A framework by Kyle Poyar

Usage Based Pricing: When to Charge by Use, per Kyle Poyar

Usage based pricing charges each customer for how much they use your product. You set a price per unit, such as a contact, an API call or a credit, and the bill grows with use. A seat price charges a fixed fee for each user instead. Kyle Poyar, who writes the Growth Unhinged newsletter, has worked on five or six switches to usage pricing (Megan and Amy Show, 2025). He says real use is the best sign of the value a customer gets (Product Growth podcast, 2024). He charges by use when the unit tracks that value, keeps growing in each account, and can be measured (Product Growth podcast, 2024).

Our verdict: charge by use if your product does work that does not need more people to log in. Put a seat fee with a usage cap under it, so the first bills stay small.

Cited from real sources 8 min read Written by Published
Best for
SaaS and AI founders deciding whether to charge by use in place of a fee per seat
Price
Free to read. Ask Gavel free.

Not sure your product should charge by use? Ask Gavel free and get what named founders did, with a link to the minute they said it.

Key takeaways

  • Usage based pricing charges for what a customer uses. A seat price charges for each user (Product Growth podcast, 2024).
  • Charge for a unit that tracks the value customers get, grows in each account, and that you can measure (Product Growth podcast, 2024).
  • Customers pay only while they use the product. So your product team now drives revenue (Megan and Amy Show, 2025).
  • To stop surprise bills, Poyar pairs a seat fee with a usage limit (Product Growth podcast, 2024).
  • Aaron Levie would keep a subscription fee under any usage price, so revenue comes back each year (Y Combinator, 2025).

The idea

What usage based pricing is

Every B2B price is built on one unit, which Poyar calls the value metric. It is the main thing that sets how much a customer pays (Product Growth podcast, 2:59). You can charge per seat, by usage, or by the results the customer gets (Product Growth podcast, 3:27).

Seats and usage are the two units most software companies use to start small with a customer and grow from there (Product Growth podcast, 7:23).

Usage fits products whose value does not depend on how many people log in. Poyar's example is Intercom's AI chat product, which answers support chats so a team needs fewer agents. If Intercom charged per login, it would earn less as its product worked better (Product Growth podcast, 8:55).

The models

Four usage based pricing models

Poyar's examples fall into four shapes, from paying per unit to paying per result.

For AI products, a flat fee is risky. A heavy user can run up your AI costs while paying you the same amount (Megan and Amy Show, 18:15). Poyar sees credits as the short-term fix and pay per result as where pricing goes next (Megan and Amy Show, 33:55). An AI agent priced per seat also tells a strange story: if the AI does the work, who is the seat (Megan and Amy Show, 42:22)?

To pick between seats, usage and results for an AI product, read Madhavan Ramanujam's AI Pricing 2x2. For an AI product sold by use with small top-ups, see Matt Swulinski on selling an AI employee.

A worked example

A worked example: a seat fee with a usage limit

Poyar's own example mixes the two units. The basic plan is $10 per seat, with a limit of 500 contacts that each seat can reach in a day (Product Growth podcast, 21:37).

One person can start cheaply and get real use out of the product. A team that later sets up automation, or tries to do ten times as much, hits the contact limit fast. That is a new way of using the product, and it is where the bill grows (Product Growth podcast, 21:59).

The goal is to give people what they need to get started, then have ways to grow the bill that feel natural (Product Growth podcast, 22:10).

When it fits

Usage based vs seat based pricing: when each one wins

Poyar would start by asking customers how they budget for a product like yours and which pricing models they would accept (Product Growth podcast, 12:44). Then he puts each unit through three tests.

  1. 1

    It tracks the value customers get.

    His example is HubSpot. More seats do not mean much more email marketing. More contacts to market to does, so contacts line up with the value (watch at 13:20). For more on picking the unit, read Patrick Campbell's Value Metric.

  2. 2

    It keeps growing in each account.

    A phone company once charged by the minute. Talk time went down while people used more data. So minutes were the wrong unit (watch at 14:24).

  3. 3

    You can measure it and enforce it.

    Some usage only the customer can see. If you cannot track the unit in your own product, you can never launch a usage price. That holds even when usage is the right model (watch at 15:28).

Seats still work for many companies, such as Salesforce and Slack (Product Growth podcast, 8:26). They fit best when people want to bring in their team, as with a shared whiteboard like Miro (Product Growth podcast, 18:32). The trouble starts when the first buyer pays for one seat from their own budget and keeps the account to themselves, so it never spreads (Product Growth podcast, 19:06).

To choose, watch how your product spreads inside a customer. Then ask if your pricing helps that or gets in the way (Product Growth podcast, 19:14).

What it changes

Why usage pricing makes your product team own revenue

With usage pricing, a customer who stops using your product stops paying. Each day they decide again whether to use it (Megan and Amy Show, 8:37). When product work raises usage, the bill rises too. So your product team starts to drive revenue (Megan and Amy Show, 8:56).

If you are considering a usage-based model, you actually have to become more of a product-led organization.
Poyar on what usage pricing asks of a company Watch at 16:25

In practice your product team works on ease of use and on a roadmap that drives more use (Product Growth podcast, 16:43). If the team does not think that way, he says you will not capture the revenue a usage model offers (Product Growth podcast, 17:04).

