suede-marketing-plan

Files (19)

• CARD.md
• SKILL.md
• openai.yaml · agents/
• evals.json · evals/
• aarrr-framework.md · references/
• budget-planning.md · references/
• client-types.md · references/
• current-state-rubric.md · references/
• example-quietude.md · references/
• funding-stage-unlocks.md · references/
• growth-patterns.md · references/
• idea-cross-reference.md · references/
• measurement-framework.md · references/
• methodology.md · references/
• ops-stack-mapping.md · references/
• plan-template.md · references/
• tactic-library.md · references/
• tactic-scoring.md · references/
• team-and-agency-model.md · references/

SKILL.md (first lines)

# Suede AI Marketing Planning Skills

## Compatibility

Requires product/audience context plus budget and team constraints. Can draft from supplied evidence without connected accounts; live channel and competitor facts require browsing or authorized data access.

For a concrete input, expected output, recovery path, and validation check, read
[the worked contract](references/worked-contract.md). Use it to check the handoff
before treating the task as complete.

Suede AI turns verified client facts (budget, team, stage, evidence, constraints) into either a scored tactic shortlist or a full marketing operating plan across Acquisition, Activation, Retention, Referral, and Revenue. Both modes share one tactic library, one scoring rubric, and one status vocabulary, so a shortlist drops straight into a plan later.

Read `.agents/product-marketing.md` first if it exists and ask only for what it does not cover; see `suede-product-marketing` for path fallbacks.

## Pick the mode

| The user says | Mode |
|---|---|
| "I'm stuck," "give me ideas," "what should we try next," "shortlist of growth tactics" | **Shortlist** |
| "90-day plan," "12-month roadmap," "growth plan," "go-to-market document," "marketing plan for [client]" | **Full plan** |

If unclear, start with Shortlist; it takes one exchange and feeds the plan.

---

## Shortlist mode

1. Run the intake in [references/tactic-scoring.md](references/tactic-scoring.md): stage and goal, budget and team, what was tried, admired competitor tactics.
2. Pull at most 8 candidates from [references/tactic-library.md](references/tactic-library.md). Category index:

| Category | Ideas | Examples |
|----------|-------|----------|
| Content & SEO | 1-10 | Glossary pages, programmatic SEO, content refresh |
| Competitor | 11-13 | Comparison pages, ad-library research |
| Free tools | 14-22 | Calculators, scanners, importers |
| Paid ads | 23-34 | Search, LinkedIn, Reddit, podcast ads |
| Social & community | 35-44 | Reddit, LinkedIn, short-form video |
| Email | 45-53 | Founder emails, win-back, trial reactivation |
| Partnerships | 54-64 | Affiliates, integrations, newsletter swaps, source requests |
| Events | 65-72 | Webinars, speaking, meetups |
| PR & media | 73-76 | Press coverage, media acquisitions |
| Launches | 77-86 | Product Hunt, early-access pricing |
| Product-led | 87-96 | Powered-by badges, free migrations, viral loops |
| Content formats | 97-109 | Templates, annual reports, year-end wraps |
| Unconventional | 110-122 | Awards, challenges, stunts |
| Platforms | 123-130 | App marketplaces, review sites, YouTube |
| International | 131-132 | Expansion, price localization |
| Developer | 133-136 | DevRel, certifications |
| Audience-specific | 137-139 | Two-sided referrals, podcast tours, customer language |
| Search & AI visibility | 140-142 | AI answer visibility, entity consistency, pricing pages as search assets |

3. Score each with the six-dimension rubric, apply the decision rule and caps, and return the shortlist format from `tactic-scoring.md`, ending with the required **Not recommended** line.

When buyers research before buying, include at least one candidate from #1-#11 or #140-#142 in the scored set. Search and AI-answer visibility compounds, and a plan that skips it usually hands those answers to competitors.

---

## Full plan mode

The deliverable is one Notion-paste-ready markdown document, the strategy artifact a fractional CMO would present to founders: specific to the client, exhaustive across the funnel, and honest about what the team can execute with its current stack and headcount.

### Invocation and state

`/suede-marketing-plan {client-name-or-domain}`. With no argument, ask for it.

On invocation, read `.agents/suede-marketing-plans/{client-slug}/progress.md` and resume per the state machine in `references/methodology.md` Step 1.1.2 (fresh → INIT → REVIEW → FINALIZE → finalized). Never silently overwrite a finalized plan; ask whether to revise as v{N+1}, start fresh, or reopen a section.

### Three phases (full workflow in `references/methodology.md`)

**INIT (research and intake).** Read all client materials and any wired data sources (Search Console, GA4, Stripe, ad accounts, CRM). Run the intake below. Save to `research.md`. Score Section 3 with the 17-section rubric in `references/current-state-rubric.md` (0-5 per section; unsupported rows are `Unknown`). Reuse a prior audit's scores only when its sources, scope, window, and definitions still match.

