investment-memo-writing

name	investment-memo-writing
description	Write an investment memo that forces a decision: explicit recommendation (no hedging), steelman against, decision log, and next steps. Use when preparing for IC or partner discussion.
license	Proprietary
compatibility	Works offline; improved with web comps and diligence evidence; optional Salesforce logging.
metadata	{"author":"evalops","version":"0.2"}

Investment memo writing

When to use

Use this skill when:

• You need to brief partners / IC with a decision-ready document
• You have diligence evidence and need to turn it into a recommendation
• You want to reduce "vibes" and increase clarity

Inputs you should request (only if missing)

• Company deck + diligence notes
• Round terms (or what's known so far)
• Firm constraints (ownership target, stage, geography)
• Comparable comps / pricing benchmarks if available

Output you must produce

A memo that:

• Opens with a clear recommendation in the first line (no hedging)
• States the best argument AGAINST the deal (steelman)
• Includes a decision log (what changed since first call, and why)
• Names top risks (max 4, ranked) and evidence for each
• Proposes decision + next steps with dates

Templates:

• assets/memo-template.md
• assets/outcome-framing.md

Writing rules

• First line = recommendation. "Recommend: Lead at $X" or "Recommend: Pass because Y". No "it depends."
• Adjectives are a tax. Replace them with evidence.
• The best argument against the deal is mandatory. Steelman it.
• If it's important, quantify it or show a proxy.
• Keep TL;DR to 5 bullets max.
• Include a decision log: what you believed at first call, what changed, why.

Memo structure (required sections)

1) Recommendation (FIRST LINE, NO HEDGING)

• "Recommend: Lead $Xm at $Ym pre for Z% ownership"
• or "Recommend: Follow with $Xm"
• or "Recommend: Pass"
• One sentence. No "lean towards" or "potentially interested."

2) Best argument against (MANDATORY STEELMAN)

• In 3-5 sentences, make the strongest case for why this deal fails.
• This is not "risks" - it's the single most compelling reason a smart investor would pass.
• Write it as if you were advising a competitor fund to say no.

3) TL;DR (5 bullets max)

4) Decision log (what changed since first call)

Date	Belief	Evidence	Change
First call			Initial thesis
[Date]			Updated because...
[Date]			Updated because...
Today			Current view

5) Why now + wedge

6) Market (buyer, budget, trigger, adoption timing)

7) Product (10x, hard to copy, deployment reality)

8) Traction (the 2-3 metrics that matter)

9) GTM (motion, cycle time, pricing, margin)

10) Competition + moat

11) Team (learning rate, decision rights, gaps)

12) Must be true (3-5 falsifiable claims)

For each claim:

• The claim
• Evidence today
• What would falsify it

13) Risks (MAX 4, ranked)

For each risk:

• The risk
• Evidence collected
• Mitigation / what we learned
• Residual concern

14) Deal + terms + ownership target

15) Next steps + decision date

Appendix

• Customer quotes (especially anti-confirmation evidence)
• Diligence notes
• Competitor analysis

Anti-confirmation section (required in appendix)

Every memo must document:

• Churned customer / lost deal call: What did we learn? What would have changed the outcome?
• Competitor / alternative user call: Why do they use something else? What would switch them?
• If these calls didn't happen, explain why and flag the gap prominently.

Salesforce logging (recommended)

• Attach the memo as a File/Note to the Opportunity.
• Update Opportunity fields: stage, amount (if relevant), close date, next step.
• Record key "must be true" claims in a structured field if you have one; otherwise in Notes.

Use salesforce-crm-ops if you need API patterns.

Edge cases

• If terms are unknown: write the memo "pending terms" and explicitly state what term ranges would change your recommendation.
• If evidence is thin: say so. Propose the smallest next diligence step that would change the decision.
• If you can't make a clear recommendation: you're not done diligencing. Go back and get the evidence you need.