Value metrics

Pricing is not a finance afterthought glued onto a finished product. It decides who you are for, which features matter, how unit economics work, and whether expansion revenue appears without a hero sales motion. Good pricing captures a fair share of the value you create — not "cost plus a bit," and not "whatever the competitor's website shows."

Customers buy outcomes. Your price should scale with something that tracks that outcome as they grow. When price and value move together, upgrades feel fair; when they do not, customers feel rented or stuck.

Patrick Campbell and others in monetization research often note that small improvements in monetization and retention can outweigh similar relative gains in acquisition — because they flow through the whole base. Treat pricing reviews as regularly as roadmap reviews.

Choosing a value metric

Good value metrics are:

1. Easy to understand at purchase time.
2. Aligned with value the customer perceives.
3. Aligned enough with your COGS that heavy users pay more (critical for AI and infra-heavy products).
4. Auditable in the product (you can meter them).

For AI products as of when this was written, pure seat pricing often fails when one power user burns orders of magnitude more tokens than another. Hybrids (base platform + included usage + overage) are common. Outcome-based pricing ("pay per successful resolution") is attractive in pitch decks and hard in contracts — start with usage you can meter cleanly.

Walking through a job-workflow API

You build Pipekit, an API for background job workflows. COGS scale with executed steps. Buyers are indie hackers, startups, and enterprises. Value metric: workflow steps executed per month (aligns with value and COGS), plus seats for the dashboard (collaboration). Seat-only would subsidize power users until margins break.

Common mistakes

• Seat-only when costs are usage. You subsidize power users and overcharge light ones — acute for AI features.
• A meter nobody understands. "Billable message unit subtype B" will not survive a purchase conversation.
• A metric you cannot audit. If you cannot meter it, you cannot bill it.

Check yourself

One power user burns 100× the tokens of a teammate. Seat-only pricing? It fails. Hybrid: platform fee + included usage + overage, so cost-to-serve shows up in the bill.

Charging per seat for a usage-heavy API. What incentive did you just create? Login sharing, and a margin hole on heavy users.

Key takeaways

• Price to capture value. Pick a unit that tracks customer value and your costs.
• Good metrics are understandable, value-aligned, COGS-aligned enough, and auditable.
• AI and infra almost always need a usage component, not seats alone.

Further learning

Next: good-better-best packaging. Map: pricing and packaging.