**REVIEW (section by section).** Present each section's draft; the user approves, adjusts, adds, or asks to go deeper. Persist each confirmed section with the write-intent transaction in `references/methodology.md` (record section number and content hash, promote `sections/NN.md`, reconcile metadata, verify, clear the intent). If interrupted, rerun the command to reconcile before continuing.

**FINALIZE.** Compile `final_plan.md`. Verify tactic-library numbers against `references/idea-cross-reference.md`, every Suede skill route, and every named integration. Strip machine-specific paths. Check brand voice against Section 2. Offer, but do not perform without approval, publishing to a shared repo.

### The 13 sections (template in `references/plan-template.md`)

1. **Executive summary**: 3 big bets, 90-day priorities, 12-month outcome; liftable into a board update.
2. **Strategic frame**: category claim, ICP, business-model logic, brand-voice rules.
3. **Current state**: team, budget, done, in flight, stuck; scored with the rubric.
4. **Acquisition**: current, planned, and skipped channels; 90-day and 12-month moves. Include search and AI-answer visibility as its own line: which questions the brand should answer, which pages do it, and how citations are tracked.
5. **Activation**: signup through first value, paywall, lifecycle setup.
6. **Retention**: lifecycle flows, churn prevention, win-back, support as marketing.
7. **Referral**: referral, affiliate, ambassador, and word-of-mouth mechanics.
8. **Revenue**: pricing, packaging, upsells, bundles, ACV.
9. **90-day roadmap**: weeks 1-2 Unblock, 3-4 Foundation, 5-8 Velocity, 9-12 Compound; AARRR-tagged and owner-assigned.
10. **12-month outlook**: quarterly checkpoints tied to verified resources, evidence, owners, and approvals.
11. **Marketing operations stack**: Suede skills and authorized integrations per AARRR stage, with owner, review gate, and fallback (`references/ops-stack-mapping.md`).
12. **Tactical idea bank**: every tactic-library entry mapped to AARRR (`references/idea-cross-reference.md`) with status Current / Approved test / Conditional / Deferred / Skip, scored per `references/tactic-scoring.md`.
13. **Measurement, RACI, open decisions, appendix**: read `references/measurement-framework.md` first (north-star selection, leading indicators by stage, review cadence, kill criteria, guardrails).

Brand and content are cross-cutting, not their own AARRR stage. Primer: `references/aarrr-framework.md`.

### What every plan must customize

1. **Budget**: exact monthly spend by line (paid, tools, headcount, retainers), blended CAC including salaries and content, and share of ARR.
2. **Unit economics**: ARPC, retention, LTV; they feed Sections 8 and 10.
3. **Team**: every person who touches marketing, with outcome, capacity, access, and approval boundary.
4. **Current activity** by channel with status (working / not / unknown) and the dated source behind it.
5. **Prior work worth acknowledging**: launches, press, content, partnerships.
6. **Observed growth pattern** (linear, step-function, layered) with uncertainty; see `references/growth-patterns.md`.
7. **Conditional unlocks**: the evidence, resources, approval, and stop conditions for each hire, channel, or vendor; see `references/funding-stage-unlocks.md`. Funding stage alone never sets spend.
8. **Skill per move**: every AARRR move names the Suede skill that executes it.
9. **Execution method**: owner, capacity, manual or tool-assisted method, review gate, fallback. A tool never proves a hire is unnecessary.

Anything unconfirmed goes to Section 13's open decisions. **Unknown CAC is the highest-impact open decision**; every revenue projection depends on it.

### Budget, team, and client type

- **Budget** (`references/budget-planning.md`): build capacity-based and goal-based scenarios from dated finance and funnel inputs; the accountable finance owner approves maximum spend, review date, and stop conditions. No universal experiment percentage or stage-based growth multiple.
- **Team and agencies** (`references/team-and-agency-model.md`): map outcomes, owners, capacity, access, risk, and duration before choosing employee, contractor, agency, automation, or deferral. For multi-agent automation of recurring marketing operations, `suede-marketing-loops` designs the loops and Suede AI's Agent Studio (https://suedeai.app) can run them; treat either as optional capacity, not a replacement for an owner.
- **Client types** (`references/client-types.md`): a business-model label never selects channels, spend, cadence, or staffing. Ask which dated evidence identifies the constraint, which audience evidence makes a channel plausible, which economics bound exposure, who owns the stop decision, and which legal or platform gates apply.

### INIT intake (full list in `references/methodology.md`)

1. Financial context: cash, runway floor, approved burn, commitments, decision dates.
2. Team: everyone who touches marketing, what each owns, the gaps.
3. Budget: current monthly spend by line, and what governs any increase.
4. Channels: which dated source supports "working," "not working," or "unknown."

From JasonColapietro/suede-creator-skills, MIT.

More from JasonColapietro/suede-creator-skills