Risk

A risk framework that produces decisions, not adjectives

"Execution risk is high" tells a committee nothing actionable. A usable framework forces every risk into a category, a score, and an owner, so the aggregate becomes comparable across deals.

Published 28 July 2026 · 8 min read

Article

Six categories that cover most failure modes

CategoryCore question
MarketDoes enough reachable demand exist at this price?
FinancialDo the unit economics work before capital runs out?
ExecutionCan this team deliver this plan on this timeline?
TechnicalIs the product buildable and are dependencies survivable?
RegulatoryCan it operate legally in the target jurisdictions?
CompetitiveWhat happens when a funded incumbent responds?

Score likelihood and impact separately

Rate each risk 1–5 on likelihood and 1–5 on impact, and plot it on a 5x5 matrix. Separating the two dimensions prevents the common collapse where a catastrophic but unlikely event and a certain but minor one both get labelled "medium".

Define the scale in words before scoring, and keep the definitions constant across deals. Without shared definitions the scores are not comparable, and comparability is the main reason to score at all.

Decide mitigate, transfer, accept or avoid

Every risk gets one of these four dispositions. An unlabelled risk is an unmanaged one.

  • Mitigate: a concrete action with an owner and a date reduces likelihood or impact.
  • Transfer: insurance, contractual terms, or structuring the instrument to shift exposure.
  • Accept: the risk is understood, priced into the terms, and recorded as accepted.
  • Avoid: the risk is structural and the deal does not proceed.

Aggregate into a verdict without hiding the detail

Weight the categories according to stage. Pre-revenue deals weight market and execution; capital-intensive deals weight financial and regulatory. Publish the weights so the score can be challenged.

The aggregate should never replace the register. A committee needs to see the two or three risks driving the score, not just the number they produce.

Re-score, don't archive

Risk profiles move. A regulatory risk scored 4 before a consultation closes may be a 2 afterwards, and a competitive risk scored 2 changes the day an incumbent ships. Re-score at each milestone and keep the history — the trend is often more informative than the current value.

Keep reading

Valuation

Valuing a company with no revenue

The methods used to value pre-revenue startups — scorecard, comparables, dilution-first and venture method — and how to document a valuation so it can be defended later.