.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.
What Is the Best Way to Design a Pricing Page for SaaS Products
A SaaS pricing page is often treated as a design exercise. Teams debate card colors, whether to mark a plan as “most popular,” and how many features fit above the fold. Buyers face a more basic question: Can we understand what we will buy, what it will cost, and what will make that cost change?
That question has become harder in 2026. Traditional SaaS still sells seats, access, and tiers. AI products add credits, actions, resolutions, token-backed usage, and outcome charges. A buyer who cannot estimate a first-year budget will either delay the decision or force the sales team into a costly explanation.
Monetizely’s position is clear: the best SaaS pricing page is a self-qualifying buying tool built around one visible primary meter, segment-specific offers, and a credible path from advertised price to total cost. Its job is not to display every product feature. Its job is to make the right buyer confident enough to proceed and the wrong buyer clear enough to opt out.
The page must help buyers answer five commercial questions
A useful pricing page reduces the work required to reach an internal buying decision. The buyer should not need a demo merely to learn whether the product is priced per user, per workflow, or per successful AI result.
The page therefore needs to answer five questions in sequence:
The central lesson is simple: a pricing page must explain the buying model before it explains the product catalog.
“Contact sales” has a legitimate role in enterprise SaaS. It should not be a substitute for explaining how the commercial model works. An enterprise buyer may accept a custom quote for data residency, implementation support, or a large deployment. That same buyer should still know the primary metric, the rough entry point, and the factors that change the price.
Slack provides a disciplined version of this approach. As displayed on September 8, 2026, Slack lists Pro at $7.25 per user per month on annual billing and Business+ at $15 per user per month, while reserving Enterprise+ for a sales conversation. It also explains that billing is based on members active during a 28-day period and that inactive members create prorated credits.
That combination matters. The public page tells a buyer what the product costs, whom it serves, and how the bill moves. Enterprise procurement can still negotiate, but the buyer does not begin in the dark.
HubSpot makes a different but equally useful choice. As displayed on September 8, 2026, its Customer Platform page shows a $7-per-seat monthly starting price for Starter, a $1,300 monthly starting price for Professional with six included seats, and a $4,700 monthly starting price for Enterprise with eight included seats. The page also states the included HubSpot Credits and the credit use tied to certain AI agents and features.
Neither company publishes every enterprise term. Both disclose enough to let a serious prospect decide whether a conversation is worthwhile. That is the standard.
Five linked choices determine whether the page sells the right offer
A pricing page cannot repair an unclear pricing strategy. It can only reveal one. Monetizely’s 5-Step Pricing Framework puts the choices in the order buyers experience them: goals and segmentation, packaging, pricing metric, price points, and operationalization. The sequence, developed in Monetizing Agentic AI, matters because a price is only credible when it follows from a defined customer, a package that fits that customer, and a meter the customer can understand and the vendor can actually bill.
A pricing page should make each of those choices visible.
The table points to a practical discipline: the pricing page is the public expression of the full commercial system, not an isolated marketing asset.
Most SaaS companies default to three cards: good, better, and best. That structure works when the core job is consistent across the market and the main differences are collaboration, controls, support, and governance.
Cursor illustrates the pattern. Its public pricing page, as displayed on September 8, 2026, offers an individual Pro plan at $20 per month, team plans beginning at $40 per user per month, and an Enterprise offer. The team plans add centralized administration, shared context, usage analytics, privacy controls, and SSO. Cursor also explains that plans include a set amount of model usage and that on-demand usage can continue after that amount is consumed.
The logic is sound because the buyer progression is clear:
- An individual developer wants productive coding help.
- A team wants shared billing, visibility, and management.
- An enterprise wants controls over identity, repositories, models, and procurement.
Monetizely’s assessment of Cursor reaches the same point: the core job remains coding, while the buying and governance requirements rise by segment. That is when tiered packaging earns its place.
