.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.
The Three Pillars of a Winning SaaS Pricing Strategy
Pricing has become harder because SaaS companies now sell several kinds of value at once. A collaboration product may charge for users, an observability platform for infrastructure and data, and an AI agent for completed work. Each choice affects far more than the price page. It shapes which customers can buy, what sales must explain, how finance forecasts revenue, and whether a successful customer becomes more profitable or less so.
The mistake is to treat pricing as a rate-card exercise. A company can change a $20 seat to $30, add a usage fee, or publish a new enterprise tier and still leave its real commercial problem untouched. The deeper question is whether the company has chosen a customer, an offer, and a meter that reinforce one another.
Monetizely's position is clear: a winning SaaS pricing strategy rests on three pillars - a precise choice of target segments, packages built around how those segments buy, and one primary pricing metric that follows realized value. Price points and billing operations matter, but they make those three decisions work; they cannot repair them after the fact.
The three pillars are:
- Choose the customers and business goal before designing the offer.
- Package around meaningful differences in buying needs, not a feature inventory.
- Anchor revenue in a primary meter that customers recognize and the business can run.
Winning pricing starts with a deliberate choice of customer and commercial aim
Every pricing model makes a trade-off. A low-friction self-serve offer can widen adoption but may cap contract value. A custom enterprise offer can capture more value per account but slows the sales cycle. Neither is inherently better. The failure begins when leaders pursue both at once without deciding which segment matters most.
Monetizely’s 5-Step Pricing Framework puts that choice first. The sequence is deliberate: goals and segmentation, packaging, pricing metric, price points, and operationalization. Goals establish what the company needs pricing to accomplish. Segmentation identifies who buys, what job each group is trying to complete, and how needs differ. Packaging then creates offers for those groups; the metric defines what is billed; price points set the actual rates; and operationalization connects the model to product data, quoting, billing, and renewal. As described in Monetizing Agentic AI, the logic matters because a company that starts with price usually ends up debating numbers before it has agreed on the customer or the value being sold.
The practical implication is simple. Management should not ask, “Should we charge per user or by usage?” until it can answer, “Which buyers are we trying to win, and what is the pricing change meant to achieve?” A company targeting 10-person teams needs a different answer from one seeking a small number of global accounts with complex compliance needs.
Consider the contrast between Cursor and Devin. Cursor’s core users - individual developers, teams, and enterprises - share a common job: writing software faster. Their needs diverge mostly in administration, security, purchasing, and governance. Devin also serves individuals, teams, and enterprises, but its lower plans can limit the amount of autonomous work a buyer can test before facing a much larger commitment. The segments may be sensible, yet the path between them can still leave a commercially important middle group without a natural offer.
Before building tiers, leaders should force agreement on three questions:
- Which segment should supply most new ARR over the next 24 months?
- Is the immediate goal faster adoption, higher average contract value, stronger gross margin, or a move upmarket?
- What job does each priority segment hire the product to perform?
A pricing model becomes coherent once those answers are written down. Without them, every internal stakeholder can argue for a different structure while believing they are discussing the same market.
Public pricing pages show that mature SaaS companies do not converge on one fashionable metric. They choose units that fit the product’s role in the customer’s business.
The exhibit makes the central point: strong pricing does not begin by copying a competitor’s meter; it begins by identifying the unit that best represents the customer’s reason to buy.
Most SaaS companies understand that small businesses and enterprises should not receive identical offers. Many still package poorly because they turn the product roadmap into a tiering scheme. Features are sorted into Free, Pro, and Enterprise based on what the vendor wants to withhold, rather than what a segment needs to purchase successfully.
A package should answer a buyer’s practical constraints. A solo user may need immediate access and a credit card checkout. A departmental buyer may need shared billing, usage visibility, and manager controls. An enterprise buyer may require SSO, audit logs, procurement support, data controls, and contractual commitments before the product can be deployed at all.
Cursor illustrates the stronger pattern. Its Free, Pro, Teams, and Enterprise offers preserve the core coding job across segments while expanding the controls needed to buy and manage the product at greater scale. Current enterprise capabilities include pooled usage, SCIM seat management, audit logs, access controls, invoicing, and priority support.
The distinction matters. A company should not force an enterprise buyer into a premium plan merely because that plan contains more product features. Enterprise buyers often pay more because their organization needs a different way to administer, secure, and govern the same core product.
Exhibit 2: Package design should follow the buyer’s purchasing reality
The table below separates package choices that support a real segment from choices that simply create artificial scarcity.
The implication is not that every SaaS company needs many tiers. It is that every package must justify its existence through a distinct buying need.
Jira demonstrates how this can work at scale. Its paid tiers retain the core work-management product, then add higher levels of automation, storage, support, uptime commitments, planning capability, administrative controls, and AI credits. As of September 3, 2026, Jira Standard includes 400 automation steps per user per month, Premium includes 750, and Enterprise includes 1,000.
HubSpot follows a related logic. Sales Hub uses paid seats for people who need full sales functionality, while its credit system lets customers add AI work without redefining the core commercial relationship. That structure recognizes that a sales leader budgets for people, yet AI agents create work that can grow faster than headcount.
