.png)
Frameworks, core principles and top case studies for SaaS pricing, learnt and refined over 28+ years of SaaS-monetization experience.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
How to Build a Beginners Framework for SaaS Pricing Techniques
SaaS pricing once looked like a choice between a monthly subscription and an annual contract. That simplicity is gone. A single vendor can now sell seats, platform access, usage credits, and completed outcomes at the same time. As of September 7, 2026, HubSpot combines seat-based plans with credits that power AI features, while Salesforce offers Agentforce through per-user licenses, per-conversation pricing, and action-based credits.
For a beginner, the risk is not merely setting a number too high or too low. The greater risk is choosing a pricing structure that teaches customers to value the product in the wrong way. A developer tool sold only by tokens can feel like cloud infrastructure. A customer-service agent sold only by seats can hide the work it completes. A monitoring product sold per employee would make little sense to an engineering leader who manages thousands of hosts.
Monetizely's position is direct: begin with one primary meter that reflects the durable unit of value a buyer recognizes, then use packaging and price points to separate customer needs. Add variable usage only when it protects margins or tracks a clear expansion path; use outcome pricing only when the product can complete a defined unit of work with limited human intervention.
A coherent price begins with the business choice leaders are trying to make
Pricing is not a rate card exercise. It is a sequence of decisions, and the order matters. Monetizely's 5-Step Pricing Framework starts with Goals and Segmentation, moves to Packaging, then Choosing the Right Pricing Metric, Finding the Right Price Points, and finally Operationalizing Pricing. The framework matters because each decision narrows the next one. A company cannot credibly select a meter before it knows which customers it intends to serve, and it cannot sustain a price until its systems can measure, bill, and explain it. Monetizing Agentic AI develops this discipline in greater depth.
Exhibit 1: The five decisions that turn a pricing idea into a working system
The sequence protects beginners from the most common error: starting with a competitor's price and working backward to justify it.
A founder who wants rapid adoption may set a different package and commitment threshold than a mature public company seeking to improve gross margin. Neither goal is wrong. Confusion begins when the product team designs for adoption, finance designs for margin protection, and sales tries to close both motions with one undifferentiated offer.
Segments deserve distinct offers when their buying motion or value differs
Segmentation does not mean creating a separate plan for every customer. It means recognizing material differences in what customers need, how they buy, and what they can reasonably pay. A five-person startup buying a collaboration tool values speed and low commitment. A regulated enterprise may value SSO, audit logs, procurement support, and contractual controls more than another feature in the core workflow.
HubSpot offers a useful current example. On September 7, 2026, its Customer Platform listed Starter pricing beginning at $7 per seat per month on annual commitment, while Professional started at $1,300 per month and included six seats; Enterprise started at $4,700 per month and included eight seats. The company is not merely charging more for more features. It is packaging for different buying motions, operating complexity, and expected scale.
Before drawing plan names or feature grids, test whether a segment is commercially distinct.
- Different job: A sales manager may need pipeline control, while a revenue operations leader needs governance and reporting.
- Different buying process: A self-serve customer can buy with a card; an enterprise buyer may require security review and a negotiated order form.
- Different value ceiling: A workflow that saves one administrator an hour a week cannot support the same price as one that improves a 500-person team's throughput.
- Different cost profile: Heavy AI or data users can create materially higher costs than light users.
Exhibit 2: Packaging should respond to meaningful customer differences
The practical implication is simple: packages should separate customer needs, not display every feature the product team has built.
Slack's public plans reinforce the point. As of September 7, 2026, Slack listed its Pro plan at $8.75 per user per month when billed monthly, while higher plans add administrative, security, and enterprise capabilities. The unit remains the user because collaboration value still centers on people participating in a shared workspace; the package changes as organizational requirements deepen.
The pricing metric is the unit a customer sees on the invoice: a user, host, credit, gigabyte, transaction, case, or completed outcome. It is the most consequential choice in the model because it determines whether revenue grows as customer value grows.
A useful beginner rule is to ask: What does the buyer believe they are acquiring more of when the product expands? If the answer is more workers using a tool, a seat is often the proper anchor. If the answer is more infrastructure monitored or data processed, an operational unit is stronger. If the answer is completed customer work, an outcome can be appropriate.
