For years, software pricing was relatively easy to understand.
You paid ₹999 a month.
Or $49.
Or $99 per user.
The bill arrived, you paid it, and that was basically the end of the conversation.
But software businesses are increasingly running into a problem with that model.
The number of people using a product doesn't always tell you how much value they're getting from it.
A two-person company might use an AI tool hundreds of times a day.
A 50-person company might barely use it.
Both companies could technically have the same number of seats, but they're not necessarily getting the same amount of value.
That's where usage-based pricing comes in.
Instead of charging primarily for access, businesses charge according to something connected to actual consumption.
API calls.
Messages.
AI tokens.
Transactions.
Storage.
Documents processed.
Bookings.
Or another measurable unit.
It's not a completely new idea. Cloud infrastructure companies have used consumption-based models for years.
What's changing is how widely the idea is moving into ordinary SaaS products.
And for smaller businesses, there is an obvious attraction:
Pay more when you're getting more value. Pay less when you're not.
But there's another side to the story.
Usage-based pricing can make revenue less predictable, complicate billing and create uncomfortable surprises for customers if the pricing isn't designed carefully.
So is it actually better?
Let's look at a real case.
What Is Usage-Based Pricing?
The simplest way to explain it is:
Your bill changes according to how much you use the product.
Imagine a traditional software product charging:
₹2,000 per month.
It doesn't matter whether you process 100 transactions or 10,000.
The bill is ₹2,000.
Now imagine the company changes the pricing to:
₹0.50 per transaction.
A customer processing 1,000 transactions pays ₹500.
A customer processing 10,000 pays ₹5,000.
The second customer is using the product much more heavily, so the software company captures more of the value it is providing.
That's the basic idea.
In reality, most companies don't make the model quite that simple.
They might combine a base subscription with usage charges, introduce free allowances, create volume discounts or place spending limits around the system.
That's why you'll often hear terms such as hybrid pricing or consumption-based pricing alongside usage-based pricing.
Why Are Smaller Businesses Interested?
For a small company, a large fixed software bill can be difficult to justify.
Imagine a five-person startup that wants to try a new business tool.
If the software costs ₹20,000 every month before the company has generated meaningful value from it, the decision becomes harder.
But if the company can start cheaply and pay more as usage grows, the risk is lower.
That's one of the strongest arguments for consumption pricing.
You don't necessarily have to make the customer commit to their future success on day one.
They can start small.
If the product works, usage increases.
And if usage increases, the software company earns more.
That creates a much more natural relationship between customer growth and vendor revenue.
A Real Example: HubSpot
One of the better-known examples is HubSpot.
According to reporting by TechCrunch, HubSpot originally sold three subscription packages priced between $3,000 and $18,000 per year during its early years.
The company struggled with churn and weak expansion revenue.
In 2011, it introduced usage-based pricing tied to the number of leads customers generated.
The idea was straightforward:
If customers generated more value from the platform, their spending increased.
According to the same report, HubSpot's net revenue retention improved from around 70% to nearly 100% by the time the company went public in 2014.
That doesn't prove that usage-based pricing alone caused the improvement.
Companies change many things as they grow.
But it is a useful example of the underlying idea:
pricing can grow alongside customer success rather than being completely disconnected from it.
A More Recent Example: ProbeStack
A more recent example comes from ProbeStack, a YC-backed observability startup.
The company originally used a seat-based model.
The reported price was $30 per user per month, with limits around data retention and ingestion.
The problem became obvious as some customers started generating dramatically more data than others.
According to the company's published customer story, some power users were ingesting around 40 times the events of the median customer.
Yet the pricing model wasn't increasing proportionally.
The result was a familiar SaaS problem:
heavy users were consuming substantially more resources without paying proportionally more.
ProbeStack introduced a usage-based option alongside its existing plans.
The company says that within the first quarter after launch:
- 19% of seat-based customers upgraded to the usage-based Pro plan
- Expansion revenue increased 34%
- The company avoided a planned six-week billing-layer rewrite
- Two of four anomaly alerts were traced to instrumentation problems rather than legitimate usage spikes
There is an important caveat here.
These are figures published in a vendor customer story, so they should be treated as company-reported results, not independently audited performance.
Still, the example illustrates why the model can make sense.
The Problem With Per-Seat Pricing
Per-seat pricing made a lot of sense when software's value was closely connected to the number of employees using it.
But that's becoming less true for some products.
Consider an AI application.
One employee might send five prompts a day.
Another could send 500.
Both technically occupy one seat.
The software company's infrastructure costs aren't necessarily equal.
The amount of value being generated isn't necessarily equal either.
This becomes even more important for products involving:
- AI inference
- API calls
- Data processing
- Cloud infrastructure
- Messaging
- File processing
- Payments
- Automation
These products have a measurable unit of consumption.
And that makes usage-based pricing much easier to design.
AI Is Making This Even More Important
Artificial intelligence has added another reason for companies to rethink software pricing.
Running an AI model costs money.
The cost can vary depending on the model, amount of computation, tokens processed and other infrastructure requirements.
That makes a simple unlimited subscription harder to manage for some AI products.
A customer who sends 20 prompts a month and a customer running thousands of AI tasks aren't necessarily equivalent from the provider's perspective.
That's why credits, usage limits and consumption-based billing have become increasingly common in AI software.
The broader SaaS market has been moving in this direction for years, but AI has made the economics much harder to ignore.
But Usage-Based Pricing Isn't Automatically Better
This is where I think the conversation gets too enthusiastic.
Paying for usage sounds fair.
But customers also want to know what their bill will be next month.
That's the biggest weakness of pure consumption pricing.
Imagine running a small business where your software bill is:
₹5,000 this month.
Then your business has a particularly busy month.
Suddenly:
₹14,000.