A tier grid fails when each customer segment needs a different product rather than a deeper version of the same product. A small law firm, a global law firm, and a corporate legal department may all use legal AI, but their workflows, service needs, and buying process can differ sharply. A page that merely puts more features in the top card will not make those offers fit.
In those cases, the pricing page should use a visible core offer and purpose-built modules or a clearly defined enterprise route. The objective is not to force every prospect into three cards. It is to prevent buyers from paying for features they will not use or being denied capabilities they require.
One primary meter makes cost understandable, while limits make it safe
The most important line on a SaaS pricing page is often not the price. It is the unit beside the price.
A company can charge for seats, active users, records, transactions, storage, API calls, workflows, messages, resolutions, or credits. Many can be measured. Far fewer make sense to the buyer. The strongest primary meter tracks the value the customer recognizes while remaining predictable enough to budget and feasible enough to invoice.
A simple decision matrix helps teams test a proposed meter before placing it on the page.
| Test for the proposed primary meter | Score 1-5 | What a score below 4 signals |
|---|---|---|
| A buyer can estimate normal monthly use before signing | The meter will create budget anxiety | |
| The unit rises as customer value rises | Revenue will lag value, or buyers will see the charge as arbitrary | |
| Product data can measure the unit consistently | Billing disputes and manual adjustments will grow | |
| The invoice can explain each charge in one sentence | Customer success will have to translate the bill | |
| The unit protects gross margin at heavy use | The company will be forced into emergency limits or price changes | |
| Sales can describe the unit without a slide deck | Self-service conversion will fall and sales cycles will lengthen |
A meter that cannot pass these tests should not be hidden behind a calculator. It should be changed.
Intercom’s Fin AI Agent demonstrates how an outcome meter can be made understandable. As displayed on September 8, 2026, Intercom lists Essential at $29 per seat per month and charges $0.99 per Fin outcome. Its documentation defines the billable outcomes, limits charges to one outcome per conversation, states that unsuccessful attempts are not charged, and offers usage reminders and hard limits.
The design principle is not simply “charge for outcomes.” Intercom can charge for outcomes because it identifies a specific, observable event and tells the buyer exactly when it counts. A vague claim that an AI agent is priced “on value” does not meet that bar.
AI products need to show whether the human or the output is the anchor
AI has made many pricing pages harder to read because vendors often combine a familiar seat price with credits that buyers cannot connect to work. The answer is not to remove usage. It is to clarify the role each charge plays.
The Agentic Monetization Spectrum, or AMS, helps make that judgment. It assesses an agent on three dimensions: zero-human ability, operational domain, and output-to-cost ratio. Zero-human ability asks how much work still requires a person. Operational domain asks whether the agent performs one task, a function-wide workflow, or work across functions. Output-to-cost ratio asks whether the value produced grows roughly with compute cost or far outpaces it. As autonomy, scope, and value rise, pricing should move away from the human seat and toward a measurable output or outcome.
For pricing-page design, AMS does not dictate a fashionable meter. It tells the operator what buyers need to see.
| Product archetype | Zero-human ability | Operational domain | Output-to-cost ratio | Best primary meter to place on the page | Required page disclosure |
|---|---|---|---|---|---|
| AI assistant used by a professional | Small to medium | Small | Linear to inflecting | Seat or active user | Included AI use, limits, and heavy-use policy |
| Coding agent with human review | Medium | Medium | Inflecting | Seat as the anchor, with usage shown separately | Included usage, spend controls, and on-demand rate |
| Customer-service agent that resolves issues | Large | Medium | Inflecting | Resolution or completed workflow | Exact outcome definition and treatment of failed attempts |
| Broad autonomous business agent | Large | Large | Inflecting to exponential | Action, workflow, or verified business result | Unit definition, attribution rule, and budget controls |
The table explains why one page should not copy another. Cursor’s seat-led design remains credible because a developer is still the central buyer and quality gate. Intercom can lead with outcomes because a successful resolution is visible to both vendor and customer.
