founder-competence-self-assessment

Work through a structured 10-area / 76-item self-assessment (AFCA) of a founder's or team's readiness for an academic spin-out or research-based startup.

SKILL.md

GitHub

Founder Competence Self-Assessment (AFCA)

Purpose

Work through a structured 10-area / 76-item self-assessment (AFCA) of a founder's or team's readiness for an academic spin-out or research-based startup.

Based on

the owner's published commercialisation guide (owner, 2025). The AFCA self-assessment tool (the owner's own). Full data: ../../references/afca-framework.md.

  • A synthesis of the EU's EntreComp and ResearchComp frameworks and peer-reviewed research on commercialisation competence
  • 10 areas, 76 items, a 1-7 scale, weighted average normalised to 0-100%

Method

  1. Work through all 10 areas in order (full item list: ../../references/afca-framework.md): Opportunity Orientation, Market & Customers, Financial Planning & Management, IP & Evidence, Venture Building, Partnerships & Networks, Communication & Influence, Execution & Roadmap, Regulatory & Quality, Founder Mindset & Teaming.
  2. Score each item on a 1-7 scale honestly — avoid over- or under-estimating; if useful, have another team member score in parallel.
  3. Calculate the weighted average per area and normalise to 0-100%: (raw score - 1)/6*100.
  4. Place each area at a level: Foundational (<25%), Intermediate (≥25%), Advanced (≥50%), Expert (≥75%).
  5. Calculate the overall score as the average of the area averages.
  6. Identify the 2-3 weakest areas — these are the most likely stumbling blocks. Plan a concrete action for each (e.g. a low IP & Evidence score → work through ip-disclosure-and-ownership-check).
  7. Repeat the assessment every 3-6 months to track progress — this is a development tool, not a one-time verdict.

Gotchas

  • Self-scoring bias is the failure mode the Method names directly (step 2): over- or under-estimating on the 1-7 scale is likely with a single scorer, but the output (a normalised percentage) can look precise enough to mask that unreliability. Have a second team member score in parallel when possible.
  • The normalisation formula is (raw score - 1)/6100, not raw/7100 — using the wrong denominator shifts an area's placement across a maturity boundary (e.g., Foundational vs. Intermediate) near the edges of the scale.
  • The overall score is an average of area averages, not a minimum, so a strong overall percentage can hide one critically weak area (e.g., IP & Evidence). Step 6 requires looking at the 2-3 weakest areas separately — the topline number alone isn't enough to plan action from.
  • Treating this as a one-off test defeats its purpose: the Method calls for repeating every "3-6 months" specifically so progress on prior action items can be checked. A single snapshot can't show whether a weak-area action plan actually worked.
  • This tool is explicitly not psychometrically validated (see "What this skill does NOT do") — presenting its 0-100% output as a certified competency score (e.g. in an investor deck) overstates its rigor beyond what the framework supports.

What this skill does NOT do

  • Is not a psychometrically validated clinical test — a self-assessment whose reliability depends on honesty.
  • Does not make a hiring or investment decision for you.
  • Does not replace the other skills in this pack for a deep dive into a given area — it acts as a map that points to the right skill.

Continue from here

  • Based on the score, go back to the relevant skill: low Market & Customers → research-opportunity-recognition; low Venture Building → founding-team-design-and-agreements; low IP & Evidence → ip-disclosure-and-ownership-check; low Execution & Roadmap → commercialisation-journey-roadmap.
  • Pack's shared guardrails: ../../CLAUDE.md
  • Overview of the full journey: ../commercialisation-journey-roadmap/SKILL.md

References

  • ../../references/case-studies.md — 7 spin-out examples from different industries and regions
  • ../../references/terminology.md — the handbook's glossary
  • ../../references/sources.md — the handbook's own source references
  • ../../CLAUDE.md — the pack's shared guardrails