The problem

How to keep usage bills from shocking customers

The risk with usage is a bill the customer did not expect. Poyar's picture is a bill that starts at $100 a month and then jumps to $10,000, with no idea what happened (Product Growth podcast, 21:01). He wants customers to feel in control of what they spend (Product Growth podcast, 21:08). Four ways he does it:

  1. 1

    Put a usage limit inside a seat fee.

    Start with a seat price and a cap on use, like $10 a seat for 500 contacts a day. The first bill stays small and easy to read (watch at 21:26).

  2. 2

    Call extra use flex or on demand.

    Never call it overage. That word sounds like a penalty for not guessing your usage up front (watch at 16:04).

  3. 3

    Tie credits to what the customer gets.

    Some credits pass your own costs through to the customer. Others are tied to the output the customer sees, and those make it easier to see the return and guess how many credits you need (watch at 27:24).

  4. 4

    Show what each unit is worth to the buyer.

    His example is Twilio, which charges per text message. A rep asks what a missed appointment costs, then shows how a reminder text cuts no-shows. Then the buyer can live with a bill they cannot predict, because more texts mean more value (watch at 28:42).

You should call it flex or on demand usage.
Poyar on naming extra use Watch at 16:04

Switching

How to move existing customers to usage pricing

Poyar's usage pricing playbook draws on the five or six switches he worked on and on talks with about 25 people who made one (Megan and Amy Show, 15:28). It covers sales pay plans and whether unused usage rolls over to the next month (Megan and Amy Show, 15:55).

A growing company often moves from seats to a mix of seats and usage, once it sees customers with few seats and a lot of use (Product Growth podcast, 29:13). At an early stage, he would not change pricing for the loyal first customers at all. Change it for new customers, who just see it as your price (Product Growth podcast, 31:30).

A bigger company has to move its current customers too. Test the plan with customers first, because you cannot change their pricing twice in a short time (Product Growth podcast, 33:27). Then look for the few accounts whose bill would jump the most. If adding usage makes a handful of customers pay 10 times more, expect pushback and lost customers (Product Growth podcast, 34:22).

I'll often want a more custom migration plan for that specific customer.
Poyar on the accounts whose bill would jump Watch at 34:38

One plan keeps their old price for the usage they have today, and they pay only for usage that grows from here (Product Growth podcast, 34:45). Another lets them stay on the old plan as long as they like, without the new features (Product Growth podcast, 35:01).

Trade-offs

Pros and cons of usage based pricing

Pros

Cons

There's actually no perfect pricing model for any business.
Poyar on choosing a model Watch at 0:00

He admits he has called seats dead himself (Megan and Amy Show, 18:47). His example of a trade-off is the flat fee. It is easy to sell, yet it cannot grow with your best customers, and heavy AI use can make some of them lose you money (Megan and Amy Show, 18:03).

Questions

Usage based pricing FAQ

What is an example of usage based pricing?

Twilio charges for each text message sent, which can cost a penny or less (Megan and Amy Show, 28:42). Intercom's Fin charges 99 cents for each support chat its AI resolves (Megan and Amy Show, 34:51).

What is a hybrid pricing model?

A hybrid charges per seat and by usage at the same time. A customer pays more as more people use the product, and more as each person uses it more (Product Growth podcast, 20:20).

When should a SaaS company switch to usage based pricing?

Poyar points to one sign: customers with few seats and a lot of use. That is when a seat company adds usage to its price (Product Growth podcast, 29:13).

The sources

Where Poyar discusses this

Two recordings from 2024 and 2025. We checked each quote on this page against his words at the linked minute on November 18, 2026.

Where experts disagree

Where experts disagree: what keeps usage revenue coming back?

Kyle Poyar

says the product does. With usage pricing, customers decide each day whether to keep using it, which is also whether they keep paying. That makes product work a source of revenue (Megan and Amy Show, 2025). A company that charges by use has to become product-led (Product Growth podcast, 2024).

Aaron Levie

says a fee does. He expects AI products to charge for the amount of work done, yet he would keep some subscription fee under it. A customer can push all of its work through in one year and not come back the next (Y Combinator, 2025).

Pick by how your customers' work arrives. If it comes back every month, such as messages, calls or support chats, meter it and put your effort into the product, as Poyar says. If it comes in one big batch, such as a contract review project, follow Levie and keep a subscription fee under the usage price.

Sources: Poyar on usage and daily use (Megan and Amy Show, 8:27) and on becoming product-led (Product Growth podcast, 16:25). Levie on pricing by the work done (Y Combinator, 20:57) and on keeping a subscription fee (Y Combinator, 22:02).

Stuck on a real call of your own? Gavel answers startup questions with what a named founder did, and links to the minute they said it. Ask Gavel free

New experts every week

Know when the next expert lands.

Gavel adds new operators to the database every week, each one with cited frameworks you can check and a note on where they disagree with the others. You found this page by searching. Get the next one by email instead.

Latest: Ethan Smith on Answer Engine Optimization

One email a week, only when new experts shipped. Unsubscribe with one click. We never sell or share email.

Related frameworks