06-reference

mostlymetrics quote to cash part 2 billing

2026-10-08·reference·source: Mostly Metrics·by CJ Gustafson
quote-to-cashbillingusage-based-pricingmeteringentitlementscfo-voicemostlymetrics-series

"Your Complete Guide to Quote to Cash: Billing" (Part 2) — CJ Gustafson (Mostly Metrics)

Series note: this is Part 2 of the 3-part quote-to-cash series. Part 1 ([[2026-10-01-mostlymetrics-quote-to-cash-part-1]]) covered Pricing & Packaging, Quoting, Contracting, and Provisioning & Entitlements. This issue covers layer 5, Billing. Per CJ's own roadmap, Part 3 (next week) covers Tax, Collections, and Rev Rec — layers 6-8.

Why this is in the vault

CJ makes the case that billing broke because the proxy for value broke — pricing used to anchor to a person ("Bill") and now anchors to tokens, credits, and usage events that behave like a currency rather than a headcount — then walks the concrete architecture a company needs before it can even attempt usage-based billing: emit events, meter them, rate them into dollars, invoice, and reconcile into finance. It's a practitioner-level sequencing of a problem RDCO only reasons about informally today.

Mapping against Ray Data Co

The most concrete connection is CJ's "day zero" point under Architecting Usage Based Pricing for Billing: you can't bolt usage billing on after the fact, because the product has to emit taggable, ID'd events from the start, or there's nothing to meter. RDCO already runs AI-gateway-metered infrastructure for Scribble Works (project_scribble_works_ops_rules — "real gateway smoke for model-request changes," CF_AIG_TOKEN via AI Gateway) and treats agent API cost as budget-controlled rather than meter-billed (feedback_api_cost_budget_controlled). CJ's chain — emit → meter → rate → bill → reconcile into finance — is the exact architecture RDCO would have to build the day any Studio product (a paid Scribble Works tier, a per-run agent product) needs to charge by usage instead of a flat cap. Right now RDCO caps spend and never meters it per-customer; this piece names precisely the gap between "we don't overspend" and "we could bill for what we spend," and the gap is architectural, not a pricing-page decision — it has to be designed into the product on day one, not retrofitted. Nothing here overturns existing RDCO practice; it sharpens the entitlements thread from Part 1 ([[2026-10-01-mostlymetrics-quote-to-cash-part-1]]) by showing what sits downstream of an entitlement decision once a product actually monetizes by usage.

The core argument

Billing's complexity grew because the unit of value shifted from a person (a seat) to an event (a token, an API call, a credit) that has no fixed worth — the same token costs differently depending on context, the way a bottle of water costs differently at a grocery store versus a stadium, and output tokens cost more than input tokens because generation is sequential while reading a prompt is parallel. Subscription billing was an invoicing problem that lived inside the ERP; usage billing is a real-time data and pricing problem that doesn't fit there, because billing needs product-level awareness of events as they happen, including late-arriving usage from a metering pipeline that runs in arrears. The required chain, in order: a product must emit well-tagged events; those events get metered (turned into usage records); usage records get rated (turned into dollars); rated charges get billed (turned into invoices); and invoices must flow down into the finance system so revenue can be proven and leakage caught. CJ's practical bar: decide which of those five steps to build versus buy, but decide it before the product ships, not after — because by the time usage-based pricing feels urgent, the event-emission design decision is already baked in or already missing.

⚠️ Sponsorship

Related