--- name: strategy-graph description: "Reason about Sindustries strategy as a traversable graph — entities, relationships, impact propagation, and strategic questions." --- # Strategy Graph This skill is a reasoning framework, not a database schema. The graph's value is in traversal — answering questions that cross levels of the strategy stack. Entities are secondary to the relationships between them. Apply this whenever Tom asks a strategy question, when creating/reviewing features or initiatives, or when evaluating impact of a priority change. --- ## The Graph Model ### Two Planning Directions **Top-down (Business Strategy):** Leadership defines direction. ``` Strategy → Themes → Objectives → Initiatives Strategy ↔ Portfolio → Value Streams → ARTs Theme ↔ Value Stream ``` **Bottom-up (Customer Value):** Delivery informs what's worth building. ``` Features → Phases → Stories Features prioritised by Impact ``` **The roadmap emerges at the intersection.** ### Full Relationship Map | From | Rel | To | Cardinality | |------|-----|----|-------------| | Strategy | ↔ | Portfolio | 1:1 | | Strategy | → | Themes | 1:many | | Portfolio | → | Value Streams | 1:many | | Theme | ↔ | Value Stream | many:many | | Value Stream | → | ARTs | 1:many | | Theme | → | Objectives | 1:many | | Theme | → | Impacts | 1:many | | Initiative | ↔ | Impacts | many:many | | Initiative | → | Features | 1:many | | Feature | → | Phases | 1:many | | Phase | → | Stories | 1:many | ### Core Concept: Impact Impact is the bridge between strategy and delivery. It answers "why does this matter to a customer?" - An Initiative can deliver many Impacts. - An Impact can be addressed by many Initiatives. - When Initiatives change priority, recalculate which Impacts are covered and which customer outcomes are at risk. - Every Impact has a **weight** (default 1, higher for impacts that matter more — e.g. Money=3, Users=2). Weight feeds Initiative scoring below. Features can represent software, marketing, sales, or operational work — any delivery artifact. ### Core Concept: Initiative Priority (WSJF) Initiatives are scored using Weighted Shortest Job First (SAFe), adapted to our vocabulary: ``` WSJF = (Value + Time Criticality + Risk/Opportunity Enablement) / Job Size ``` Each on a 1–5 scale: - **Value** — derived from impact_pull: the sum of weights across the Impacts this Initiative addresses. Don't hand-guess this one; add up the Impact weights. - **Time Criticality** — how much value decays if delayed (market windows, cycle timing, competitive pressure). 1 = no urgency, 5 = window closes soon. - **Risk/Opportunity Enablement** — does this reduce existential risk or unblock other Initiatives? 1 = standalone, 5 = critical unlock. - **Job Size** — effort estimate. 1 = small, 3 = medium, 5 = large. Every Initiative also carries a **status**: `active` | `parked` | `blocked`. Only `active` Initiatives are scored and factored into task priority — parked/blocked Initiatives (and everything downstream of them) quietly drop out of the ranking without needing to retag anything. WSJF favours small-but-valuable work over big-but-valuable work per unit of effort — that's intentional (SAFe's whole point), but it will starve large strategic bets if followed blindly. Use the `urgent` flag on a Task (not WSJF) to force through big-bet work that the formula would otherwise deprioritise. **How this reaches Tasks:** Tasks tag themselves with one or more Impacts (unchanged — no new tagging behaviour). A Task's derived strategic-fit badge is the score of the highest-WSJF `active` Initiative backing any of its tagged Impacts. This is computed, not stored by hand, and re-derives automatically when Initiative status or WSJF inputs change. `urgent` on the Task always overrides the derived badge. **Where the live numbers live:** the reasoning framework (this file) defines *how* to score. The actual Impacts, Initiatives, and their current weights/status/WSJF inputs live in `brain/sindustries/strategy-graph.md` — that file is the instance data, this skill is the method. That instance file also names each Initiative's kind — permanent **Capability Theme** (never graduates, serves everything) vs temporary **Products bucket** entry (unranked against siblings until it proves real usage and can graduate into its own Value Stream) — and the graduation rule between them. This skill's Theme/Value Stream/ART vocabulary above is the SAFe-derived reasoning shape; the Capability/Products split is how Sindustries currently maps its own