Stripe · Strategy · Senior · 45 to 60 minutes

Should Stripe make token billing customer facing?

Stripe provides payments, billing, and other financial infrastructure for internet businesses. Answer this product management case and get it graded line by line against a weighted rubric.

Pricing strategy
Customer segmentation
Platform strategy
Roadmapping
Systems thinking

The situation

You are a senior PM in Stripe's billing business, focused on usage-based pricing tools for software companies. A fast-growing wave of AI customers and prospects is asking for token billing because their own model providers charge them per token, and Stripe's sales team sees this as an urgent gap. Tokens are easy to count, easy to explain, and useful for controlling runaway compute costs and margins. But some product leaders worry that if Stripe makes token billing the default customer-facing pattern, many AI application companies will end up putting model usage and markups on invoices in a way that weakens their differentiation and makes them look like resellers of commodity models. Others argue Stripe should stay neutral and let customers decide. Assume the team can make one major AI billing bet this planning cycle. Your VP wants a strategy that decides what Stripe should build and position first for AI pricing, who it is for, and how to balance short term demand from customers with a durable point of view on how AI products should charge over time.

What to produce

A 2 page strategy memo.

Your answer must

Recommend Stripe's product strategy for AI billing, including whether token billing should be supported as backend metering, customer-facing pricing, both, or neither, and for which customer segments.
Compare at least three pricing bases relevant to AI, such as tokens, outputs, outcomes, seats, credits, or hybrid models, and explain the tradeoffs for customer value communication, implementation complexity, and business defensibility.
Propose a phased roadmap for the next 12 to 18 months, including what Stripe should build first, what should wait, and any migration path for customers that start with one model and evolve to another.
Define success metrics and leading indicators that cover customer adoption, invoice clarity, and protection of customer unit economics or margins.
Identify the biggest risks and second-order effects of your recommendation, including how Stripe's choice could influence customer positioning and future competitive dynamics.

Constraints

Stripe can meter usage and generate invoices, but it does not control customers' end user workflows or the pricing of the underlying AI models they buy.
The strategy must account for at least two distinct customer types, such as model providers and AI application companies, or clearly justify focusing on one first.
Customers need a way to manage changing model costs over time, even if Stripe does not expose those costs directly on invoices.
Do not assume Stripe can reliably measure or charge on end business outcomes for most customers.

Format

Use headings: Recommendation, Customer segments, Option analysis, Roadmap, Metrics, Risks.

How Stripe actually played it

A Stripe leader argues that token billing is useful backend infrastructure but a bad customer-facing pricing model, because it lets costs define the price. The post says the invoice should state the value delivered, not a markup on model costs. A pointed read for anyone deciding how to charge for AI.

Use usage metering internally, price to value externally.
If price follows cost, your value is just a margin on someone else's model.
The invoice is part of your product positioning.

Built from Why I tried to kill token billing (and why we kept it) by Stripe.