Our view is that packaging should create a credible path from first purchase to broader deployment. The buyer should be able to see why moving up is necessary, what additional value it unlocks, and why the new price matches the organization’s expanded needs.
The primary meter must follow realized value, while price and systems make it credible
The third pillar is the pricing metric: the unit a customer sees on an order form and invoice. In classic SaaS, the default was often the named user. That logic remains sound when software improves the productivity of an identifiable person. It breaks down when the product monitors infrastructure, processes large volumes of data, or completes work without a person directing every step.
A primary meter should pass four tests:
Datadog’s model is instructive because technical environments grow through hosts, containers, events, and data, not through software seats. Its billing documentation states that host-based products may use a high-water-mark plan or a monthly commitment with hourly overage, while product-specific units cover containers, custom metrics, devices, and other monitored assets. The meter tracks the thing creating both customer value and service cost.
Intercom Fin sits at the other end of the spectrum. Fin charges for an outcome when it resolves a customer issue, completes a procedure handoff, or performs another defined result. The buyer is not paying for model tokens or agent logins. The buyer is paying when the product does something that would otherwise demand customer-service labor or leave a customer waiting.
Agentic AI moves the meter toward work completed only when the work is measurable
Agentic AI makes the metric decision more acute because the product may shift from helping an employee to doing part of the employee’s job. The Agentic Monetization Spectrum, or AMS, provides a disciplined way to read that shift. It scores an agent on three dimensions: zero-human ability, or how little human involvement remains; operational domain, or whether the agent handles one task, one end-to-end function, or several functions; and output/cost ratio, or how far the customer value of the output exceeds the cost to produce it. As autonomy, domain breadth, and output value rise, a seat becomes a weaker anchor and output or outcome pricing becomes more compelling.
For this exhibit, Small = 1, Medium = 2, and Large or Exponential = 3. The score is directional rather than mechanical; the pattern across the three dimensions matters more than the total.
The scoring supports a firm conclusion: AI does not make outcome pricing universally superior. It makes the choice of primary meter more consequential.
Cursor’s current architecture is especially useful. It maintains a clear seat-based purchase for developers and teams, then adds included usage and on-demand consumption for expensive model activity. The primary commercial relationship remains a seat because the human developer still owns the work and evaluates the output.
By contrast, Intercom can define and count a successful customer-service outcome at the conversation level. Its pricing rules specify that customers are charged at most once per conversation and not charged for unsuccessful attempts. Those conditions make the meter understandable enough to serve as the main economic anchor.
Exhibit 4: A meter is only strategic if the company can operate it
A pricing decision is incomplete until product, finance, sales, and customer success can run it consistently.
The table shows why billing is not back-office work. An unmeasurable or unexplained metric is not a sophisticated pricing strategy; it is a future renewal dispute.
The strongest SaaS pricing models make an explicit choice. They do not use seats because seats are familiar, usage because usage is fashionable, or outcomes because outcomes sound customer-friendly. They choose a primary meter based on the product’s role in the buyer’s work, then use packages and price points to capture differences in willingness to pay.
Monetizely’s position is therefore not “move to usage” or “stay with seats.” It is more demanding: choose one primary meter that reflects the value customers recognize, build packages that match the way priority segments buy, and use secondary charges only to solve a specific economic or deployment problem.
Operators should act on that position now:
Make pricing a board-level growth decision. Tie the pricing architecture to the company’s annual ARR plan, margin targets, and segment strategy rather than treating it as a website or sales-enablement project.
Create a migration policy before changing the rate card. Decide which customers remain on legacy terms, which move at renewal, and which receive a new offer immediately. Installed-base pricing is a portfolio decision, not an exception-management exercise.
Measure revenue quality by segment and meter. Track expansion, discounting, support cost, gross margin, and renewal performance separately for seat-based, usage-based, and outcome-based revenue.
Put pricing review into product-roadmap governance. A new AI capability, workflow, integration, or automation may change what customers perceive as the unit of value. Product strategy and monetization strategy should change together.
Use discount data as a signal of structural weakness. Repeated concessions in one segment often indicate that the package, meter, or buyer definition is wrong - not that sales needs more discretion.
Sources and notes
https://www.amazon.com/Monetizing-Agentic-AI-Handbook-Transformation/dp/B0H7Z13VKJ/
Monetizely, “Goals and Segmentation,” “Packaging,” “Choosing the Right Pricing Metric,” “Finding the Right Price Points,” “Operationalizing Agentic AI Pricing,” and “The Agentic Monetization Spectrum,” accessed September 3, 2026. (getmonetizely.com)
Cursor and Devin official pricing and billing documentation, accessed September 3, 2026. (cursor.com)
Atlassian Jira and HubSpot Sales Hub official pricing pages, accessed September 3, 2026. (atlassian.com)
Intercom Fin AI Agent pricing and outcome definitions, accessed September 3, 2026. (intercom.com)
Datadog official pricing and billing documentation, accessed September 3, 2026. (datadoghq.com)