Skip to main content

Finance Performance

Is the loop working — or pretending to?

The Gauge​

Performance pairs every "what good looks like" with a matching "what bad looks like." A target without a paired failure mode is optimism. A failure mode without a target is fear. Both together is a gauge.

Every gauge below names a threshold, a warning, a bad state, and a decision flow that ends in CFO action. The CFO owns every alert. The agent or report-up role acts on it.

Paired Gauges​

G1. Close cycle​

  • Good: Books closed in ≤ 3 business days from period end. Real-time close in flight.
  • Warning: Day 4 with reconciliation still open.
  • Bad: Day 7+ with material unreconciled balances.
  • AI-augmentation: close-agent runs anomaly detection against ledger as soon as last journal posts.
  • Crypto-instrument: on-chain reconciliation reduces wallet-to-ledger lag to near-zero.

G2. Truth drift​

  • Good: Variance ≤ 0.5% of cash position between bank/custody/wallet sum and ledger cash.
  • Warning: Variance 0.5–1.5% at any daily check.
  • Bad: Variance > 1.5% at period-end close.
  • AI-augmentation: reconciliation-agent flags every drift > 0.25% within the trading day.
  • Crypto-instrument: chain-as-audit-trail makes wallet drift detectable in seconds.

G3. Runway​

  • Good: Total runway ≥ 18 months; runway-in-stables ≥ 6 months independent of bank rails.
  • Warning: Total runway 12–18 months OR stables runway < 6 months.
  • Bad: Total runway < 12 months OR stables runway < 3 months.
  • AI-augmentation: FP&A-agent reforecasts runway nightly with current burn and pipeline.
  • Crypto-instrument: stables held in segregated custody decouple runway from bank-failure risk.

G4. Mark drift​

  • Good: Non-stable token exposure marked daily; drift band ≤ 2% of total treasury per asset.
  • Warning: Drift 2–5% on any tracked asset.
  • Bad: Drift > 5% without hedging or sell decision logged.
  • AI-augmentation: treasury-agent pulls oracle prices and alerts on drift band breach.
  • Crypto-instrument: on-chain oracles deliver mark-to-market in real time; no end-of-day batch.

G5. DSO​

  • Good: DSO ≤ 35 days.
  • Warning: DSO 35–50 days OR concentration > 30% in top customer.
  • Bad: DSO > 50 days OR one customer > 50% of AR.
  • AI-augmentation: AR-agent sequences dunning and proposes payment-rail offers.
  • Crypto-instrument: stablecoin invoicing settles in minutes, collapsing DSO for crypto-native counterparties.

G6. AP discipline​

  • Good: Early-payment discounts captured ≥ 90% when offered; late-payment rate ≤ 2%.
  • Warning: Discount capture 70–90% OR late rate 2–5%.
  • Bad: Discount capture < 70% OR late rate > 5%.
  • AI-augmentation: AP-agent batches and sequences payments to hit discount cutoffs.
  • Crypto-instrument: on-chain payment rails reduce settlement latency to one block.

G7. Agent overrides​

  • Good: CFO or named human overrides agent decisions on ≤ 5% of completed actions.
  • Warning: Override rate 5–15% — calibration drift.
  • Bad: Override rate > 15% — agent is mis-tuned and burning trust.
  • AI-augmentation: receipt log captures every override with the reason; FP&A-agent retunes monthly.
  • Crypto-instrument: on-chain receipts make overrides immutable and auditable.

G8. Tier discipline​

  • Good: Allocation within ±5% of target by tier; experiments funded from protected budget.
  • Warning: Drift 5–10% in any tier OR experiments funded from core.
  • Bad: Drift > 10% OR three consecutive months of tier breach.
  • AI-augmentation: capital-allocation-agent tags every disbursement to tier and alerts on drift.
  • Crypto-instrument: smart-contract escrow can lock experimental capital and release on milestone — protection becomes mechanical.

G9. Audit-trail completeness​

  • Good: 100% of state-changing transactions carry timestamp + actor + evidence link.
  • Warning: < 99% completeness on monthly sample.
  • Bad: < 95% OR any material transaction missing evidence.
  • AI-augmentation: audit-agent samples nightly and flags gaps before close.
  • Crypto-instrument: chain-as-audit-trail produces immutable evidence by default for any on-chain leg.

Decision Flows​

Every "bad" or "warning" trigger above routes to a named decision flow. Each flow ends in an action, not a review.

  • G1 close cycle slips → controller escalates to CFO by day 4 → CFO decides hold or release with caveat → board narrative adjusted same day.
  • G2 drift > 1.5% → controller pauses period close → CFO opens forensic ticket with audit-agent → resolution logged within 48 hours or audit committee notified.
  • G3 runway warning → CFO triggers raise-prep workflow OR cost-cut workflow within five business days → board notified within ten.
  • G4 drift > 5% → treasurer presents hedge or sell plan to CFO within 24 hours → CFO authorizes → executed by treasury-agent within the trading day.
  • G5 DSO bad → AR lead presents top-five overdue plan → CFO authorizes write-off, collection escalation, or rail-switch offer → action booked within five business days.
  • G6 discount capture bad → AP lead surfaces process or cash-constraint root cause → CFO decides cash injection or process fix → action logged.
  • G7 override > 15% → CFO commissions agent-tuning sprint with FP&A lead → agent retrained or scope cut → re-baseline within 30 days.
  • G8 tier breach → CFO restores tier discipline at next allocation cycle OR formally re-bases the 70/20/10 split with board sign-off → no silent drift.
  • G9 audit gap → controller patches sample → CFO escalates to audit committee if pattern → external auditor briefed.

Alert Owners​

  • G1 close — owned by controller; runs through the daily close stand-up; escalates to CFO by day 4.
  • G2 drift — owned by controller; surfaced by the reconciliation agent; escalates to CFO within four hours.
  • G3 runway — owned by FP&A lead; weekly CFO sync; CEO and board notified within ten days if unresolved.
  • G4 mark-to-market — owned by treasurer; treasury-agent pings the moment a drift band breaks; CFO decides within twenty-four hours.
  • G5 DSO — owned by AR lead; weekly collections review; CFO acts at end of week.
  • G6 discount capture — owned by AP lead; monthly AP review; CFO acts at month-end.
  • G7 overrides — owned by FP&A lead; monthly agent review; CFO and agent owner act together.
  • G8 tier discipline — owned by FP&A lead; quarterly capital-allocation review; CFO escalates to board on breach.
  • G9 audit-trail — owned by controller; pre-close audit sweep; CFO and audit committee respond to gaps.

Warning Signals​

  • Three consecutive months of close slipping by one day, even within the ≤ 3-day target.
  • A single counterparty crossing 25% of AR.
  • Stablecoin redemption window stretching from same-day to T+1 at any custodian.
  • Agent override rate trending up week over week, even below 5%.
  • Allocation tier drift trending up two months in a row, even within ±5%.

The CFO acts on warnings before they become bad. Acting late is the failure.

Context​

  • Principles — What the gauges are measuring against.
  • Process — The workflows these gauges instrument.
  • Positions — Who owns which gauge.
  • Platform — Where the gauges are read from.
  • Scoreboard — How finance gauges roll into the wider business picture.