Profit (to date)<br/>β‰ˆ RM8,200"]
    NW --> OD["πŸ”» Owner Draws<br/>tangled in personal acct"]

    NP --> GP["Gross Profit<br/>RM8,563 (66% margin)"]
    NP --> OX["πŸ”» Operating Expense<br/>light β€” software ~RM105/mo"]

    GP --> REV["🟣 Revenue (to date)<br/>RM12,950"]
    GP --> COGS["πŸ”» COGS<br/>RM4,387"]

    REV --> ORD["Orders Γ— AOV"]
    ORD --> NORD["# Orders<br/>(volume)"]
    ORD --> AOV["Avg Order Value<br/>(price Γ— mix)"]

    NORD --> CH1["πŸ‘₯ Resellers β‰ˆ 97%<br/>Muz 93% Β· Kio Β· Haikhal<br/>THE engine + the risk"]
    NORD --> CH2["🧡 Threads<br/>RM0 β€” untapped"]
    NORD --> CH3["πŸ“± TikTok<br/>RM0 β€” untapped"]
    NORD --> CH4["πŸ›’ Direct β‰ˆ 3%<br/>Najiha"]

    AOV --> PR["Price ladder<br/>R10β†’R60, two tiers"]
    AOV --> MIX["Product mix<br/>+ bundles / stacks"]

    COGS --> LAND["Landed cost / unit<br/>supplier waterfall"]
    COGS --> STOCK["Stock availability<br/>(stockout = lost sale)"]

    OX --> FIX["Fixed burn<br/>Lovable RM105 Β· Cursor Β· domain Β· phone"]
    OX --> VAR["Variable<br/>ad spend Β· shipping Β· payment fees"]

    NW --> ASSET["🏭 Productive Assets<br/>printer · brand kit · ad accounts<br/>(lower future cost / raise price)"]
    NW --> TRUST["πŸ›‘οΈ Trust Equity<br/>COA / Verify Β· packaging Β· brand<br/>(raises conversion at every node)"]

    classDef good fill:#1f7a3d,stroke:#0d3,color:#fff;
    classDef cost fill:#7a2d2d,stroke:#d33,color:#fff;
    classDef lever fill:#2d4a7a,stroke:#39d,color:#fff;
    class NW,NP,GP,REV good;
    class OD,OX,COGS,LAND,STOCK,FIX,VAR cost;
    class CH1,CH2,CH3,CH4,PR,MIX,ASSET,TRUST lever;

The equation, in one block

Net Worth Growth   = Net Profit βˆ’ Owner Draws
Net Profit         = Gross Profit βˆ’ Operating Expense
Gross Profit       = Revenue βˆ’ COGS
Revenue            = (# Orders) Γ— (Average Order Value)
# Orders           = Agents + Threads + TikTok + Website demand
Average Order Value= Price (ladder, correct tier) Γ— Product mix
COGS               = Units sold Γ— (landed cost + pack cost) / unit
Operating Expense  = Fixed burn + Variable (ads + shipping + fees)

Two nodes don’t show up in this period’s profit but compound everything above β€” Productive Assets (lower future cost / raise price power) and Trust Equity (raise conversion at every channel). Stage 2 of the ladder exists to fund exactly these.

Node β†’ Lever table

Each leaf, the number that lives on it, the lever that moves it, and who pulls it (pillar β†’ owner). This is the action layer.

🟣 Revenue side β€” grow the top

NodeCurrent read (real)Executional leverOwner
# Resellers3 (Muz, Kio, Haikhal) β€” Muz = 93% of all revenueRecruit to β‰₯4; repeatable agent offer; grow Kio/Haikhal off the Muz dependency🟣 Distribution β†’ Community & Agents
Revenue / resellerNow measurable per-customer (orders logged) β€” Muz 12,030 Β· Kio 370 Β· Haikhal 150Add agent: field so Bases group it automatically; track monthly🟠 Data Model
Threads demandRM0 β€” untapped3–5 research-led posts/week β†’ Verify/COA β†’ WhatsApp close🟣 Distribution β†’ Threads
TikTok demandRM0 β€” untapped (the old β€œcampaign” was fake data)Weekly video batch; hand DMs to a reseller🟣 Distribution β†’ TikTok
Direct (Najiha-type)RM400 (3%)Verify page β†’ WhatsApp close SOP🟣 Distribution
Price (AOV)Reta RM400 dropship; public ladder unset; early sales went at RM350Complete R10/R20/R30 public + dropship; hold ~23% below Olymp🟒 Capital β†’ Pricing Model
Product mixReta = ~95% of revenueBuild Wolverine blends (5 possible); price + sell GHK/CJC; stacks🟣 + 🟒

πŸ”» Cost side β€” keep more of it

NodeCurrent read (real)Executional leverOwner
Landed cost/unitReta RM133; 66% realised gross marginSource down the waterfall BFF(Reta)β†’TCIβ†’Uther; update #costing on every PO🟠 Operation β†’ Suppliers
Stockout (lost sales)Reta down to 4; GHK-Cu = 0 (sold out)Restock Reta now; reorder-point discipline🟠 Operation β†’ Restock SOP
Fixed burnLight β€” Lovable RM105/mo (Janoshik COA / domain / phone not yet logged)Log all recurring costs; cut tools that don’t earn🟒 Capital
Variable costshipping + payment fees (no ad spend)Batch shipping; watch fee leakage; keep any future ROAS β‰₯2.0🟒 + 🟣

🏭 Compounding nodes β€” make every future cycle better

NodeWhy it compoundsExecutional leverOwner
Productive assetsPermanently lower COGS/opex or raise price powerStage 2: buy in-house packaging/printer, owned brand kit, ad accounts🟒 Capital (funds) β†’ 🟣/🟠 (use)
Trust equityRaises conversion at every revenue node at onceCOA/Verify everywhere; research-grade packaging; consistent brand🟣 Distribution β†’ Brand pipeline

Where the constraint is right now (2026-06)

Read the tree against today’s reconciled reality. The books are now clean (RM12,950 revenue, RM8,563 gross profit) β€” so the constraint has moved. In order:

  1. Commingled cash / founder ↔ company (THE constraint now). Every ringgit sits in the founder’s personal account and the founder funded purchases personally β€” so the business has no clean cash figure and can’t safely fund its own next cycle. Settle the founder ↔ company balance and separate a business account before anything else. β†’ Stock & Revenue Reconciliation Part D.
  2. Brutal reseller concentration. Muz alone is 93% of revenue; resellers are 97%; TikTok/Threads/website are RM0. If Muz stalls, revenue stalls. Widen the base (recruit, grow Kio/Haikhal) and stand up the dormant channels. This is now the biggest growth risk.
  3. Reta is the whole business and it’s nearly out. ~95% of revenue is Retatrutide and the shelf is down to 4 vials; GHK-Cu is sold out. Restock Reta β€” a stockout here is a near-total revenue stop.
  4. Price ladder incomplete + dead inventory. R10/R20/R30 unpriced; CJC/SS-31/MOTS-C sitting unsold. AOV and mix upside for low effort.

The agent's standing question

Each week: β€œWhich leaf node, if I moved it, adds the most to net-worth growth for the least effort β€” and am I allowed to (Stage rule, ROAS β‰₯2, research-use, buffer intact)?” Pull that lever. Re-check next week β€” the constraint moves as you fix it.

Graph View

Backlinks