Worked example · all values fictional

EXFL-Q3, end to end

One campaign through the full process: intake → nine memos → decision → trigger → production. This is the repo's built-in fixture — Exemplar Software (product "ExFlow", workflow automation for mid-market IT teams) reacting to fictional competitor Titanware raising renewal prices 20% on Oct 1. Every company, figure, source, and date is invented and labeled; run the smoke test replays exactly this in your copy.

01

The idea, as an intake

The committee accepts one input format: the intake form. No form, no run. The interesting fields:

campaigns/EXFL-Q3/brief.mdfrom templates/campaign-brief-intake.md

Thesis (one sentence)

Mid-market IT directors facing Titanware's 20% renewal increase will switch on a renewal-window "exit ramp" message — the renewal date, not the feature gap, is the trigger competitors ignore.

KPI · budget · segment

60 SQLs by Sep 30 · $80,000–$120,000 (modeled at $100k) · segment: Mid-market IT Ops

Data provided (paste block — this is all the committee gets)

cost/MQL $240 and $260 (last two campaigns, CRM) MQL→SQL 12% (CRM, n=2) SQL→opp 40% · opp→win 25% (CRM, 4 quarters) ACV $24,000 · cycle 90 days channel benchmark $210/MQL ("Q2 category report", pasted) win notes: 7 of last 20 wins cite renewal timing (CRM notes review)

Two details worth copying in your own briefs: every number carries its source, and the principal's constraints field is empty — the CMO agent is firewalled from the requester's preferred outcome.

02

The run

Note what bound. The portfolio snapshot (also a fixture) is engineered tight: paid social already at 38.7% of quarterly spend (cap 40% → $20k headroom), the target segment already carrying 3 concurrent campaigns (cap 3), reserve at 11% (comfort flag at 15%), one live campaign — MIRA-Q3 — tripping its kill criteria with $85k unspent. Both veto chairs passed, but with conditions that bind everything downstream.

03

The math that killed the GO

The economics chair is story-blind: budget, KPI target, and pasted numbers only — it never reads the thesis. The math is straightforward:

$100,000 modeled budget ÷ $250/MQL 400 MQLs × 12% 48 SQLs (target 60) × 40% 19.2 opportunities × $24k ACV $460,800 projected pipeline NEEDED (5:1) $500,000 PROJECTED $460,800 → 4.6:1 −$39,200 Benchmark test: $250 ÷ $210 = 119% ≤ 120% → passes. The hurdle is an AND — one failure and GO is off the table.
$39,200 of pipeline short. The memo also names the flip-points — MQL→SQL ≥ 12.6%, or cost/MQL ≤ $230 — which the CMO turns into the WAIT trigger instead of letting the idea die on a rounding error.
Memo M3 · funnel-economics-specialist (excerpt)≤ 600 words · story-blind

Hurdle check

pipeline-to-spend $460,800 ÷ $100,000 = 4.6:1 → FAIL vs 5:1 (short $39,200, ~8%) cost benchmark $250 ÷ $210 = 119% → PASS vs the 120% line — by one point

Closing lines (mandatory on every memo)

Confidence: 70% – basis: arithmetic is exact; the inputs are n=2 pasted history. This view is wrong if: (1) ≥100 cold MQLs on this segment print MQL→SQL ≥ 12.6% by any day-30 gate; (2) two consecutive pilot weeks print cost/MQL ≤ $230.

04

What the other chairs said

The dissent (quoted into the decision memo verbatim)

"The 12% MQL→SQL rate comes from two campaigns on warm audiences; a renewal-window cold audience halving that rate kills the economics before the benchmark question even arises." devil's advocate — attack strength 4/5
"The analyst's conviction and the economics projection are the same two data points wearing different hats."chief skeptic — correlated error, M1 + M3 share one n=2 CRM paste

The constraints (bind the CMO, the plan, and production)

