pricing-model-selection

Skill: Pricing-Model Selection

The pricing model is how the customer is charged. It is downstream of the value metric (what you charge per) — so sketch the metric (see value-metric-design) before locking the model. This skill traces a product to its best-fit model and names the runner-up.

Step 0 — One opinion up front

Hybrid is the default answer for anything consumption-driven. A committed base fee + an included allowance + metered overage gives the vendor a revenue floor and the customer a predictable bill while still aligning price with value at the margin. Reach for a pure model only when the product is genuinely simple (flat capability, predictable use) or genuinely metered (clean per-event value).

Step 1 — Characterize the value delivery

Answer three questions:

1. Is value continuous (ongoing) or one-time/episodic? Continuous → subscription family; episodic → transactional; one-time → perpetual + maintenance.
2. How much does consumption vary across customers? Low variance → seat or flat/tiered; high variance → usage or hybrid.
3. Does value scale with the number of people, or with the capability itself? People → per-seat; capability → flat-rate/tiered.

Step 2 — Run the model-selection tree

Traverse ../../knowledge/pricing-decision-trees.md §1 to a leaf. Record the path you took.

Step 3 — Apply the model-specific cautions

• Per-seat: caps at seat count and can penalize adoption (teams ration logins). If AI does the work, per-seat shrinks the account as the product succeeds — reconsider.
• Usage-based: bill-shock risk + unpredictable revenue. Add an included allowance, caps, and usage alerts; consider the hybrid instead.
• Freemium: a model only with a measured conversion path AND a bounded cost-to-serve. Absent either, use a time-boxed free trial.
• Hybrid: size the included allowance to the median customer; price overage to be felt but not punitive.

Step 4 — Stress-test against the failure question

Ask: "If this product becomes 10× more valuable to the customer, does our revenue grow with it?" If the model says no, you've picked a metric/model that caps the company — go back to Step 1.

Step 5 — Hand off

• The dollar impact (LTV, margin, cash) of the chosen model → finance.
• The value metric decision → value-metric-design (do it alongside, not after).
• WTP validation of the price points the model implies → willingness-to-pay-research.

Output

A model recommendation with: the tree path, the runner-up and why it lost, the model-specific cautions that apply, and the failure-question check.