Playbook
Product Marketing Context
Builds a reusable profile of your product, audience, positioning, and proof that every other playbook can draw on.
- Best for
- The first step on any new product, client, or market.
- What you get
- A complete product marketing context document.
Use it in three steps
- 1
Copy the prompt
One click copies a version that works on its own in any chat. It tells the AI to ask about your business first and to cover anything the original setup files would have provided, and it includes 1 reference guide the playbook relies on.
↓ Download as a .md file↓ Original files (.zip)
If your AI app limits pasted text, attach the .md file to your chat. The .zip holds the original files for Claude Code and other agent tools. How to use the lab
- 2
- 3
Answer its questions
The AI will ask about your product, audience, and goal. Answer in a few plain sentences, then ask for the deliverable you need. Specific answers produce specific work.
Read the full playbook
Product Marketing Context (Global Foundation)
Operating rules:
skills/en/references/operating-rules-global.md(esp. rules 2, 3, 7).
Foundation skill for the GLOBAL marketing cluster — runs BEFORE every other global skill (00-29). Creates
.agents/product-marketing-context-global.mdwith product, customer, positioning context tuned to a specific region. Parallel toproduct-marketing-context(the Vietnam foundation); this one serves 4 regions: US, EU, SEA, LATAM.
For newbies
Who is this skill for?
| Audience | Concrete example |
|---|---|
| Founder going international | Already in Vietnam, expanding to US/EU/SEA |
| Global SaaS / DTC brand | Multi-region launch needing region-specific context per market |
| Agency serving international clients | Onboarding a US, EU, SEA, or LATAM client |
| Marketer entering a new market | Setting up positioning for a new geography |
Who is this NOT for?
- Vietnam-only marketing -> Use
product-marketing-context(VN foundation) instead - Personal brand (founder/coach/creator) -> Use
22-personal-brand-context-global(when available) - Single ad campaign -> Use
02-campaign-brief-globaldirectly; you do not need a full context file
30-second pre-read
This skill produces ONE region-tagged file (.agents/product-marketing-context-global.md) with 12 sections describing your product, customers, competitors, positioning. Pick 1 of 4 region variants (US, EU, SEA, LATAM) — the variant tunes currency, platforms, regulations, persona framework. Every global skill (00-29) reads this file before working — like a medical chart every doctor reads first.
3 common errors
- Wrong region variant — "I sell in 5 countries" -> pick PRIMARY market first; multi-region later.
- Mixing currencies — Mixed USD/EUR/local without a primary anchor confuses downstream skills.
- Skipping regulations — GDPR/CCPA/PDPA/LGPD shape data collection; ignoring breaks campaigns at launch.
Why do you need this skill?
Without this file, every global marketing skill has to re-ask: What is the product? Who is the customer? Which region? Currency? Regulations? Platforms? Repetitive, slow, inconsistent.
With .agents/product-marketing-context-global.md:
- Answer ONCE at project start
- Every skill reads it automatically -> straight to work
- Region defaults (currency, platforms, regulations) propagate everywhere
- No drift between skills — copy and report use the same persona, currency, metrics
Workflow
Step 1: Check existing file
|-- exists -> ask which section to update
|-- missing -> Step 2
Step 2: Choose region variant (US / EU / SEA / LATAM)
Step 3: Collect info per variant (12 sections)
Step 4: Create file `.agents/product-marketing-context-global.md`
Step 1: Check existing file
Check whether .agents/product-marketing-context-global.md already exists.
- If it exists -> read the file, summarize, ask user which section to update. Only collect info for that section.
- If it does not exist -> proceed to Step 2.
Step 2: Choose region variant
Ask the user one question: "Which is your PRIMARY region: US, EU, SEA, or LATAM?"
Decision tree
Where is your primary market?
|-- North America (USA, Canada)
| --> 01-us.md (USD, FTC + CCPA, Meta + Google + TikTok)
|-- European Union (UK, DE, FR, ES, IT, NL, etc.)
| --> 02-eu.md (EUR, GDPR + EU AI Act, Meta + Google + LinkedIn)
|-- Southeast Asia EXCEPT Vietnam (SG, MY, ID, TH, PH)
| --> 03-sea.md (USD or local, PDPA, TikTok + Shopee/Lazada)
|-- Latin America (BR, MX, AR, CO, CL)
| --> 04-latam.md (USD or local, LGPD, WhatsApp + Meta + Mercado Libre)
|-- Vietnam only
| --> Use `product-marketing-context` (VN foundation) — NOT this skill
|-- Multi-region (>2 regions equally important)
--> Pick the LARGEST market as primary; note others in section 12
Detection logic
- User says explicitly ("we sell in the US") -> load matching variant
- User vague ("we sell internationally") -> ask which country has the most revenue today
- Truly multi-region -> pick primary market; add secondary markets in section 12 (Goals)
Step 3: Collect info per variant
After choosing the variant, read the matching file in variants/ and walk through each section:
- Brief explanation — what this section captures and why it matters
- Ask 1-3 questions — one section at a time, NEVER all at once
- Confirm — read back; user confirms
- Move to next section
Notes:
- Every variant has 12 sections (matching the VN foundation structure)
- All 4 variants share the same skeleton but have DIFFERENT defaults, examples, and region-specific notes
- Region cues: currency format, platform list, regulation pointer, cultural context
Variant summary:
variants/01-us.md— USD, FTC + CCPA, Meta + Google + TikTok, JTBD + ICPvariants/02-eu.md— EUR, GDPR + EU AI Act, Meta + Google + LinkedIn, GDPR-compliant ICPvariants/03-sea.md— USD or local, PDPA, TikTok + Shopee/Lazada + LINE, mobile-first personavariants/04-latam.md— USD or local, LGPD, WhatsApp + Meta + Mercado Libre, WhatsApp-first persona
Step 4: Create the file
Save to .agents/product-marketing-context-global.md (create the .agents/ folder if missing).
