Credits / drawdown model

What is a credits / drawdown model?#

The credits / drawdown model is a monetization model that sits between pure usage-based pricing and fixed subscriptions. Customers prepay a commitment (in dollars or credits), then usage events consume that balance over time.

Instead of billing entirely in arrears like pure pay-as-you-go, you collect cash upfront while still charging in proportion to usage. This can reduce procurement friction (one PO funds a wallet), support multi-meter products, and improve revenue predictability without fully flattening value into seats.

Key definitions#

• Credit: A prepaid unit that authorizes consumption of defined value (for example, 1 credit = 1,000 tokens, 1 API call, or weighted usage units).
• Wallet / balance: The prepaid pool where purchased credits live.
• Commitment: The prepaid amount (in dollars) or prepaid quantity (for example, 10M credits/year).
• Drawdown (burn): The mechanism/rate by which usage deducts credits from the wallet.
• Overage: Usage above committed credits, handled via auto top-up, pay-as-you-go fallback, or throttling.
• Auto-reload (top-up): Automatic credit purchase when balance falls below a threshold.
• Breakage: Purchased credits that expire unused.

Mental model#

"The Prepaid Transit Card"

Think of drawdown like a prepaid MetroCard: you load $20 first, then spend that value as you ride. You are not renegotiating each trip, and you are not locked into a flat monthly amount regardless of use. You keep flexibility on what/when you consume; the seller gets committed cash upfront.

When should you use a credits / drawdown model?#

Decision criteria#

Equations and rules of thumb#

• Recognized revenue: (Credits consumed x price per credit) + recognized breakage
• Effective unit price: $/unit = pack price / units represented by that pack
• Discounting rule: Keep credit-to-usage mapping stable. Offer discounts by changing pack price/commitment size, not by changing what one credit buys.
• Expiration baseline: A 12-month expiry window is common in B2B because it limits open liabilities while remaining customer-reasonable.

Why does a credits / drawdown model matter?#

Credits/drawdown shares risk better than either extreme model:

• Versus pure pay-as-you-go: You avoid full vendor cash-flow risk from postpaid billing.
• Versus pure subscriptions: Customers avoid paying fixed fees unrelated to actual usage.

This makes it useful for products with variable demand and meaningful per-use costs:

• Budget plus flexibility: Customers commit spend once, then allocate it dynamically.
• Improved cash flow: You collect earlier, often before costs are fully incurred.
• Enterprise-friendly buying motion: One commitment can be easier for procurement than many variable invoices.
• Usage alignment: Heavy users fund more value/cost; light users are not forced into oversized fixed subscription plans.

Tradeoff: the model only works when metering, wallet transparency, policy terms (expiry/rollover/refunds), and accounting controls are robust.

Worked example: the discount that wasn't#

A customer signs a $120,000 annual commitment. List is $0.10 per credit; the commit earns a 20% volume discount, so credits are granted at $0.08 — 1,500,000 credits.

Actual consumption lands at 85,000 credits/month, or 1,020,000 for the year:

The customer negotiated 20% off and ended up paying 18% above list for every credit they actually used. Nobody notices this in year one. Everybody notices it at renewal, and the conversation you get is not about price — it is about trust.

What this does to revenue recognition. Under ASC 606 you cannot book the $38,400 of expired credits when the contract is signed. If your cohort history supports a 68% redemption estimate, you recognize breakage in proportion to redemption — roughly $1.47 of revenue per $1.00 of credits redeemed. Estimates get revised: if utilization improves to 90% next year, you reverse breakage you already recognized.

The design lesson. Utilization is the number that decides whether this model builds a durable account or a one-year one. Two guardrails cost nothing and prevent most of the damage:

• Size the commit to observed usage, not to the discount tier the customer wants. If the pilot burns 85,000/month, sell 1,020,000 — not 1,500,000.
• Publish utilization back to the customer monthly. A buyer who watches their own burn rate renews; a buyer who discovers 32% waste at renewal churns or demands the whole overage back.

How do you implement a credits / drawdown model?#

Inputs you need#

• Value metric: The exact usage events you will bill (tokens, API calls, minutes, GB, etc.).
• Unit economics by event: Marginal cost per event/feature (including stress scenarios).
• Usage curves and seasonality: Historical burn by segment to size commitments and alert thresholds.
• Segment and procurement context: SMB vs. mid-market vs. enterprise tolerance for variance, approvals, and budget controls.

Risks and anti-patterns (and fixes)#

References and links#

Sources#

• Lehrskov-Schmidt, U. (2023). The Pricing Roadmap: How to Design B2B SaaS Pricing Models That Your Customers Will Love. Houndstooth Press.
• Howatson, A. (2024). The usage economy: Strategies for growth, smart pricing, and effective technology management. Usage Economy Publishing.
• Nagle, T. T., Hogan, J. E., & Zale, J. (2016). The strategy and tactics of pricing (6th ed.). Routledge.