Exhibit 3: Established SaaS vendors show how the primary meter follows the product
| Product and public pricing as of September 7, 2026 | Primary meter | Why the meter fits | What a beginner should learn |
|---|---|---|---|
| Slack Pro: $8.75 per user per month, billed monthly | User | More participants create more collaboration value | Use seats when adoption by people is the core value driver |
| Datadog Infrastructure Pro: $15 per host per month, billed annually; logs ingestion: $0.10 per GB | Host and data volume | Infrastructure scale and telemetry volume drive use and cost | Use operational units when software observes or processes technical assets |
| Snowflake | Credits consumed for compute and related services | Customers consume compute while loading, transforming, and querying data | Use consumption when the buyer already manages a variable resource budget |
| Intercom Fin: $0.99 per outcome | Resolved customer need or completed procedure | The buyer values support work completed, not AI access alone | Use outcomes when success can be defined and verified |
| Cursor Teams Standard: $40 per user per month, with included usage and on-demand usage beyond it | User, with usage guardrails | Developers buy a work tool, while model use creates variable cost | Keep the user as the anchor when humans remain central to the workflow |
Sources: Slack, Datadog, Snowflake, Intercom, and Cursor official pricing pages and documentation, accessed September 7, 2026.
The pattern is not “move everything to usage.” The pattern is to bill on the unit that best connects a buyer's budget, the product's value, and the supplier's economics.
Datadog offers a strong example of a meter that reflects the operating environment. Its documentation states that host-based products can use a high-water-mark model or a hybrid plan with a monthly commitment and hourly charges above that level. The company records host counts over time because a single static seat count would not reflect an autoscaling cloud environment.
Snowflake makes the same point from a data-platform perspective. Its official documentation defines a credit as the unit consumed while customers use compute resources and notes that credit consumption stops when compute is suspended or idle. For a data team, credits are familiar because they correspond to workload intensity rather than the number of employees with logins.
Autonomy moves agent pricing from access toward completed work
AI changes the metric decision because a product can now do part of a worker's job. The Agentic Monetization Spectrum, or AMS, helps make that distinction visible. It scores an agent on three dimensions: zero-human ability, meaning how little human work remains; operational domain, meaning whether the agent handles a task, a workflow, or work across functions; and output/cost ratio, meaning whether the value of output rises faster than the cost to deliver it. A low score supports a seat or access model. A high score supports a meter closer to output or outcome because the customer is buying completed work rather than assistance.
The AMS does not replace commercial judgment. It prevents a familiar mistake: calling software an “agent” while pricing it as though the human user still performs nearly all of the work.
Exhibit 4: AMS scores indicate how far a product can move beyond seats
| Product or archetype | Zero-human ability | Operational domain | Output/cost ratio | Total | Primary pricing direction |
|---|---|---|---|---|---|
| Cursor Teams coding assistant | 1 - small | 1 - small | 2 - inflecting | 4/9 | Per-user plan, with included usage and on-demand overage |
| Intercom Fin support agent | 2 - medium | 2 - medium | 2 - inflecting | 6/9 | Per successful support outcome |
| Salesforce Agentforce service deployment | 2 - medium | 2 - medium | 2 - inflecting | 6/9 | Per conversation or action credit, with a defined billable event |
| High-autonomy, cross-functional operations agent | 3 - large | 3 - large | 3 - exponential | 9/9 | Contracted price tied to a verified business outcome |
The first three scores are Monetizely assessments based on the products' published pricing and stated roles; the fourth is a forward-looking operating archetype.
Cursor demonstrates why a seat can remain the primary meter even in an AI-heavy product. On September 7, 2026, Cursor listed Teams Standard at $40 per user per month, while its pricing materials also described included model usage and on-demand billing after that usage is consumed. The developer remains accountable for the code, reviews the output, and directs the work. The seat therefore remains the commercial anchor, while usage protects the company's variable model costs.