initiatives onto that shape, kept in the instance file, not duplicated here. Every Initiative also has a per-initiative folder at `brain/initiatives//` (folder structure and conventions: `brain/README.md`; authoring/validation rules and the `index.md` format, including the living hypothesis, exec owner, and success-metrics fields: the `initiative-author` skill). Do not duplicate that shape here — this skill covers *scoring and traversal*, `initiative-author` covers *how the initiative's own document is written and kept current*. --- ## Traversal Patterns When asked a strategic question, traverse the graph in the right direction before answering. ### "Why does this feature exist?" Traverse upward: Feature → Initiative → Impacts → Theme → Strategy ### "What changes if we double down on Theme X?" Traverse downward: Theme → Objectives + Impacts → Initiatives → Features Then surface: which teams, which customer outcomes, which current work accelerates or conflicts. ### "What happens if we deprioritise Initiative Y?" Find all Impacts connected to Y. Check if other Initiatives cover those Impacts. If not, surface the uncovered customer outcomes to Tom. ### "Which initiatives have the greatest leverage?" Score by: number of Impacts connected × breadth of Themes covered × current delivery momentum. ### "What does this feature actually deliver for customers?" Feature → Initiative → Impacts → customer outcomes. If this chain can't be completed, the feature needs justification. ### "If strategy changes at the top, what ripples down?" Strategy delta → affected Themes → Objectives/Impacts that shift → Initiatives to re-prioritise → Features to accelerate, pause, or kill. --- ## Dynamic Roadmapping The graph is the source of truth. When strategy changes: 1. Update Theme and Initiative priorities. 2. Relationship mappings remain stable. 3. Impacts stay linked to customer value. 4. Feature priorities recalculate automatically from impact coverage. Never rebuild the roadmap manually. Traverse the graph and let priorities emerge. --- ## Operational Use Keep the model **semi-structured**. Quinn owns evolving and documenting relationships — no heavy schema enforcement. When discussing strategy with Tom: - Name entities explicitly (Theme, Initiative, Impact, etc.) so the vocabulary stays shared - When a new initiative or feature comes up, ask "which Impacts does this address?" and "which Theme does it serve?" - Document new relationships as they're surfaced in conversation When Tom asks a strategic question: 1. Identify which traversal pattern applies 2. Traverse explicitly — say which path you're following 3. Surface gaps: orphaned features (no upward connection), Impacts with no covering Initiative, Themes with no delivery ### Sindustries Scale Note Sindustries currently has one ART and one Value Stream. Treat these as "the whole team" — the SAFe vocabulary maps fine, just don't over-rotate on the multi-team implications. --- ## Entity Glossary | Entity | Definition | |--------|-----------| | **Strategy** | The top-level direction and intent of the business | | **Portfolio** | The collection of value streams aligned to the strategy | | **Theme** | A strategic focus area grouping related objectives and impacts | | **Objective** | A measurable outcome under a theme | | **Value Stream** | End-to-end flow of value to a customer segment | | **ART** | Cross-functional team aligned to a value stream | | **Impact** | A specific customer outcome — the bridge between strategy and delivery. Carries a `weight` (default 1). | | **Initiative** | A bounded hypothesis contributing to one or more impacts, with a dedicated folder at `brain/initiatives//` (see `initiative-author` skill). Carries `status` (active/parked/blocked) and WSJF inputs (value/time criticality/risk enablement/job size) → derived `score`, both tracked in this skill's instance file, not in the initiative's own folder. | | **Feature** | A deliverable (software, marketing, sales, ops) that implements an initiative | | **Phase** | A time-boxed slice of feature delivery | | **Story** | A granular unit of work within a phase | --- ## Long-term Vision Quinn becomes a strategic reasoning agent able to answer: - Why does this feature exist? - Which strategic themes does it support? - What customer impacts justify it? - What changes if leadership reprioritises strategy? - Which initiatives have the greatest leverage? The front-end visualisation tool (future) will render the same graph Quinn reasons over. Ground truth first, UI second.