# Product Marketing Context (Global) — [Product name]
> Region variant: [US / EU / SEA / LATAM]
> Primary currency: [USD / EUR / SGD / BRL / etc.]
> Last updated: [YYYY-MM-DD]
> Compatible with: Claude Code, ChatGPT, Gemini, any AI agent
---
[12 sections following the chosen variant]
---
## 13. Paid Ads (optional; skip if you do not run ads yet)
> Ads skills (`10-reverse-kpi-global`, `53-tracking-setup-global`, `54-media-plan-global`, `21-ads-audit-global`) read this block. Where a value is unknown, write `[REAL DATA NEEDED]`; do not guess.
- **Ad accounts:** (platform, account name, ID. NEVER write tokens or passwords)
- **Pixel / CAPI status:** (not installed / installing / verified). **Last verified:** [YYYY-MM-DD] (how to check: see `53-tracking-setup-global`)
- **Target CPA / ROAS:** [REAL DATA NEEDED] (how to derive: see `10-reverse-kpi-global`)
- **Monthly budget:**
- **Approver for budget and account changes:** [name]
## 14. Agent permissions
> **Only a human edits this section. Agents never edit it; they stop and ask when an action conflicts.** See `skills/en/references/operating-rules-global.md` rule 7. An `execute` row is a standing authorization under rule 7: its Limit column must state the allowlist, spend ceiling, per-run limit and expiry date.
| Action | Agent may (read / draft / propose / execute / never) | Limit | Approver |
|--------|------------------------------------------------------|-------|----------|
| Read reports and dashboards | read | Only the accounts listed in section 13 | - |
| Draft copy, briefs, captions | draft | Follow brand voice (section 10); no claims without proof (section 11) | [name] |
| Change budget / bid / target | propose | One change block at a time; execute only after the approver says yes | [name] |
| Publish a public post | propose | Approver reviews the final version before it goes live | [name] |
| Upload a customer list | never | - | - |
## 15. Learnings (max 20 active rows)
> Agents may only **append proposed rows** (Status = `proposed`); the approver moves a row to `active`. Over 20 active rows, retire the expired ones.
| Date | Learning | Evidence (observed / reported / inferred) | Source | Status (active / retired) | Retire when |
|------|----------|-------------------------------------------|--------|---------------------------|-------------|
| [YYYY-MM-DD] | [one sentence] | observed | [link / report name] | active | [condition or date] |
How other skills use this file
In every global marketing skill (00-marketing-plan-global, 05-ad-copy-global, etc.), add this block under "Collect info":
## Collect info
### Step 0: Check global context file
Check whether `.agents/product-marketing-context-global.md` exists:
- **Yes** -> Read everything; pull product / customer / region / currency / regulation context. Do NOT re-ask.
- **No** -> Suggest user run `product-marketing-context-global` first.
If user wants to continue anyway, ask 2-3 minimum questions about product + region.
Skills that will read this file: 00-marketing-plan-global, 01-content-calendar-global, 02-campaign-brief-global, ... 29-* (all global cluster skills).
Quality checklist
- File saved at
.agents/product-marketing-context-global.md - Region variant tagged at the top (US / EU / SEA / LATAM)
- All 12 sections completed (or at least 8 core: 1, 2, 4, 5, 6, 9, 10, 12)
- Currency format matches region (e.g.
$9.99for US,€9,99for EU) - Voice of customer (section 9) uses REAL quotes — not paraphrased
- Anti-persona explicit (section 7); 3 concrete differentiators (section 6)
- North Star Metric defined (section 12); last-updated date in file
- If running ads: section 13 has a tracking verify date and a named approver (no tokens or passwords)
- Section 14 Agent permissions is filled in by a human; no write action left blank
- Section 15 Learnings rows carry evidence (observed / reported / inferred) and a retire condition
Related skills
product-marketing-context— Vietnam foundation (parallel)00-marketing-plan-global— reads this file first22-personal-brand-context-global— personal brand foundation (when available)
Global Foundation Skill | Over Powers Agency | v1.0.0