# Deck 21 · Strategic Deep Dives

**Status:** Draft (not yet built)
**Saved:** 2026-06-28 by Jeff (verbal)
**Owner:** Plex / OMX Strategy

---

## One-line thesis

**The deep dives we already do informally — long-form, multi-source, multi-mode analyses of a market, customer, supplier, or strategic question — codified as a repeatable internal product.** Briefing Room covers the standing brief; QBR covers the recurring customer review; this deck covers the deep, one-off strategic investigation.

## The wedge — why now

- The deep dives we've already done for strategy (segment analyses, competitive landscape teardowns, M&A targets, market sizing for new categories) have all been bespoke
- Each takes 1-4 weeks; each is high-impact; each draws on the same Briefing Room engine + Plex-CI + Snowflake + IBP data
- Codifying turns the deep dive from "Jeff's analyst time" to "anyone in ExCo can commission one"
- The output is a long-form artefact (deck or report) — different shape from Briefing Room's brief

## What this deck covers

1. **The deep-dive shape** — scoping question → research plan → multi-source synthesis → strategic recommendation
2. **Reuse of the Briefing Room engine (Deck 18)** — 5-mode underneath, but at much greater depth and synthesis
3. **Long-form output** — strategic narrative + supporting data + boardable recommendation
4. **Standing deep-dive types** — categorisation of the recurring shapes:
   - Customer deep dive (top-tier accounts beyond QBR — strategic relationship review)
   - Competitor deep dive (Plex-CI plus narrative)
   - Category deep dive (entry, expansion, exit)
   - M&A target deep dive
   - Geography deep dive
   - Trend / signal deep dive (specific weak signal → "is this a thing")
5. **Cadence + commissioning** — how ExCo asks for one; how the Strategy team scopes; how it's delivered
6. **Library of past deep dives** — every dive lives in the knowledge base; can be re-read, refreshed, extended

## What this deck explicitly does NOT do

- Not the standing brief (Deck 18 — Briefing Room — runs short-form 5-mode)
- Not the routine QBR (Deck 18 — QBR Pack)
- Not real-time data (Deck 17 — AI Semantic Conversations covers ad-hoc questions)

---

## Problem framing (what's broken)

- **Each deep dive starts from zero** — same engine, but rebuilt every time
- **Strategy team capacity is bounded** — no scaling beyond Jeff + analysts
- **Past work decays** — last year's customer deep dive is in someone's downloads folder
- **No commissioning protocol** — ExCo asks ad-hoc; scope creeps; deliverables don't match the question
- **No "refresh" pattern** — a deep dive done 6 months ago might be 80% still valid, but it's not surfaced

## Benefits (the value story)

| Lever | Mechanism | Sizing approach |
|---|---|---|
| **Strategy capacity** | Codified vs artisan | 2-3x more deep dives per year with same team |
| **Decision quality** | Faster, sharper, more grounded ExCo decisions | Strategic |
| **Reuse compounding** | Library + refresh pattern | Year-over-year leverage on past work |
| **Cross-deck synergy** | Deep dives feed Decks 14 (IBP), 15 (Marketplace), Plex-CI signals back, etc. | Strategic |
| **Knowledge retention** | Every dive becomes searchable corporate memory | Hard to quantify, very valuable |

---

## Layout candidates from the gold standard

- **Cover** — strategy lead at desk with deep-dive deck open; multiple browser tabs reflecting the multi-source nature
- **Problem vector grid (4)**: Always-from-zero / Capacity-bound / Past-work-decays / No-commissioning-protocol
- **Standing deep-dive types** — 6-tile grid (Customer / Competitor / Category / M&A / Geography / Trend)
- **The dive workflow** — scope → plan → research → synthesis → recommendation
- **Library visualisation** — past dives, by type, with refresh status (current / aging / stale)
- **Commissioning protocol** — ExCo ask → Strategy scopes → committed deliverable → review cadence
- **Architecture pipeline** — Briefing Room engine + Plex-CI + Snowflake + external feeds + AI synthesis → deliverable
- **Roadmap** — codify shape + first 3 dives (90d) → library + refresh pattern (6mo) → ExCo self-commissioning (12mo)
- **The ask** — Strategy team capacity envelope + AI tooling + library infrastructure

