Satoshi-Precision Operating Model

The unifying principle. Every cost element in the merchant’s operating model — Cost of Goods Sold, Customer Acquisition Cost, SG&A — and every physical and digital movement that drives those costs is tracked at satoshi precision (sub-cent, ledger-grade) end to end. IoT extends the movement-tracking surface beyond POS-only to every meaningful physical event in the store. The Virtual Store Manager surfaces this granular truth; merchants for the first time know what every cent of every cost actually was, where it went, and what it produced.

Why this article exists

The satoshi-cost-accounting substrate article named sub-cent unit cost on the perpetual ledger as a Canary innovation. That was the right starting move but undersold the principle. Satoshi precision is not a ledger feature — it is the operating-model commitment that makes the entire 13-module spine internally consistent and externally credible.

If the perpetual ledger is satoshi-precise but COGS rolls up at penny precision, the precision falls out the bottom. If COGS is satoshi-precise but CAC and SG&A are guessed monthly from accounting summaries, the merchant cannot answer the only question that actually matters: “Did this customer make me money, and how do you know?”

The Virtual Store Manager’s promise is that every dollar in and every dollar out of the merchant’s business is decomposed to its originating event with full audit trail. That promise requires satoshi precision through the entire stack, not just on inventory cost. This article names that commitment and walks each cost dimension that has to honor it.

The four cost dimensions

1. COGS at satoshi precision

Already grounded in satoshi-cost-accounting. Every unit’s landed cost is tracked in millisatoshis (cost_msat BIGINT) alongside fiat cents (cost_usd_cents INT), with the FX rate captured at posting time. Every movement on the perpetual ledger carries both representations. Variance, write-offs, shrink, and markdown all post against the satoshi side as the authoritative cost basis; fiat is the display projection.

What this enables: per-unit margin queryable in real time at full precision, even for items where the per-unit margin is fractional cents (e.g., bulk consumables, low-value inventory, fractional packaging).

2. CAC at satoshi precision

Customer Acquisition Cost is the most-faked number in SMB retail. Most merchants compute it as total marketing spend ÷ new customers this month — an aggregate metric that hides every actionable fact.

Satoshi-precision CAC decomposes:

• Channel-level spend posted at touch time. Every ad impression / click / placement charge, every SEO content piece’s amortized cost, every email send’s compute + delivery cost — each posted as a channel-specific event on the perpetual ledger at sub-cent precision.
• Attribution computed against the actual customer journey on the ledger (first touch, last touch, multi-touch with weights). The customer journey is itself a stream of perpetual events: page views, cart adds, abandonments, returns, eventual purchase.
• Per-acquired-customer rollup as a derived projection. CAC for customer X = sum of channel spend events attributed to X’s journey. Reported in fiat for human consumption; backed by satoshi-precise underlying events.

What this enables: merchants can ask “what did Customer X cost to acquire and what’s their lifetime value so far?” and get a real answer with audit trail back to the originating ad clicks. They can compare channels at the level of return on satoshi spent rather than return on dollar spent rounded.

3. SG&A at satoshi precision

Selling, General & Administrative is the catch-all that absorbs every overhead element merchants do not bother to decompose. Satoshi-precision SG&A decomposes:

• Labor cost per shift per minute — driven by v3.L Labor module’s time-clock event stream. Each minute on the clock posts a labor-cost event on the perpetual ledger at the employee’s precise rate (with shift differentials, overtime, etc.) in millisatoshis.
• Rent allocated per minute per square foot — fixed monthly rent expense decomposed into minute-by-square-foot accruals; reportable per location, per department within a location, per fixture.
• Utilities per device per minute — electrical / network / cloud costs allocated to specific devices (POS, cameras, IoT sensors) by measured consumption (where IoT enables measurement) or attribution (where it does not).
• Software costs per agent-tool-call — every Canary tool call gas- metered through the satoshi-cost GasMeter (already shipping per Goose / GRO-117). External SaaS subscriptions (QuickBooks, Gusto, etc.) accrue per-active-user per-day against the merchant’s wallet.
• Cost-of-tool-not-Canary — for every package the merchant integrates with under integrated-hybrid, the monthly subscription cost is captured as a per-day SG&A accrual against the merchant’s perpetual ledger.

