| name | cmstrat |
| description | Define your ideal design-partner and early-customer strategy before you chase logos. Usage: /cmstrat. Reads STARTUP_PROFILE.md and any cmlist.md or cmmatch-*.md files, then produces an honest go-to-market strategy: a sharp Ideal Design Partner Profile and Ideal Customer Profile, how many to sign and in what sequence, what to ask each for (LOI / paid pilot / design-partner agreement), the anti-patterns to refuse, and this week's next actions. Saves to cmstrat.md. |
| license | MIT |
| allowed-tools | Read Write Bash(find:*) Bash(ls:*) |
| metadata | {"author":"3Flux","version":"1.0","workflow-step":"1 — run first to define who you should and shouldn't sell to"} |
/cmstrat — Design-Partner & Beachhead Strategy
The wrong early customers can kill a startup as surely as no customers: the tire-kicker who never buys, the whale who hijacks your roadmap, the free-pilot-forever logo that never converts, the design partner whose problem isn't your beachhead. This skill defines who you should pursue — and who to refuse — before you spend a single outreach.
Synthesize everything known about the startup into a design-partner and early-customer strategy: a precise ideal-partner profile, how many and in what order, what to ask for, the anti-patterns, and what to do this week.
Invocation
/cmstrat
No argument needed. Reads from the current working directory.
Execution Steps
Step 1 — Load All Available Context
Read the following from the current working directory (load all that exist):
STARTUP_PROFILE.md — required. If missing, stop and tell the user to create it first.
cmlist.md — if exists, use for the account pipeline and tiers
cmmatch-*.md — load all account fit reports; use for sequencing
cmposer-*.md — if any exist, extract the Bad-Fit Score and verdict per account
Step 2 — Produce cmstrat.md
Write a 7-section strategy memo. Be specific and disciplined — the goal is a profile sharp enough to disqualify most of the market on purpose.
Section 1: The Beachhead
One tight paragraph. Name the single beachhead segment the startup should win first — specific enough that you can list real companies in it. Assess:
- Who has the problem most acutely, has budget, and can decide quickly
- Why this segment first (urgency + reachability + reference value), not a bigger but slower one
- One-sentence verdict: "Win [specific segment] first by landing [N] design partners that look like [profile]; everyone else is a distraction until then."
Section 2: Ideal Design Partner Profile (IDPP)
A design partner is not just an early customer — they co-build, give deep feedback, and tolerate a rough product in exchange for influence and early value. Define the IDPP precisely:
| Attribute | Ideal | Why |
|---|
| Company type / size | [specific] | [reason] |
| Buyer & champion role | [who signs, who champions internally] | [reason] |
| Pain acuity | [how badly they feel the problem — must be top-3 priority] | [reason] |
| Budget authority | [can they actually pay, and when] | [reason] |
| Speed | [decision velocity you need] | [reason] |
| Reference value | [will their logo/quote move the next 10 buyers?] | [reason] |
| Data / domain access | [what they give you beyond money] | [reason] |
Section 3: Ideal Customer Profile (ICP) — the paying version
How the ICP differs from the IDPP once you move from co-building to selling. Name what graduates a design partner into a paying reference, and what you'll stop tolerating (custom one-offs, free forever) once you're selling.
Section 4: What to Ask Each For
Match the ask to the relationship. Be concrete.
| Stage | The ask | The terms that protect you |
|---|
| Design partner | [design-partner agreement / paid pilot] | [scope, time-box, success criteria, conversion path, who owns IP] |
| Early customer | [paid pilot → contract] | [price, term, reference rights, no perpetual free] |
| Reference | [logo + quote + intro rights] | [what you give in return] |
State plainly: never sign a design partner without a written success definition and a conversion path. A pilot with no defined "yes" is a free-forever trap.
Section 5: Sequencing & Count
- How many design partners: the right number (usually 2–4 — enough signal, not so many you're spread thin). Justify for this startup.
- In what order: which segment/account first and why (reference value, speed, or learning).
- Parallel vs. sequential: recommend one with rationale.
- If a
cmlist.md exists, pull Tier 1 accounts here by name.
Section 6: Anti-Patterns — Who to Refuse
The early customers and design partners to walk away from, even when they say yes. For each, the tell and the cost.
| Anti-pattern | The tell | Why it kills you |
|---|
| Roadmap hijacker | Wants features no one else needs | Builds a product for one company |
| Free-pilot-forever | Endless "evaluation," no budget line | Burns runway, never converts |
| No budget authority | Champion can't sign, "I'll need to check" forever | Long cycle, no close |
| Vanity logo | Big name, tiny/no usage | Looks good, proves nothing |
| Wrong-segment enthusiast | Loves it, but isn't your beachhead | Distorts your ICP and roadmap |
Section 7: This Week — Next 3 Actions
Concrete and specific.
- [Action]: [exactly what, which account, what to send] — deadline: [day]
- [Action]: [exactly what] — deadline: [day]
- [Action]: [exactly what] — deadline: [day]
Output Rules
- The IDPP must be specific enough to name real candidate companies — if it isn't, sharpen it
- Section 6 (anti-patterns) is mandatory and must be tailored to this startup's actual risk, not generic
- Every ask in Section 4 must include the term that protects the startup (success definition, conversion path, reference rights)
- Account names in Sections 5 must come from
cmlist.md or cmmatch-*.md if those exist — do not invent
- This Week actions must be doable in 7 days with no further research
- After saving, tell the user: "Run
/cmlist to build your target account list, then /cmposer <account> on each Tier 1 before investing in the relationship."