---

## Open questions to resolve

1. **Past deep dives — count + locations** — what have we done? Where are they? (Likely in `00-INBOX/` historical folders + Plex internal drives)
2. **Canonical deep-dive types** — confirm the 6-shape list; add/remove
3. **Output format** — PPT, PDF, Lens dashboard, all three?
4. **Commissioning gate** — anyone in ExCo, or specific roles?
5. **Refresh trigger** — time-based (6/12 month review), signal-based (Plex-CI material change), or both?
6. **Library tech** — Confluence, Sharepoint, custom on Snowflake?
7. **AI synthesis vs human-in-loop** — first draft AI, human-final? Or AI structures, human writes?
8. **Connection to Deck 18 Briefing Room** — engine reuse; sequencing

## Audience

**Primary:** CEO/GM + ExCo + Head of Strategy.
**Secondary:** Strategy analysts (engine maintainers).
**Tertiary:** Sales Directors (when dive supports a customer or sector pursuit).

## Reference

- Memory: **Deck 18 Company Research + QBR** — shares the Briefing Room engine
- Memory: **Deck 17 AI Semantic Conversations** — same Anthropic + Snowflake stack
- Memory: **Strategic Deep Dives we did for the strategy work** (Jeff verbal 2026-06-28) — informal pattern to codify
- Memory: **Plex-CI** — competitor + market signals as inputs
- Memory: **Briefing Room owns Company Research; 5 modes = one brief, not five** (`reference_omx_project_umbrellas`)
- External references: McKinsey deep-dive shape, BCG case-narrative pattern, Bain situation-complication-resolution-question

---

## Research deepening (background-agent, 2026-06-28)

### Canonical strategic-consulting deep-dive shapes

| Firm | Shape | Anatomy | When it's used | Source |
|---|---|---|---|---|
| **McKinsey** | **MECE + Pyramid Principle (Minto)** | Top: governing thought / answer-first. Then 3 supporting arguments, each MECE (Mutually Exclusive, Collectively Exhaustive). Each argument = 3 sub-supports. SCQA: Situation → Complication → Question → Answer. | Default for any strategic recommendation deck | Barbara Minto, *The Pyramid Principle* (1987); https://www.mckinsey.com/featured-insights ; https://www.barbaraminto.com/ |
| **McKinsey** | **MECE Issue Tree** | Decompose ambiguous question into hypothesis-testable sub-questions; assign owner + data source to each leaf | Scoping phase of a deep dive | McKinsey *Problem-Solving* internal training (public via Ethan Rasiel, *The McKinsey Way*) |
| **McKinsey** | **7-S Framework** | Strategy / Structure / Systems / Shared values / Style / Staff / Skills | Org-strategy deep dives | https://www.mckinsey.com/business-functions/strategy-and-corporate-finance/our-insights/enduring-ideas-the-7-s-framework |
| **McKinsey** | **Three Horizons** | H1 defend core / H2 emerging opportunities / H3 viable options | Growth strategy deep dive | Baghai/Coley/White, *The Alchemy of Growth* (1999); https://www.mckinsey.com/featured-insights/the-next-normal/three-horizons |
| **BCG** | **Growth-Share Matrix** | Stars / Cash Cows / Question Marks / Dogs by market share x growth | Portfolio deep dive | https://www.bcg.com/about/overview/our-history/growth-share-matrix |
| **BCG** | **Experience Curve / Strategic Sector** | Cost-position vs accumulated experience | Cost-position deep dive | https://www.bcg.com/publications/2013/growth-business-unit-strategy-experience-curve-bcg-classics-revisited |
| **BCG** | **Smart Simplicity (six rules)** | Org complexity diagnosis | Org deep dive | Yves Morieux, *Six Simple Rules* (2014) |
| **Bain** | **RAPID decision rights** | Recommend / Agree / Perform / Input / Decide | Decision-quality deep dive | https://www.bain.com/insights/rapid-decision-making/ |
| **Bain** | **Founder's Mentality / Net Promoter** | Internal compass + customer voice | Transformation deep dive | Chris Zook + Fred Reichheld; https://www.bain.com/founders-mentality/ |
| **Bain** | **Repeatable Models** | Identify what scales; cut what doesn't | Strategy deep dive | Zook + Allen, *Repeatable Models* (2012) |
| **Porter (HBS)** | **Five Forces** | Buyer / Supplier / Substitute / Entrant / Rivalry | Sector/category deep dive | https://hbr.org/1979/03/how-competitive-forces-shape-strategy |
| **Porter** | **Value Chain** | Inbound logistics → operations → outbound → marketing → service + support activities | Operating-model deep dive | Porter, *Competitive Advantage* (1985) |
| **Christensen (HBS)** | **Jobs-to-be-Done + Disruption** | Customer hires product for a job; low-end vs new-market disruption | Category-entry / threat deep dive | https://hbr.org/2003/01/the-innovators-solution-discovering-and-developing-disruptive-products |
| **Rumelt** | **Kernel of good strategy** | Diagnosis → Guiding Policy → Coherent Action | Strategy QA pattern | Richard Rumelt, *Good Strategy / Bad Strategy* (2011) |
| **Wardley** | **Wardley Mapping** | Value chain plotted against evolution (genesis → commodity) | Technology / capability deep dive | https://medium.com/wardleymaps |
| **Blue Ocean (INSEAD)** | **Strategy Canvas + Four Actions** | Eliminate / Reduce / Raise / Create | Repositioning / new-category deep dive | https://www.blueoceanstrategy.com |