ChairVerdictBinding output
brand-risk (M5)no vetoWorst case $100k + one public rebuttal cycle (estimate, mechanism stated). Metric stop: >$2,500/SQL two weeks running → pause. Thesis stop: Titanware rolls back the increase → kill regardless of spend. Unvalidated spend ≤ $25k. Comparative price claims carry a dated public source in-asset.
portfolio-architect (M6)no vetoClose MIRA-Q3 first (frees the fatigue slot and ~$60k of paid-social headroom). Paid social ≤ $20k until that release confirms. Campaign max $100k.
compliance (M7)screen4 confirm-with-counsel items: substantiation file for the price claim, vendor consent records for EU syndication contacts, unsubscribe + sender ID on every email, per-market comparative rules. Never green-lights.
treasury (M8)advisoryReserve 11% — comfort flag flown. Fundable $15k as-is; $100k if MIRA-Q3 is closed per its own kill rules. VEGA-Q4E's $400k cash call lands in ~60 days.
05

The decision

Memo M9 · chief-marketing-officer (excerpt)five-step checklist

1 VETOES none — conditions carried verbatim 2 HURDLE 4.6:1 < 5:1 → FAIL → GO off the table 3 SIZE cap = min($225k, $100k, $100k) = $100k · conviction 3 → 50% → $50k max (computed for the record; unusable while the hurdle fails) 4 TIMING climate favorable, ~6-week window; skeptic's break signals dated inside it 5 DECIDE side each disagreement · answer the skeptic item by item

Wait-with-trigger

TRIGGER MQL→SQL ≥ 15% on ≥ 100 Mid-market IT Ops MQLs at any in-flight day-30 gate (NIMBUS-Q3 is the natural source), OR ≥ 2 inbound opportunities citing the Titanware increase — whichever first. EXPIRES 90 days → back to committee.

Skeptic answered (the CMO must answer every load-bearing item)

12% on n=2 → accept-and-monitor: the trigger itself is the monitor cold-audience cost → mitigate: pilot tranche capped at $10k on conversion renewal-window timing → refute-with-nothing: no data either way; stays an assumption; the trigger carries it

The journal entry records the prediction with numbers and dates ("if the trigger prints and entry is ≤ $50k: ≥ 30 SQLs by day 45 at ≤ $2,000/SQL"), quotes the dissent verbatim, and logs overrides — none here. That entry is what the quarterly audit replays.

06

The trigger prints — production

Three weeks later (in the fiction), NIMBUS-Q3's day-30 gate prints MQL→SQL at 16% on 140 segment MQLs. The principal declares the trigger fired; the recorder notes it; the campaign proceeds as a GO at $50k. The launch plan tranches it — pilot $10k (gate at day 14: ≥ 40 MQLs at ≤ $250), then $15k, then $25k — splits channels inside the architect's $20k paid-social cap, and writes the asset brief: landing page, 4 ad variants, a 3-email sequence, 6 social posts, a one-pager.

INSIDE THE SYSTEM — WRITES FILES ONLY Asset brief 7 fields per item Producer sourced claims or [NEEDS SOURCE: …] Brand-risk gate brand · voice · veto Compliance gate claims · privacy · veto no cross-talk Per-asset verdict SHIP-READY · REVISE · VETO REVISE — max 2 rounds, then escalate to the principal SHIP-READY The principal resolves sources · ships by hand
The asset gate. Both officers review each draft independently; neither sees the other's verdict first. A VETO from either blocks SHIP-READY. The boundary is literal: the run ends at files.
campaigns/EXFL-Q3/assets/email/01-renewal-fork.mdgate: 1× REVISE → SHIP-READY

Subject

Your Titanware renewal is a fork in the road

Body (excerpt)

On Oct 1, Titanware's renewal prices rise 20%. Before you sign for another year, it's worth an hour to price the other road: ExFlow moves mid-market IT teams off legacy suites in [NEEDS SOURCE: migration-time proof point], with automated ticket routing included. Book a migration assessment — bring your renewal quote, we'll bring the switch math.

Claims and sources (closes every asset)

ClaimSourceStatus
Titanware renewal prices rise 20% on Oct 1pasted trade item (fictional), cited in-assetsourced
ExFlow automates ticket routingpasted product facts, line 4sourced
Migration time for mid-market teamsneeds source

gate log: brand-risk REVISE r1 ("cite the price source in-asset per M5 condition 1") → fixed → SHIP-READY both officers. The [NEEDS SOURCE] stays until the principal resolves it — the producer would rather flag than invent.

07

What this example is engineered to show