MarginPulse · scenario library v0.1

Five AI pricing scenarios to test before you promise “unlimited”.

This library turns the calculator into a practical launch review. Copy a scenario, enter current provider prices, and decide whether the feature needs a cap, tier, add-on, fallback model, or different plan price.

Open calculator Download CSV starter Copy worksheet kit

How to use this without fake precision

Verify current model prices on official provider pages; MarginPulse does not claim live rates.
Start with baseline usage, then test one downside scenario that would actually hurt margin.
Record the calculator memo in the launch doc before publishing plan copy or sales promises.
Count demand only when a real founder/team requests the worksheet, shares a pricing use case, or expresses purchase intent.

1. Baseline paid account

default plangross margin

Use when the product team has a normal expected usage pattern for paid customers.

Input	Starter value
Calls/customer/month	400
Input tokens/call	900
Output tokens/call	350
Decision	Launchable only if margin clears target after other COGS.

Memo line: “Baseline usage preserves/does not preserve target gross margin at the current plan price.”

2. Heavy-user overage

power userscap design

Use when a minority of customers could consume much more AI than the median user.

Input	Starter value
Calls/customer/month	2,000
Input tokens/call	1,200
Output tokens/call	600
Decision	If fragile, add a hard cap, fair-use rule, or paid overage.

Cap copy: “Includes X AI actions/month; higher usage requires upgrade or add-on.”

3. Free-trial abuse

trialanti-abuse

Use before offering free AI credits, ungated demos, or public beta access.

Input	Starter value
Trial accounts	100
Calls/account/month	100
Plan price	0 during trial
Decision	Trial usage is pure cost; limit actions and monitor abuse.

Guardrail: “Free trial includes a small AI action allowance; production-scale usage needs a paid plan.”

4. Team-seat account

B2B SaaSseat pricing

Use when one paying workspace can contain many users generating AI calls.

Input	Starter value
Seats/workspace	10
Calls/seat/month	250
Workspace plan price	Enter effective price/customer
Decision	If per-seat usage is uncapped, workspace pricing can hide loss-making accounts.

Question: “Is the cap per workspace, per seat, or both?”

5. Premium model fallback

model switchingquality tiers

Use when high-quality or long-context calls cost materially more than default calls.

Input	Starter value
Default calls	350
Premium calls	50
Premium output tokens	2–4× baseline
Decision	Reserve expensive calls for higher tiers or explicit add-ons.

Tier copy: “Standard AI uses the default model; premium/long-context AI is limited to higher tiers.”

6. Purchase-intent qualifier

validationsales signal

Use in replies to separate useful feedback from vanity attention.

What AI feature are you pricing?
Which scenario is the real risk: heavy users, free-trial abuse, team seats, model switching, or something else?
Would you use a worksheet/template pack before launch?
If yes, acceptable one-time price: CHF 19 / CHF 29 / CHF 49 / other.

Count only concrete answers, worksheet requests, or purchase intent. Do not count likes, views, generic praise, or directory acceptance.

Scenario-to-action map

If calculator says…	Pricing action	Copy/action before launch
Healthy margin	Keep price, monitor usage	Publish limits clearly and track AI actions/account.
Below target	Cap, tier, add-on, or higher plan price	Do not use “unlimited” language without a working cap.
Fragile margin	Model fallback or hard usage boundary	Move premium AI to higher tier or explicit add-on.
Underwater	Do not launch at that price	Raise price, reduce included usage, or narrow the feature.

Mercury Loop artifact. No live vendor-price claims, no fake demand, no personal identity. Created 2026-05-22T07:37:28+02:00.