Intercom Fin sits further along the spectrum. Its pricing page listed $0.99 per outcome as of September 7, 2026. An outcome occurs when the customer confirms resolution, does not seek more help after Fin responds, or Fin completes a workflow. Intercom charges no more than once per conversation. That definition is not a marketing flourish. It is the operational rule that makes outcome pricing credible.
Salesforce provides a useful contrast. Its Agentforce pricing page listed $2 per conversation and $500 per 100,000 Flex Credits, with one action consuming 20 credits, or $0.10. A conversation is easier to observe than a full business outcome; an action is even easier to count. Those choices make sense when customers want a clear usage record but the vendor cannot yet contract around a universally accepted definition of completed work.
Price points should test value capture before they protect every edge case
Price comes fourth because a price point cannot repair a weak segment definition, a confusing package, or a poor meter. Once those choices are set, rate setting becomes a disciplined question: how much value is created, what alternatives does the customer have, what cost does the supplier carry, and what commercial goal takes priority?
A beginner should use competitor prices as reference points, not instructions. Intercom's $0.99 outcome price, for example, may look low until a customer calculates the annual spend at scale.
Exhibit 5: Outcome pricing requires an explicit volume conversation
| Successful Fin outcomes per month | Annual spend at $0.99 per outcome |
|---|---|
| 5,000 | $59,400 |
| 20,000 | $237,600 |
| 50,000 | $594,000 |
The table shows why a buyer needs spend visibility and why a vendor needs a clear volume commitment before calling outcome pricing “simple.”
Rate-setting should answer four questions in order:
A support agent that resolves a low-value FAQ and one that completes a complex insurance claim may produce the same “resolution” count but very different value. The answer is not to abandon the outcome meter. The answer is to define the outcome narrowly enough for the initial use case, then package more complex workflows separately.
A price model only becomes real when product events, sales terms, invoices, and customer reports use the same definitions. Cursor's August 21, 2026 pricing policy, for example, defines an active user for enterprise true-ups by user access during the final 30 days of a true-up period. That rule turns “active seat” from a sales phrase into an auditable contractual measure.
Every variable or outcome-based offer needs three operating controls:
- A billable-event definition: State precisely what counts, what does not count, when an event is reversed, and who can dispute it.
- Customer spend visibility: Show usage, included volume, remaining balance, and projected charges before the invoice arrives.
- A finance-grade record: Retain the event data, applicable rate, discount, and contract rule needed to reconcile the invoice.
Datadog's host-metering rules demonstrate the value of precision. Its published high-water-mark method measures hosts hourly and excludes the highest 1% of readings to reduce the effect of short spikes. Customers may still debate price, but they can understand the rule.
Beginners gain leverage by making one durable commitment
Monetizely's position is not that every SaaS business should copy a consumption leader or rush into outcome pricing. The stronger move is to make the commercial promise visible in the primary meter. A collaboration product should start from the people collaborating. An infrastructure product should start from the infrastructure managed. An autonomous support agent should start from verified customer work completed.
- Choose a primary meter that the executive buyer can explain in one sentence to finance. If the explanation requires a product demo, the meter is too obscure.
- Set a 12-month rule for changing the model. Do not redesign pricing after every customer objection; change it only when segment evidence, usage data, or economics show that the primary meter is failing.
- Separate core SaaS value from AI cost exposure. Keep the main promise legible, then use included credits, committed usage, or overage limits to manage variable cost.
- Treat pricing authority as a cross-functional operating decision. Product, finance, sales, and customer success should share one definition of the meter and one source of truth for its data.
- Measure renewal behavior by package and meter, not only by total ARR. A growing account can still be poorly priced if customers repeatedly downgrade, dispute usage, or need exceptions to renew.
Footnotes
- https://www.amazon.com/Monetizing-Agentic-AI-Handbook-Transformation/dp/B0H7Z13VKJ/
- HubSpot and Slack official pricing pages, accessed September 7, 2026. (hubspot.com)
- Datadog and Snowflake official pricing pages and billing documentation, accessed September 7, 2026. (datadoghq.com)
- Cursor official pricing pages, documentation, and pricing policy, accessed September 7, 2026. (cursor.com)
- Salesforce Agentforce and Intercom Fin official pricing pages and documentation, accessed September 7, 2026. (help.salesforce.com)