Skip to content

AI agent skills for startup validation, competitive intelligence, and planning

feat: add positioning frameworks reference (Dunford/Moore/Neumeier/JTBD/Ries)

Ferdinando Bonsegnadefde8f6766d
diff --git a/startup-positioning/references/frameworks.md b/startup-positioning/references/frameworks.md
new file mode 100644
index 0000000..fa5e449
--- /dev/null
+++ b/startup-positioning/references/frameworks.md
@@ -0,0 +1,132 @@
+# Positioning Frameworks Reference
+
+Canonical definitions for the 5 positioning frameworks used in this skill. Apply them in order: Dunford's process first (primary engine), then express the result through Moore and Neumeier formats, grounded by JTBD thinking, and sanity-checked against Ries & Trout principles.
+
+---
+
+## 1. April Dunford — Obviously Awesome
+
+Positioning defines how your product is the best in the world at providing something that a well-defined set of customers cares a lot about.
+
+**The 5 components (work through in this order):**
+
+1. **Competitive Alternatives** — What would customers do if your solution didn't exist? Not just direct competitors — include manual processes, spreadsheets, hiring someone, doing nothing. This grounds the positioning in reality.
+
+2. **Unique Attributes** — What do you have that the alternatives don't? Be specific and factual. Not "better UX" but "drag-and-drop workflow builder that requires zero code." Capabilities, features, expertise, integrations, data, speed, etc.
+
+3. **Value** — What do those attributes enable for the customer? Map each attribute to a customer benefit. Attributes are features; value is the outcome. "Drag-and-drop builder" → "Non-technical teams can build workflows without waiting for engineering."
+
+4. **Target Customer** — Who cares the most about this value? The best customers are those for whom your unique value is critical, not just nice-to-have. Define by characteristics that make them care more than others.
+
+5. **Market Category** — The context you position your product in so the value is obvious. Three options:
+ - **Existing category**: You're a better version of something people already buy (e.g., "CRM for real estate agents")
+ - **New sub-category**: You take an existing category and add a qualifier (e.g., "AI-powered recruiting platform")
+ - **New category**: Rare and expensive to do. Only if nothing else frames your value well.
+
+**Adapted 10-step process for AI-guided positioning:**
+
+1. Understand your best customers — who loves your product and why
+2. Form the positioning team — in our case: the AI + the founder
+3. Align on vocabulary — define terms, abandon preconceptions about current positioning
+4. List your true competitive alternatives — EVERYTHING customers would do instead
+5. Isolate your unique attributes — what you have that alternatives don't
+6. Map attributes to value themes — attribute → "so what" → customer outcome
+7. Determine who cares most — best-fit customer characteristics
+8. Find your best market frame — category that makes value obvious
+9. Layer on a trend — only if genuine, optional
+10. Capture the positioning — multiple formats: Moore statement, Onliness, elevator pitch
+
+**Common pitfalls by component:**
+
+- **Alternatives:** Only listing direct competitors. Miss the status quo — manual processes, spreadsheets, hiring someone, doing nothing. These are often the real competition.
+- **Attributes:** Listing features that alternatives also have. That is table stakes, not differentiation. If two competitors also have it, it is not unique.
+- **Value:** Staying at the feature level instead of translating to outcomes. "AI-powered" is an attribute. "Get answers in seconds instead of hours" is value.
+- **Target:** Too broad ("SMBs") or demographic ("25-34 year olds") instead of characteristic-based. Define by behaviors, situations, and needs that make them care more.
+- **Category:** Inventing a category nobody searches for. If customers cannot Google it, they cannot find you.
+
+---
+
+## 2. Geoffrey Moore — Crossing the Chasm Positioning Statement
+
+**The template:**
+
+> **For** [target customer]
+> **who** [statement of need/opportunity],
+> **the** [product name] **is a** [market category]
+> **that** [statement of key benefit].
+> **Unlike** [primary competitive alternative],
+> **our product** [statement of primary differentiation].
+
+**Field definitions:**
+
+- **Target customer** = Dunford's Best-Fit Customers
+- **Need** = The job to be done (from JTBD analysis)
+- **Category** = Dunford's Market Category
+- **Key benefit** = Dunford's Value Themes (pick the strongest)
+- **Alternative** = Dunford's top Competitive Alternative
+- **Differentiation** = Dunford's Unique Attributes (most defensible)
+
+**Example:**
+
+For mid-market engineering teams who need to track deployment reliability across microservices, DeployWatch is a deployment observability platform that surfaces breaking changes before they reach production. Unlike generic APM tools, DeployWatch correlates deployments with performance regressions in real-time.
+
+---
+
+## 3. Marty Neumeier — Zag (Onliness Statement)
+
+**Basic template:**
+
+> Our [brand/product] is the only [category] that [point of radical differentiation].
+
+**Extended template (6 elements):**
+
+> - **WHAT:** The category you are in
+> - **HOW:** Your point of differentiation
+> - **WHO:** Your target audience
+> - **WHERE:** Your market geography
+> - **WHY:** The customer need you fulfill
+> - **WHEN:** The trend or moment that makes this relevant
+
+**Example (Harley-Davidson):**
+
+"The only motorcycle manufacturer [WHAT] that makes big, loud motorcycles [HOW] for macho guys and macho wannabes [WHO] mostly in the United States [WHERE] who want to join a gang of cowboys [WHY] in an era of decreasing personal freedom [WHEN]."
+
+**How to use as validation:**
+
+- If you cannot complete the basic statement convincingly, your positioning lacks a clear differentiator
+- If "only" feels like a stretch, you have not found genuinely unique ground yet
+- The extended version forces completeness — every element should feel natural, not forced
+- Try completing it from the customer's perspective: would THEY describe you this way?
+
+---
+
+## 4. JTBD — Jobs to Be Done
+
+**Core concept:** Customers do not buy products — they "hire" products to get a job done. The job is stable even as products come and go.
+
+**Three types of jobs:**
+
+- **Functional:** The practical task (e.g., "track my team's project status")
+- **Social:** How they want to be perceived (e.g., "look organized and in control to stakeholders")
+- **Emotional:** How they want to feel (e.g., "feel confident nothing is falling through the cracks")
+
+**Application to positioning:**
+
+- JTBD broadens the competitive alternative set — ALL solutions hired for the same job compete, regardless of category
+- Understanding the emotional and social jobs reveals positioning angles that feature comparisons miss
+- The language customers use to describe their job IS the language your positioning should use
+- Underserved jobs (important but poorly satisfied) = strongest positioning opportunities
+
+---
+
+## 5. Ries & Trout — Positioning Principles
+
+**Core principles from "Positioning: The Battle for Your Mind":**
+
+- **Positioning lives in the customer's mind, not your product.** It is about perception, not reality. You cannot force a position — you claim one that is available.
+- **Simplicity wins.** In an over-communicated world, only simple messages get through. One idea, clearly stated.
+- **The mental ladder.** Customers organize brands on mental ladders by attribute. You need to claim a rung. If the top rung is taken, find a different ladder.
+- **First beats best.** Being first in a category (in the customer's mind) is more powerful than being better. If you cannot be first in an existing category, create a subcategory where you ARE first.
+- **Consistency over time.** Positioning takes time to stick. Do not change it every quarter. Pick a position and hold it.
+
+**Application:** These principles act as a final sanity check. After building positioning through Dunford's process, test it: Is it simple? Does it claim one clear rung? Is that rung available? Would it stick over time?