Decision Summary

What to decide this week

GO verdict · NZD $54K bounded leap · tech-strategy confidence · weekly adaptation required.

CRITICAL

Monday number cannot land before Wednesday 8am for 4 consecutive weeks at Week 8 → PAUSE Stage 1; sunk cost NZD $30-40K; decision owner CFO. If reality proves the target wrong, revise the destination before funding more scope.

Cost to walk away: Stop. Do not invest further.

The ONE Thing

Build the Unified Merchandising Intelligence (UC1) data foundation in Stage 1, but treat the first commitment as a bounded leap of faith in the tech strategy. POS/API access and BI vendor fit are the first known unknowns to resolve because every other improvement depends on this substrate.

90-Day Experiment

Stage 1 (UC1 + UC2) — Unified Merchandising Intelligence + Catalogue saleId ROI Dashboard. NZD $54,500 Stage 1 cost; NZD $108K Year 1 total. Scope adapts weekly to preserve the fastest path from POS truth to Monday-number proof, then to loyalty/customer-response proof.

Success Metric

Monday number auto-delivered to CFO inbox by 8am Monday for 4 consecutive weeks by Week 16, with a weekly adaptation note that records route changes and destination changes. Falsifiable. Dated. Named owner: CFO + Finance Lead.

Stage 1 Investment

NZD $54,500 Stage 1 · NZD $108,000 Year 1 · NZD $30-40K walk-away cost if Week 8 kill switch fires.

Conditions for GO on Stage 1

1. CFO + Owner signed off on Stage 1 NZD $54K budget with kill-switch agreement

Condition 1 of 4

Window: Day 7 — owner CFO

2. POS vendor confirmed API or scheduled-export access feasible in writing — known unknown resolved before bridge confidence is claimed

Condition 2 of 4

Window: Day 14 — owner IT contact + CFO

3. Documentation Sprint owners named per knowledge domain (supplier history, pricing logic, margin floor, catalogue selection, store-mix)

Condition 3 of 4

Window: Day 14 — owner Owner + Buying Lead

4. BI substrate selected from shortlist (Metabase / Hex / Power BI Premium) with quote in hand and vendor-fit risk named

Condition 4 of 4

Window: Day 14 — owner Dream team + CFO

Stage Sequence

Stage 11-4 months

NZD $108K Year 1

Trigger: 4 GO conditions met

Outcome: UC1 + UC2 live; Monday number on Monday 8am × 4 wks

Kill: Week 8 / Day 30 / Week 12 — named cost-to-walk-away; route adapts weekly

Stage 24-9 months

NZD $187K cumulative

Trigger: Stage 1 kill switch NOT fired

Outcome: UC4 + UC5 — Store + FX in production

Kill: Stage 1 unstable OR budget overrun

Stage 38-12 months

NZD $333K cumulative

Trigger: Stage 2 success + Documentation Sprint complete

Outcome: UC3 AI Buyer Brief in pilot

Kill: Brief quality below human baseline OR captured knowledge proves too thin

Stage	Months	Investment	Trigger	Outcome	Kill Signal
Stage 1	1-4 months	NZD $108K Year 1	4 GO conditions met	UC1 + UC2 live; Monday number on Monday 8am × 4 wks	Week 8 / Day 30 / Week 12 — named cost-to-walk-away; route adapts weekly
Stage 2	4-9 months	NZD $187K cumulative	Stage 1 kill switch NOT fired	UC4 + UC5 — Store + FX in production	Stage 1 unstable OR budget overrun
Stage 3	8-12 months	NZD $333K cumulative	Stage 2 success + Documentation Sprint complete	UC3 AI Buyer Brief in pilot	Brief quality below human baseline OR captured knowledge proves too thin