**OMX canonical six deep-dive shapes (proposed mapping):**

| Deep-dive type | Recommended core framework | Supporting frameworks | Output length |
|---|---|---|---|
| **Customer** | Bain Founder's-Mentality lens + Five Forces from customer's POV | RAPID for buyer decision-rights; share-of-wallet model | 20-40 slides |
| **Competitor** | BCG Experience Curve + Porter Five Forces | Wardley map if tech-led | 25-40 slides |
| **Category** | McKinsey Three Horizons + Blue Ocean Strategy Canvas | TAM/SAM/SOM; Christensen JTBD | 30-50 slides |
| **M&A target** | Bain Repeatable-Models + Synergy decomposition (cost / revenue / capability) | DD checklist; Rumelt kernel test | 40-80 slides |
| **Geography** | McKinsey 7-S applied to region + Porter cluster diamond | Market entry mode matrix (Greenfield / JV / Acq) | 25-40 slides |
| **Trend / signal** | Wardley map + Christensen disruption test | Rumelt kernel; weak-signal triangulation | 15-25 slides |

### Deep-dive process gates (industry consensus)

1. **Commission** — written brief from sponsor (one page, governing thought + 3 sub-questions, success criteria, deliverable shape, decision being made)
2. **Scope + hypothesis tree** — issue tree (MECE) with named hypothesis at each leaf; data sources identified per leaf; review gate with sponsor
3. **Research / data** — 5-mode Briefing Room engine + Plex-CI + Snowflake + IBP + interviews; ~50-60% of total time
4. **Synthesis** — pyramid up from data to governing thought; "so-what" test at each layer
5. **Recommendation** — Rumelt kernel: Diagnosis / Guiding Policy / Coherent Action
6. **Story** — Minto pyramid; SCQA opening; answer-first
7. **Review** — peer challenge (red team); sponsor pre-read 48hrs before
8. **Deliver** — boardable deck + memo + supporting data appendix
9. **Refresh trigger** — re-test diagnosis quarterly via signal monitoring (Plex-CI + news)

### Library / refresh-pattern tech options

| Tech | Fit | Cost |
|---|---|---|
| **Confluence + Atlassian Intelligence** | Familiar; weak semantic search | USD ~6-11/user/mo |
| **Notion + Notion AI** | Modern UX; good linking; weak governance | USD ~10/user/mo |
| **SharePoint + Copilot for M365** | Already in OMX tenant; semantic search via Copilot | USD ~30/user/mo (Copilot licence) |
| **Custom on Snowflake + Anthropic** | Sits inside Briefing Room engine; full reuse | Build cost ~USD 40-80k; runs on existing infra |

OMX-fit: SharePoint as system-of-record (governance + permissions) + Snowflake-backed semantic search + Anthropic synthesis for the "refresh this dive" workflow.

