Usage-Based Pricing: Pay for What You Use, Not Seats or Tiers

Usage-based pricing charges based on consumption (API calls, storage, compute, messages) rather than fixed subscriptions. Revenue scales automatically with product usage, no contract amendments needed. 60% of SaaS companies now use or experiment with usage-based pricing.¹ Twilio charges per message, Datadog per host, and Snowflake per compute credit.

The power of usage-based pricing is that customers pay per unit of consumption. Start free or cheap, scale costs with value received. No sales friction for expansion:

Common Usage Metrics

The key: usage metric should correlate with value delivered.

When Usage-Based Pricing works

Best Fit Products

Usage-Based Pricing Examples

Twilio: Pay Per Message, No Friction

Start with $20 credit. Pay per SMS, call, or API request. Twilio grows revenue automatically as developer apps gain users, with no negotiations and no contract amendments.²

How It Works

Lessons

1. Start with a low entry point to remove evaluation barriers. Twilio’s $20 starting credit eliminates initial friction. By the time finance notices the charges, the product is already in production and growing.
2. Tie your pricing metric directly to the value delivered. Messages sent, API calls made, these directly correlate with customer success. When customers succeed, Twilio’s revenue grows automatically.
3. Design for frictionless scaling without approvals. Usage grows without procurement involvement or contract amendments. Revenue expansion happens silently in the background.
4. Align your incentives with customer success. When customer success equals vendor revenue, both parties work toward the same goal.

Datadog: Per-Host Monitoring

Per host monitored. As infrastructure grows, so does revenue. Datadog (140%+ NRR) captures 40% of ARR growth from consumption expansion, not new logos.³

How It Works

Lessons

1. Tie pricing to metrics that naturally grow with customer success. Hosts, users, storage. These expand as companies grow. Datadog captures 40% of ARR growth from consumption expansion alone.
2. Make scaling invisible by eliminating contract negotiations. Just add more hosts. No procurement cycle, no approval needed, no friction.
3. Choose metrics where more usage directly means more value delivered. More hosts means more infrastructure to monitor, which means more monitoring value, which means more willingness to pay.
4. Leverage the constant nature of infrastructure growth. Companies rarely shrink their infrastructure. Natural expansion is baked into the business model.

Snowflake: Compute as Currency

Compute separated from storage. Pay for what you query. Snowflake’s consumption model means idle data costs little, but active analytics drives growth: more analysis equals more revenue.⁴

How It Works

Lessons

1. Separate fixed costs from variable costs to reduce idle anxiety. Snowflake separates compute from storage. Customers don’t worry about data sitting unused because storage is cheap. Compute charges only when queries run.
2. Align charges with the activity that generates value. More analysis means more insights, which means more business value. Customers pay when they extract value, not when they store potential.
3. Create a predictable correlation between usage and revenue. When the formula is clear (usage = value = revenue), both sides can plan accordingly.

Frictionless Expansion Is the Real Value

When a customer’s usage doubles, revenue doubles. No sales call. No contract amendment. No procurement approval. Usage-based pricing removes every barrier between customer growth and revenue growth. The magic isn’t fairness; it’s frictionless expansion.

Action Items

1. Name what you’re selling in one word: API calls? Messages? Hosts? Storage? If you can’t name the unit in one word, your customers can’t understand it either. The unit should be something customers already think about. “Credits” doesn’t count.
2. Check the correlation: Does more usage actually mean more value for customers? Plot customer usage against their stated satisfaction or retention. If high-usage customers churn as much as low-usage, you’re charging for activity, not value.
3. Build bill shock protection before launch: Set up alerts at 50%, 75%, 90% of expected spend. Add a spending cap option. Test the experience of hitting a limit. If the first time a customer sees their bill is when it arrives, you’ve lost their trust.
4. Model the downside: What happens if a customer’s usage drops 50%? If your revenue drops 50% too, can you survive that? Run the scenario with your finance team. Consider minimum commitments if the variance breaks your model.
5. Find where customers start optimizing against you: Talk to your highest-usage customers. Ask: “Have you ever tried to reduce your usage to save money?” If yes, you’ve learned where your price is too high. If no, you might be leaving money on the table.