feat(skills): add marketing skills collection

shared/skills/pricing/references/pricing-models.md

@@ -0,0 +1,66 @@
+# Pricing Models
+
+The eight core ways to structure *how* you charge. This is distinct from the value metric (what unit you charge on) and the tier structure (how you package). Most real products **combine** two or more of these.
+
+## Contents
+- The 8 Pricing Models
+- Combining Models
+- The Value/Price Ratio
+- The Low-Price Retention Counterpoint
+
+---
+
+## The 8 Pricing Models
+
+| Model | How it works | Best when | Reference |
+|-------|-------------|-----------|-----------|
+| **Flat-rate** | One price, one product, everyone pays the same | Simple product, one persona, you want zero pricing friction | Basecamp |
+| **Usage-based** | Pay for what you consume (metered) | Value scales directly with volume; consumption is variable and easy to meter | Stripe |
+| **Tier-based** | Good-better-best packages at set prices | Distinct segments with different needs and budgets | Kinsta |
+| **User-based** | Price per seat/user | Value grows as more people in the org use it (collaboration) | Notion |
+| **Feature-based** | Price gated by which capabilities are unlocked | Clear feature tiers map to willingness to pay | Intercom |
+| **Credit-based** | Buy a bucket of credits, spend them on actions | Usage is lumpy or bursty; you want prepaid commitment and simple mental accounting | Audible |
+| **Outcome-based** | Pay per result delivered (resolution, task completed) | You can measure and attribute the outcome, and the outcome is what the buyer actually wants | Intercom Fin, Zapier |
+| **Hybrid** | Deliberate mix (e.g. platform fee + usage, or seats + credits) | A single model under- or over-charges different customers | Drift |
+
+### When to reach for each
+
+- **Flat-rate** — reach for it first if you can. It's the easiest to sell, easiest to understand, easiest to forecast. The tradeoff: you leave money on the table with your biggest customers.
+- **Usage-based** — the fairest model when consumption tracks value, but revenue is less predictable and buyers fear a surprise bill. Pair with spend caps or alerts.
+- **Tier-based** — the default for self-serve SaaS. Lets one page serve SMB through mid-market.
+- **User-based** — only if value genuinely rises with headcount. If it doesn't, seats punish adoption (teams share logins to avoid paying).
+- **Feature-based** — powerful for segmentation, but don't gate the feature that delivers your core value; gate the ones that separate casual from serious users.
+- **Credit-based** — good for AI/actions-based products where each action has a cost. Credits decouple price from a single unit and make prepayment feel natural.
+- **Outcome-based** — the emerging model for AI agents (charge per resolved ticket, per automation run). Highest trust because the buyer only pays when they win — but only viable when the outcome is measurable and clearly attributable to you.
+- **Hybrid** — where most mature products end up. A base platform fee for predictability plus a usage/outcome component for upside.
+
+---
+
+## Combining Models
+
+These aren't mutually exclusive. Common combinations:
+
+- **Tiers + per-user** — seats within each package (most B2B SaaS)
+- **Platform fee + usage** — predictable base, variable upside (Twilio-style)
+- **Seats + credits** — pay per person, then top up credits for heavy actions
+- **Feature tiers + outcome** — unlock capabilities by tier, charge per result on top
+
+Pick the primary model from the value metric, then layer a second only if a single model clearly mis-prices a real segment.
+
+---
+
+## The Value/Price Ratio
+
+Aim for roughly a **10:1 value-to-price ratio** (Ryan Kulp): the customer should perceive about **10x more value than they pay**. This is the buffer that makes the purchase feel obvious rather than negotiated, and it leaves headroom to raise prices later as you add value.
+
+If you can't articulate 10x value, the problem is usually the offer or the positioning, not the price point.
+
+---
+
+## The Low-Price Retention Counterpoint
+
+Charging too little is not the safe choice. **Low prices hurt retention** (Patrick Campbell / ProfitWell data, echoed by operators like Josh Pigford of SpyFu and Tyler Tringas): under-priced customers churn *more*, not less, because a low price signals low value and attracts the least-committed, most price-sensitive buyers.
+
+Related: the **discount-asker signal** — customers who negotiate for a discount tend to churn at roughly **2x** the rate of full-price customers. Discounting to close a deal often buys a customer who leaves anyway.
+
+**Implication:** when in doubt, price higher. It's easier to grandfather a price down than to claw one up, and a higher price selects for better-fit, longer-retained customers.