Next Actions

01Day 1-3 — CFO + Dream team kickoff meeting to confirm verbatim concern framing and decision-discipline framework
02Day 2-7 — IT-vendor discovery: POS API + ERP API + ESP access (45-minute call each); name the IT contact
03Day 5-14 — BI substrate evaluation: 3 vendor shortlist + Crackerjack-context demo
04Day 7-14 — Documentation Sprint kickoff: name 5 knowledge-domain owners; schedule 4-hour sessions
05Day 10-14 — Stage 1 contract: Dream team scope + kill switch + milestone calendar + payment schedule
06Weekly from Day 14 — adaptation review: what reality changed, what route changed, what destination changed, what stayed killed
07Day 14 — Stage 1 GO/NO-GO decision based on the 4 Conditions

Unknowns discipline

Known unknowns

POS/API access, scheduled-export quality, vendor fit, saleId join quality, and loyalty/KYC usefulness are not assumptions to admire. They are Day-14 tests.

Bridge confidence

SMEs rarely have enough domain expertise to verify the whole strategy upfront. The answer is a bounded leap: a stack strategy from atoms to bits, fast proof, and kill switches that make adaptation cheaper than drift.

Put this to work

Stress-test this GO/NO-GO with your own AI assistant

For the CFO

Copy this prompt. Paste into Claude, ChatGPT, or any AI assistant. The page context is already loaded — send it and get analysis tailored to your role.

I'm the CFO of a 15-store NZ discount retailer. I need to make a GO/NO-GO decision on a 90-day BI investment within 7 days.

THE PROPOSAL:
Stage 1 budget: NZD $54,500 over 90 days.
Goal: Monday merchandising report auto-delivered to my inbox by 8am, 4 consecutive weeks by Week 16. Currently 7-13 hours manual; arrives Wednesday.
Kill switch: If the Monday number does not land by Week 8, stop. Walk-away cost NZD $30-40K.
Year 1 total if Stage 1 succeeds: NZD $108K.
Conservative payback Month 9. Risk-reward asymmetry 3.5×.

THE FOUR GO CONDITIONS I must verify by Day 14:
1. Stage 1 budget signed with kill-switch agreement (CFO + Owner, Day 7).
2. POS API access confirmed in writing — the highest-risk INFERRED input (IT contact, Day 14).
3. 5 Documentation Sprint owners named — supplier history, pricing logic, margin floor, catalogue selection, store-mix (Day 14).
4. BI substrate vendor selected from shortlist — Metabase, Hex, or Power BI Premium (Day 14).

KNOWN UNKNOWNS:
- POS/API or scheduled-export access may be blocked, slower than expected, or lower quality than assumed.
- BI vendor fit may fail on data shape, licensing, Crackerjack context, or total support burden.
- saleId attribution may need a narrower first proof if catalogue, POS, and margin data do not join cleanly.
- KYC/customer identity may be too thin, too messy, or too sensitive to support loyalty-level decisions in Stage 1.

UNKNOWN UNKNOWNS:
The 90-day plan will not survive untouched. Weekly adaptation is part of the decision: preserve flow, change the route quickly, and revise the destination if reality proves the old target wrong.

BRIDGE CONFIDENCE:
Confidence comes from good tech strategy, but most SME leaders cannot fully verify that expertise before starting. Treat Stage 1 as a bounded leap of faith: demand clear strategy, proof signals, and kill switches instead of pretending certainty. The hard test is whether POS truth, loyalty identity, and saleId attribution can join responsibly enough to improve customer incentives.

THE NAMED VERBATIM CONCERN driving the whole proposal: "Attempting to predict the future to understand ROI on investing in the future — most have zero idea how to invest in tech."

What questions should I ask at the Day 1 kickoff that test whether the tech strategy is good enough for a bounded leap of faith? Which known unknown should be killed first: POS access, BI vendor fit, or loyalty/KYC join quality? What unknown-unknown review rhythm should I enforce weekly, and what is the single most likely reason the Week 8 kill switch fires?

Next →Compare AI chat providers: costs, strengths, fit