Replace the 30% metered markup model for managed token billing with Prepaid Token Banks at 15% markup. Users purchase dollar-denominated credit packs ($10/$25/$50/$100) with tiered bonuses at higher amounts. No expiration at launch. Optional auto-top-up. Per-task cost preview before execution. BYOT remains the hero path.
Context
DR-2026-001 established the pricing model pivot to team-based modules ($10-25/mo) with BYOT as the primary model and managed billing as a convenience option. The initial token economics analysis (v1.0) recommended a 30% flat metered markup on Anthropic's published rates.
A joint BizOps + VP Product review identified that prepaid Token Banks at 15% deliver better unit economics and UX than 30% metered:
BizOps verdict: Token Banks at 15% delivers better unit economics than 30% metered while feeling cheaper to users.
VP Product verdict: Aligns with brand promise ("your keys, your costs"), protects BYOT positioning, competitively differentiated.
Options Considered
Option A: Prepaid Token Banks at 15% Markup [CHOSEN]
Structure: Dollar-denominated credit packs with 15% markup and tiered bonuses.
Credit Pack
Price
Bonus
Effective Markup
$10
$11.50
None
15%
$25
$28.75
None
15%
$50
$57.50
$5 bonus credits
~4.5% effective
$100
$115.00
$15 bonus credits
~0% effective
Pros:
Eliminates credit risk entirely (cash collected before service rendered)
Reduces Stripe fees (one charge per top-up vs. recurring metered charges)
Better UX: no surprise bills, predictable spending, user always in control
15% feels fair vs. industry-standard 20-40% markups
No expiration is a trust signal and competitive differentiator
Per-task cost preview creates transparency that builds trust
Natural BYOT funnel: power users self-select to zero markup at higher spend levels
Cons:
Lower per-user margin than 30% model (~$2-3/user/mo less on moderate spend)
Requires credit balance tracking infrastructure
Bonuses at $50/$100 reduce effective margin on large packs
No expiration means potential long-tail balance liability
Option B: Daily Credit Card Charges (Rejected)
Structure: Charge users daily for actual API usage at a markup.
Rejected because: Stripe fees explode on daily micro-transactions. A $0.50 daily charge costs $0.31 in Stripe fees (62%). Even weekly billing creates unacceptable fee drag on light-to-moderate users.
Option C: Status Quo 30% Metered Markup (Rejected)
Structure: Monthly metered billing at 30% markup on Anthropic's published rates (the v1.0 recommendation).
Rejected because:
30% markup creates higher BYOT switching pressure (users more likely to get their own API keys)
Metered model introduces credit risk (Legionis fronts API costs before billing)
Monthly metered charges incur Stripe fees on every billing cycle
30% feels expensive to cost-aware users, damaging brand perception
"Surprise bill" UX creates anxiety and churn risk
Rationale
The core insight: prepaid at a lower markup generates better total lifetime economics than metered at a higher markup.
At 30% markup, users switch to BYOT sooner (at ~$50/mo spend, the $15/mo markup fee triggers switching). At 15% markup, the switching threshold is higher (users tolerate $7.50/mo at $50 spend), meaning more users stay on managed billing longer, generating more total revenue over their lifecycle.
Additionally, prepaid eliminates the two biggest operational risks of metered billing:
Credit risk: With metered billing, Legionis fronts API costs and bills later. With prepaid, cash is collected first.
Stripe fee drag: Fewer, larger transactions reduce the impact of Stripe's $0.30 fixed fee per transaction.
What Changes
Immediately
Token economics document: Updated to v2.0 reflecting Token Bank model
Pricing strategy: Updated to v2.0 with Token Bank details in Key Decisions table