Pricing Models

---
name: pricing-models
description: Playbook for choosing a pricing model, designing tiers, setting the free-vs-paid line, and estimating willingness to pay with a real method. Used by pricing-strategist.
when_to_use: |
  Use when choosing a pricing approach, designing tiers and packaging, setting
  the free-vs-paid line, or sizing willingness to pay. Primarily for the
  pricing-strategist agent.
allowed-tools: Read, Write, WebSearch, WebFetch
---

# Pricing Models — playbook

Pricing is the lever that connects product value to a business. A clear,
honest model beats a clever, fragile one.

## 1. The five pricing models — pick by how value grows with use

- **Subscription** — recurring fee for ongoing value. Predictable revenue;
  retention is the metric. Fits products used continuously.
- **Usage-based** — pay for what you consume (API calls, GB, seats). Aligns
  cost to value; harder to forecast. Fits products where heavy users get
  much more value than light users.
- **Freemium** — free tier + paid upgrade. Works ONLY when the free tier
  meaningfully limits power-user value AND attracts the right audience.
  Fails when free is "good enough" forever or attracts the wrong crowd.
- **One-time** — single payment. Simple. No retention lever; growth must
  come from new buyers. Fits one-shot value (a course, a tool, a template).
- **Hybrid** — combinations (subscription + usage overage, freemium +
  enterprise tier). Powerful but complex; use only when one model alone
  leaves obvious value uncaptured.

Pick by matching the model to HOW VALUE GROWS WITH USE:
- Value grows with use → usage-based has a real case.
- Value is ongoing → subscription.
- Value is one-shot → one-time.

## 2. Tiers — packaging the value

- **2–4 tiers** is the sweet spot. More than 4 confuses; fewer than 2 leaves
  money on the table.
- **Name tiers by user type, not by size.** Starter / Growth / Enterprise is
  fine; Bronze / Silver / Gold is lazy.
- Each tier must have ONE clear **anchor feature** — the reason to upgrade.
- Avoid the "middle-tier dominant" pattern — the middle tier should be
  intentionally less attractive than the upgrade so the upgrade has pull.

## 3. The free-vs-paid line

The free tier is a feature, not a leftover. Design it on purpose:
- Gives away enough to prove value.
- Withholds enough that real users want to upgrade.
- Attracts the right audience (not freeloaders the product cannot serve).
- "No free tier" is a valid choice — say so plainly with the reason (sales
  cycle, low-margin, regulated market, etc.).

## 4. Willingness to pay (WTP) — cite a method, always

Three credible methods. State which one was used:
- **Van Westendorp** — 4 survey questions yielding a price range (too cheap /
  cheap / expensive / too expensive). Cheap to run; survey-based.
- **Conjoint analysis** — users choose between bundles at different prices;
  reveals what they actually trade off. Requires real respondents and care.
- **Competitor anchoring** — price relative to a comparable competitor at a
  known price. Quick but inherits competitor's pricing logic — useful as a
  sanity check, weak as a sole basis.

A WTP number without one of these is a guess — label it so or do not include
it.

## 5. Price elasticity & psychology (light)

- Round numbers feel anchored ($99 vs $97). Round when honesty matters; cut
  if the buyer is price-sensitive.
- Annual vs monthly: discount annual ~15–20% to reward commitment.
- Three-tier pages: position the recommended tier middle with a clear "best
  value" cue.
- Beware decoy tricks if the product is sold to discerning buyers — they
  smell it.

## 6. Honesty rules
- Every price has a stated reason: cost-plus, value-based, competitor-anchored.
- Every WTP claim cites its method.
- "We do not know the price yet — here is the experiment that would tell us"
  beats a confident wrong number.

## 7. Output shape

A pricing & packaging plan that shows: the chosen model and why; the tiers
with anchor features; the free-vs-paid line with intent; the WTP rationale and
range; and 1–2 monetization experiments that would reduce the largest
remaining uncertainty.