Billing Model Migration: Test Revenue Before You Switch

Changing a billing model can raise revenue on a spreadsheet and still fail with customers. The useful test is what each customer would have paid, what it cost you to serve them, and whether they could have predicted the difference.

Before moving from seats or a flat subscription to usage, credits, or a hybrid plan, replay actual accounts through the proposed rules. This gives sales, finance, product, and engineering one set of invoices to challenge.

1. State the reason for the change

Pick the problem you need to solve: a few heavy accounts have poor gross margin, smaller customers cannot justify the entry price, or expansion is invisible in a fixed fee. Write down the metric that should improve and the customer behavior you want to encourage. “Increase revenue” alone cannot tell you whether a new bill is fair or durable.

2. Choose a unit customers can verify

A billable unit should connect to value or a real service cost. Seats work when more people mean more value. API calls may be easy to count but misleading when one call costs far more than another. For AI products, tokens, model type, or compute time may better reflect variable cost. Define what counts, what is excluded, and when an event is recorded. Test retries and duplicate events before using the metric on an invoice.

3. Replay three real customer segments

Take at least one new account, one typical account, and one high-volume account. For each, calculate three recent months under both the current and proposed plan. Include base fees, allowances, tiers, overages, discounts, and prepaid credits. Compare revenue and gross margin, but also the largest month-to-month jump the buyer would see.

For example, suppose a customer pays a $1,000 monthly subscription and uses 40,000 units in a normal month, then 100,000 during a launch. A proposed $500 base fee with 50,000 included units and $0.02 per extra unit produces a $500 normal bill and a $1,500 launch bill. The average may look attractive to both sides. The launch month is the trust test: could the customer see the approaching overage, and did their own business gain enough value to justify it? Use your actual unit costs and contracts; these numbers are only a calculation example.

4. Design the shock absorbers

If the replay exposes a surprise bill, change the plan before changing the price page. An included allowance, a volume tier, a prepaid balance, a spending alert, or a negotiated commitment can make growth more predictable. Each has a tradeoff. A generous allowance protects the buyer but can hide unprofitable usage. Prepaid credits improve spending control, but cash collected at purchase is not automatically revenue earned; finance must track the remaining obligation. Lago supports subscription and usage charges together, as well as wallets with paid and granted credits, top-ups, and alerts. See hybrid plans and Credits & Spend Control.

5. Test the operating path

Run the proposed model end to end: send usage, apply the rate and allowance, produce an invoice, correct a bad event, and reconcile the result. Check what customers see during the month and what support can explain after the bill arrives. If a rate change requires re-instrumenting the product or editing customer-specific code, include that cost in the decision.

6. Plan the transition account by account

Identify existing contracts, renewal dates, price protections, and customers who would pay materially more. Decide whether the new model starts with new customers, at renewal, or through an opt-in migration. Give customers a worked example using their own usage, plus clear notice of overages and credits. Watch gross margin, expansion, downgrades, disputes, and usage after launch. A pricing change that improves one month of revenue but teaches customers to suppress useful usage is not a win.

The model comparison belongs in Usage-based vs seat-based pricing. This exercise answers the next question: whether your chosen model works for the customers you already have.