The software might have delivered considerably more value.
But from the customer's perspective, the bill has become unpredictable.
That's not necessarily a good experience.
The Best Models Usually Have Guardrails
This is why many companies don't simply say:
"Pay for every unit you consume."
Instead, they build guardrails around usage.
For example:
Base fee + usage
₹999 per month + ₹0.10 per transaction.
Included usage
₹1,999 per month including 10,000 API calls.
Additional usage is billed separately.
Usage tiers
First 10,000 units: ₹X
Next 50,000: ₹Y
Everything above that: ₹Z
Spending limits
The customer can specify the maximum amount they're willing to spend.
These approaches give the company the economic benefits of usage-based pricing while making bills easier to predict.
The Most Important Decision: What Are You Measuring?
This is where pricing gets interesting.
A company shouldn't choose a usage metric simply because it is easy to count.
It needs to choose something that makes sense to the customer.
For an email platform, emails sent might work.
For an API product, API calls might work.
For cloud storage, gigabytes stored could work.
For a booking platform, bookings might make sense.
For an AI product, tokens are measurable—but they aren't always easy for ordinary customers to understand.
The ideal metric is something that is:
Easy to understand + closely connected to value + difficult to manipulate.
If customers can't understand why they're being charged, the pricing model has already failed.
Why This Can Help Small Businesses
There's a practical benefit here that doesn't get enough attention.
Usage-based pricing can reduce the entry barrier.
A small company doesn't necessarily need to buy a large package just because it expects to grow.
It can start with relatively low usage.
Then increase spending when the product becomes useful.
That can make experimentation easier.
For a small business, that's valuable.
You don't want to spend ₹50,000 testing a product that might save you ₹10,000.
But if you can spend ₹2,000 to test it, the decision becomes much easier.
It Also Gives Software Companies a Better Expansion Path
From the software company's perspective, the model solves another problem.
Imagine a customer starts at ₹1,000 per month.
Under a fixed subscription, they might remain at ₹1,000 for years.
Under a usage-based model, the relationship can grow naturally.
Customer usage increases.
Revenue increases.
The customer doesn't necessarily have to renegotiate a new contract every time.
That's one reason usage-based models are attractive for businesses with a strong land-and-expand strategy.
TechCrunch's analysis of public SaaS companies found that companies using usage-based pricing tended to have stronger growth characteristics, although that correlation shouldn't be interpreted as proof that pricing alone caused the performance.
The Hidden Cost: Billing Infrastructure
There's another side that customers don't see.
Usage-based pricing is harder to operate.
A company needs to accurately track usage.
Then it needs to connect that usage to customers.
Then calculate charges.
Then generate invoices.
Then handle refunds.
Then deal with disputes.
Then make sure usage data and billing data agree.
This becomes particularly complicated when customers have custom contracts.
That's exactly what happened to many SaaS companies that tried to manage usage-based billing manually.
For example, Common Room reported managing hundreds of subscriptions through spreadsheets before implementing dedicated infrastructure for usage-based billing. The company said it had even been failing to bill for some usage-based features because its existing process couldn't reliably track them.
So the pricing change isn't just a pricing change.
It can become a systems project.
The Small Business Lesson
This is where I think smaller SaaS companies should be careful.
You don't need usage-based pricing just because larger companies are adopting it.
If your product doesn't have a natural usage metric, forcing one into the pricing page can make the product harder to understand.
A simple ₹999 monthly subscription can sometimes be better.
The question isn't:
"Should we switch to usage-based pricing?"
The better question is:
"Does customer value naturally increase with measurable usage?"
If the answer is yes, usage-based pricing is worth testing.
If the answer is no, a traditional subscription may still be the better model.
A Practical Example
Imagine you're building an AI document-processing product.
Your old pricing is:
Starter — ₹999/month
Unlimited documents.
That sounds great.
Until one customer uploads 50 documents and another uploads 50,000.
Your infrastructure costs aren't remotely comparable.
Instead, you might introduce:
Starter — ₹999/month
Includes 500 documents.
Then charge a smaller amount for additional processing.
Now the pricing scales with usage.
The small customer doesn't have to subsidize the huge customer.
And the business has a way to grow revenue when customers increase their activity.
But I'd still put safeguards in place.
For example:
"We'll alert you when you've reached 80% of your monthly allowance."
That's a small feature with a big impact on customer trust.
So Why Are Businesses Making the Switch?
After looking at the examples, I think there are four main reasons.
1. Better alignment with value
Customers pay more when they're getting more from the product.
2. Easier entry
Smaller customers can start without committing to an expensive plan.
3. Natural expansion
Revenue can increase automatically as customers grow.
4. Better economics for high-usage products
Companies aren't forced to give unlimited access to customers whose usage is dramatically higher than average.
But there is one major requirement:
The usage metric has to make sense.
My Takeaway
I don't think usage-based pricing means subscriptions are disappearing.
The reality is more interesting.
We're moving toward a mixture of models.
Some products will remain subscription-based.
Some will charge per seat.
Some will use pure consumption pricing.
And a growing number will combine the three.
For small businesses, that flexibility can be useful.
A company doesn't necessarily need to lock itself into a large software bill before it knows whether the product will deliver value.
And for software vendors, usage-based pricing creates a way to grow alongside customers instead of constantly renegotiating what they pay.
But there's a catch.
Good usage-based pricing should feel predictable even when the usage isn't.
If customers understand what they're paying for, can see their usage and receive warnings before unexpected charges appear, the model can work surprisingly well.
If they can't, the same pricing strategy can quickly become a source of frustration.
That's probably the biggest lesson from the businesses that have made the switch:
Don't charge for usage just because you can measure it. Charge for a unit of usage that customers actually understand as value.
And for a small business deciding whether to make the switch, that's the question I'd start with.