5. Why a *pure* flat fee breaks (and what fixes it)

# 5. Why a *pure* flat fee breaks (and what fixes it)

The manifesto's **flat $20/seat** is the right *anchor* — one number, trivially conveyable. The problem is only
the **tail**: a pure flat fee with no guardrail. Margins below are the model's exact output (`npm run economics`),
all layers included (living-domain seats carry the 5-loop reference research portfolio, §3b):

| Scenario | Revenue | COGS (shipped) | Margin |
|---|--:|--:|--:|
| Solo — living domain (5-loop portfolio + light chat) | $20 | **$6.79** | **66%** |
| Solo — heavy creative (variable ~$12.5/seat) | $20 (flat only) | **$16.72** | **19%** ← the tail |
| Room — 5 seats, 1 box, living domain | $100 | **$22.49** | **78%** |

The median seat is healthy (66–78%). The danger is the **heavy tail**: research intensity and generation volume
are both unbounded, so a power user's variable COGS can run to $12–60/seat — and on a pure flat fee that lands
on us, dragging a $20 seat to 19% and, far enough out, **negative** (see `usage-sweep.png`: flat-only crosses
zero around ~$16 of variable COGS and keeps falling).

**The fix isn't a higher flat price for everyone, and it isn't a throttle.** It's **usage-based billing on top of
the flat fee** — the chosen model below. The flat fee stays the whole story for ~80% of users; the heavy tail
simply pays for what it pulls and becomes *more* profitable, not a loss.

---

Raw Markdown source