### Sources cited (PR-013)

- https://www.barbaraminto.com/
- https://www.mckinsey.com/business-functions/strategy-and-corporate-finance/our-insights/enduring-ideas-the-7-s-framework
- https://www.mckinsey.com/featured-insights/the-next-normal/three-horizons
- https://www.bcg.com/about/overview/our-history/growth-share-matrix
- https://www.bcg.com/publications/2013/growth-business-unit-strategy-experience-curve-bcg-classics-revisited
- https://www.bain.com/insights/rapid-decision-making/
- https://www.bain.com/founders-mentality/
- https://hbr.org/1979/03/how-competitive-forces-shape-strategy
- https://hbr.org/2003/01/the-innovators-solution-discovering-and-developing-disruptive-products
- https://www.blueoceanstrategy.com
- https://medium.com/wardleymaps
- Books: Rumelt, *Good Strategy/Bad Strategy* (2011); Zook & Allen, *Repeatable Models* (2012); Baghai/Coley/White, *The Alchemy of Growth* (1999); Porter, *Competitive Advantage* (1985); Minto, *The Pyramid Principle* (1987); Rasiel, *The McKinsey Way* (1999); Morieux, *Six Simple Rules* (2014)

---

## Vectors + visuals

### Lucide icon choices (Ask Max set compatible)
- **Customer dive:** `users-round` / `handshake`
- **Competitor dive:** `swords` / `target`
- **Category dive:** `layers-3` / `grid-3x3`
- **M&A dive:** `git-merge` / `combine`
- **Geography dive:** `map` / `compass`
- **Trend / signal dive:** `radar` / `trending-up-down` / `eye`
- **Scoping / brief:** `clipboard-list` / `file-question`
- **Issue tree (MECE):** `git-branch` / `network`
- **Synthesis / pyramid:** `triangle` / `pyramid` (or `mountain`)
- **Recommendation:** `flag-triangle-right` / `lightbulb`
- **Library / refresh:** `archive-restore` / `refresh-cw` / `clock-arrow-up`
- **Commissioning gate:** `shield-check` / `check-circle-2`

### Image concepts (NZ context, 4-6)
1. **Cover** — Strategy lead at desk, large monitor showing a finished deep-dive deck title page ("Category Deep Dive: Hybrid-Work Furniture, NZ — H2 2026"); multiple browser tabs reflecting Stats NZ, MBIE, IBISWorld, Plex-CI; Auckland through window, ceramic mug with NZ-design ferns.
2. **Six-shape grid** — 2x3 tile grid for the six standing deep-dive types; each tile = Lucide icon + name + one-line definition + "engine reuse: Briefing Room 5-mode". Soft palette per OMX exec brief.
3. **The dive workflow ribbon** — left-to-right ribbon: Commission → Scope (MECE tree) → Research → Synthesis → Recommendation → Story → Review → Deliver → Refresh. Each step icon-led; gate-review markers between steps.
4. **Library visualisation** — bookshelf metaphor with past dives as books; spine colour = type; faded spines = aging/stale; bright spines = current. Last-refreshed dates on a few key NZ-relevant ones ("BTS School Channel — refreshed Apr 2026", "N3 Buying Group — current").
5. **Pyramid + SCQA poster** — Minto pyramid with OMX-flavoured example governing thought ("OMX should enter $X NZ category via Y route because Z"); 3 supporting points; SCQA strip across top.
6. **Commissioning brief one-pager** — mock A4 of an ExCo commissioning brief: sponsor, question, sub-questions, success criteria, decision-to-be-made, deliverable shape, due date. Reinforces "no ad-hoc; every dive starts with a one-pager."