What this enables: per-transaction true cost (COGS + allocated SG&A + allocated CAC). Per-store true contribution margin. Per-product-per-store true profitability. None of this is currently visible at SMB tier; merchants run on guesses.

4. IoT-tracked movement (the new perpetual-event source)

The 2002-vintage perpetual ledger ingested only POS movements. Modern retail has 100× the meaningful events available, gated only by instrumentation. The Internet of Things is the instrumentation.

Movement sources that an IoT-enabled Canary install captures as ledger events:

• Foot traffic — door counters, computer-vision people-counters posting customer-arrival and customer-departure events; conversion rate (visitors → transactions) computable in real time per minute per store.
• Dwell-and-flow — in-store positioning beacons posting dwell-time events per zone per visitor; “sections people walk past without stopping” becomes a measurable KPI.
• Shelf state — RFID tags or computer-vision shelf monitors posting out-of-stock detection events the moment a shelf empties, not at next cycle count.
• Cold-chain — temperature sensors on refrigerated units posting temperature-event streams; deviation events trigger automatic shrink reason codes on the perpetual ledger.
• Door / safe / drawer events — physical access events on cash drawers, back-room doors, safes; cross-referenced against employee clock-state for security and shrink attribution.
• Equipment health — meter readings, fault codes, maintenance events on every device in the store; preventive-maintenance events posted to the A Asset Management module’s bubble.
• Customer device interactions — receipt-printer events, POS terminal events, kiosk interactions — all already partially captured; IoT extends to peripherals not previously instrumented.

Every IoT event is a movement on the perpetual ledger with a typed schema. The ledger’s verb taxonomy (the-canonical-movement-verbs) extends to accommodate the new verbs (foot_traffic_arrival, shelf_oos_detected, cold_chain_excursion, equipment_fault, etc.). Each event is satoshi-stamped with its capture cost (the marginal cost of the IoT sensor reading) and contributes to the merchant’s SG&A allocation if the sensor is itself a metered service.

What this enables: a perpetual-ledger view of the store that includes everything happening in the physical space, not just the cash-register events. The Virtual Store Manager can answer questions like “how many people walked past the new endcap and how many converted?” at the same fidelity it answers “what was today’s shrink in dairy?“.

The end-to-end commitment

Read together, these four dimensions describe a merchant operating model where:

1. Every cost is decomposed to its originating event. No more “marketing spend $4,200 / divided by 23 customers / $182 CAC.” Instead: “Customer X cost $173.42 to acquire — here are the 17 channel events that produced them, here’s the conversion path, here’s the attribution weighting we used.”

2. Every event is captured at satoshi precision in real time on the perpetual ledger. Display rounds to fiat for humans, but the underlying truth is sub-cent and audit-traceable.

3. Every roll-up is derivable from the events. The merchant’s monthly P&L, COGS schedule, marketing performance report, labor productivity scorecard, and contribution-margin-by-product report are all projections of the same perpetual-ledger event stream. There is no separate truth in a marketing dashboard or an HR system that conflicts with the ledger.

4. Every cost has audit trail. A merchant can drill from a P&L line down to the originating event in seconds. Auditors, investors, lenders all consume the same perpetual-ledger truth as the operator.

What this means for the spine

This principle re-reads the existing 13 modules as instruments of cost decomposition:

Spine module	COGS contribution	CAC contribution	SG&A contribution	IoT contribution
T Transactions	Sale at fiat-display price	First/last touch capture for attribution	Payment-processor fees per transaction	Receipt-printer + terminal events
R Customer	Customer identity for attribution rollup	Per-customer journey reconstruction	n/a	Customer-device interactions
N Device	n/a	n/a	Equipment health → maintenance accrual	Device telemetry as canonical IoT source
A Asset Mgmt	n/a	n/a	Anomaly cost attribution	IoT bubble per device
Q Loss Prev	Shrink → COGS write-off	n/a	Investigation labor	Door/safe/drawer events as detection inputs
C Commercial	Cost-update events at landed unit cost	n/a	Buyer labor allocation	Cold-chain → reason-code shrink
D Distribution	Receipt + transfer + RTV at cost	n/a	DC labor allocation	Shelf-state events for OOS detection
F Finance	RIM cost-complement / Cost Method posting	Marketing-spend-event posting + attribution	All overhead accruals + GL roll-up	n/a
J Forecast	Demand signal for replenishment	Customer-frequency forecast feeds CAC payback	n/a	Foot-traffic forecast inputs
S Space/Range	Planogram cost basis at slot precision	Conversion attribution per zone	Fixture allocation	Dwell-and-flow per zone
P Pricing/Promo	Price/markdown events that revalue ledger	Promo-attributed acquisition	n/a	Dynamic price reaction to IoT signal
L Labor	n/a	n/a	Time-clock events at minute precision	Workforce IoT (badge, location)
W Work Exec	Cross-domain exception cost attribution	n/a	Investigation + remediation labor	Cross-IoT exception correlation