HubSpot shows the middle ground. As displayed on September 8, 2026, its Customer Platform combines subscriptions and seats with HubSpot Credits for AI-powered work. The page identifies concrete credit use: 50 credits for a customer-agent conversation resolved, 10 credits for a data-agent smart-property run, and 100 credits for a prospecting-agent recommended outreach for one lead.
That is a defensible architecture because the primary subscription buys the platform and the credits explain incremental AI work. The page would be weaker if it presented credits without naming what each credit-consuming action does.
Enterprise pricing still needs a visible starting point
Enterprise software has real variation. A multinational company may require implementation help, a dedicated environment, a security review, custom data retention, or service levels that a smaller customer does not need. Publishing a single all-in enterprise number can mislead.
Yet “custom pricing” should not be the first and only answer. Enterprise cards should disclose at least four facts:
Salesforce’s Agentforce page offers a useful example of showing the meter even when customers will involve sales. As displayed on September 8, 2026, Salesforce lists $500 per 100,000 Flex Credits, $2 per customer-facing conversation, and $5 per user per month for its Agentforce User License, while explaining that credits are consumed by defined actions and can be tracked in Digital Wallet.
The commercial model may be complex, but the buyer is not asked to guess what is being counted. That distinction protects trust.
The pricing page becomes credible only when the product and invoice agree
A polished page cannot compensate for product rules that finance cannot enforce. If a plan says “unlimited” but the company quietly throttles heavy users, the page will generate support tickets. If the page promises a usage cap but the billing system cannot stop charges, it will create disputes. If sales can discount an offer into a different shape than the page describes, the public architecture loses authority.
Before launch, operators should verify that four things match:
The point is operational, not cosmetic: every phrase on the page should survive contact with the invoice.
Monthly price cards are useful for comparison. Most B2B buyers still need to explain a longer commitment to finance. A simple total-cost view makes the page more credible because it forces the vendor to show what changes as adoption grows.
Consider a support team using Intercom Essential with Fin. At the September 8, 2026 displayed rates of $29 per full seat per month and $0.99 per Fin outcome, the usage component can quickly exceed the seat charge.
| Monthly operating profile | Full-seat cost | Fin outcome cost | Estimated monthly total | Estimated three-year total |
|---|---|---|---|---|
| 10 seats and 250 outcomes | $290 | $247.50 | $537.50 | $19,350 |
| 10 seats and 1,000 outcomes | $290 | $990 | $1,280 | $46,080 |
| 25 seats and 5,000 outcomes | $725 | $4,950 | $5,675 | $204,300 |
The table does not argue against usage pricing. It shows why a pricing page must help buyers model it before procurement discovers the variable portion later.
Operators should treat pricing pages as decision infrastructure
Choose the buyer decision you want the page to enable. Decide whether the page should drive self-service purchase, qualified demo requests, enterprise expansion, or product-led trial adoption. Do not ask one page to optimize all four equally.
Set one executive owner for the commercial story. Product marketing may publish the page, but product, finance, sales, and customer success must approve the same package rules, meter definitions, and upgrade path.
Measure page quality through buyer behavior, not aesthetic preference. Track plan-card clicks, calculator use, trial-to-paid conversion, sales-cycle time, pricing-related support tickets, and the share of deals that require nonstandard explanations.
Review the page whenever the invoice changes. A new credit rule, AI limit, add-on, minimum commitment, or overage policy is a pricing-page change, not merely a billing-system update.
Footnotes
- https://www.amazon.com/Monetizing-Agentic-AI-Handbook-Transformation/dp/B0H7Z13VKJ/
- Slack, official pricing page. (slack.com)
- HubSpot, official Customer Platform pricing page. (hubspot.com)
- Cursor, official pricing page and pricing documentation. (cursor.com)
- Intercom and Salesforce, official AI-agent pricing and billing documentation. (intercom.com)