Every cell in this matrix becomes a real product capability the VSM can expose. None require new modules — they are projections of the existing spine when read against the satoshi-precision principle.

What this means for product, sales, and engineering

Product. Every new feature decision asks: does this preserve satoshi precision end to end? If a roll-up is added that hides the underlying event chain, it does not ship. The VSM must always be able to drill from the answer back to the events.

Sales. The pitch becomes specific: we are the only retail platform that knows what every cent of every cost actually was, where it went, and what it produced. This is not “we have a dashboard” — it is “you will know things about your business no SMB merchant has ever known.” The Clarks-style diagnostic (canary-retail-diagnostic-archetype) becomes the demo: prize-sized opportunities with sub-cent audit trail.

Engineering. The substrate for this is mostly already in place (stock-ledger, satoshi-cost-accounting, perpetual-vs-period-boundary). The remaining engineering work is: (a) extend the ledger verb taxonomy to include CAC-attribution events and IoT events; (b) build the projection layer that rolls perpetual events into the merchant-readable P&L, CAC report, contribution-margin report; (c) integrate the first IoT sources (door counters, shelf RFID, cold-chain sensors) as Phase-1 parallel observers under the staged-migration pattern.

Open questions

1. Attribution model defaults. What’s the default attribution model for CAC at install? First-touch, last-touch, multi-touch with linear weighting, or merchant-configurable? Defaults matter — most merchants will never change them.

2. IoT vendor strategy. Which IoT verticals does Canary integrate with first? Foot traffic (Density, RetailNext), shelf RFID (Impinj, Avery Dennison), cold-chain (Verkada, monnit) — same integrated-hybrid route question as accounting tools. Pick a starter vendor per category for v3 work.

3. SG&A allocation policy. Per-merchant allocation rules for rent, utilities, software costs are not standardized. Default to the simplest defensible split (per-square-foot, per-device, per-day) and surface as merchant-overridable.

4. Per-customer LTV computation cadence. Real-time, daily, on-demand? Affects compute cost. Real-time is the goal but may be a v3+ capability.

5. Privacy posture for in-store positioning. Foot-traffic and dwell-and-flow capture customer signals that have privacy implications in some jurisdictions. Posture: aggregate-only by default; per-individual tracking opt-in per merchant per jurisdiction.

Related

• Stock Ledger — the perpetual-inventory movement ledger — the substrate this principle is read against
• Satoshi-Level Cost Accounting — the COGS-side of the precision principle (this article generalizes it)
• Retail Accounting Method — RIM, Cost Method, OTB — the period-summary side that satoshi-precision flows into
• Perpetual-vs-Period Boundary — the staged migration that makes parallel-then-cutover possible
• The Canary Retail Spine — 13 Modules — every module is an instrument of cost decomposition under this principle
• GrowDirect Viewpoint — VSM on a Perpetual Ledger
• The Goose — Canary’s GasMeter / L402 satoshi-metering implementation; the production proof point already shipping
• Canary Retail Diagnostic — Archetype — what the satoshi-precision rollup looks like for a real merchant
• Module Manifest Schema — every module’s contribution to the precision principle is declared in its manifest

Sources

• Goose / GRO-117 — production GasMeter shipping satoshi-precision per-tool metering
• L402 protocol — Lightning auth + payment substrate for sub-cent metered access
• Satoshi-cost-accounting article — COGS-side foundation
• Retek RMS perpetual-inventory ledger pattern — the substrate the principle is built on
• IBM BCS 2006 Clarks Retail Diagnostic — the “where every cent went” diagnostic frame this principle delivers on
• IoT instrumentation industry references — Density, RetailNext, Impinj, Verkada, monnit (vendor research deferred to v3 work)