SaaS Pricing Strategies: Seat, Usage, Hybrid, and Outcome Models, Price Increases, and Discounts

Pricing decisions come in five parts: what you charge on, which model wraps it, how you package tiers, how you raise prices, and how you control discounts. Each has a way to fail quietly, and the first often decides the rest. This guide covers all five with current examples from vendors' own pricing pages and one useful calculation for each.

It is written for sellers. If you are buying software, the contract negotiation guide covers price caps and renewal terms. Raising customer lifetime value through expansion is in the CLV guide, and reducing cancellations is in the churn guide.

Choose the value metric first

A value metric is what the price scales with. The three questions that pick it:

• Does the customer's value rise as this metric rises?
• Can the customer predict their bill?
• Does your cost rise with it, or does it stay flat while the customer grows?

The last question has become harder in 2026 because AI features carry real per-use costs. Compare four models, with examples from public pricing pages as of September 2026:

Examples

GitHub announced on April 27, 2026 that all Copilot plans would move to usage-based billing on June 1, 2026. Premium request units were replaced by GitHub AI Credits, consumed by token usage at each model's listed API rates. Plan prices stayed put (Pro $10 a month, Pro+ $39, Business $19 per user, Enterprise $39 per user), and the monthly Pro and Pro+ plans include credits equal to their price. Code completions do not use credits. GitHub said the premium-request model was no longer sustainable and that it had absorbed much of the rising inference cost. Its plans page states that one credit equals $0.01.

Salesforce's Agentforce pricing page offers three ways to buy. Flex Credits cost $500 per 100,000, and an Agentforce action takes 20 credits, or $0.10. Conversations cost $2 each. Per-user licensing exists too, with the Agentforce user license at $5 per user per month and requiring Flex Credits. One vendor is running credit, per-conversation, and per-user pricing side by side.

Intercom's pricing page combines a helpdesk priced per seat, from $19 to $132 a month depending on plan, with its Fin AI agent at $0.99 per outcome. It counts an outcome when a customer confirms resolution, does not ask for more help after Fin responds, or Fin completes a workflow, and charges once per conversation. The customer pays nothing if Fin cannot answer or the customer asks for a human.

The pattern across the three: the seat stays for the human part of the product, and a metered unit covers the part that costs money per use.

Why flat seats break with variable cost

Illustration: a $39 seat where the monthly inference cost per user is $4 for a light user, $15 for a typical one, and $60 for a heavy one. The hybrid alternative is a $19 base including $10 of usage, with usage above that billed at 1.3 times cost. These figures are our assumptions.

The flat seat loses money on the heavy user, and the hybrid plan makes $24 on the same user. But the hybrid plan earns less from light users and can look more expensive to the customer, which is a cost worth pricing for. The choice depends on how wide your usage spread is. If your heaviest users cost many times the median, flat seats will not survive it.

Deciding

• Use seats when the cost per user is stable and the product is used by people at similar intensity.
• Use usage or a hybrid when the cost or value depends on volume, and put a committed allowance in the base fee so the customer can budget.
• Use outcome pricing only when both sides can measure the outcome the same way. Intercom counts a customer who does not ask for more help as an outcome, which a customer could dispute.
• Whichever you pick, send alerts before customers cross an allowance, and offer a spending cap.

Raising prices: the break-even math

A price increase pays off if the customers who leave cost you less than the customers who stay pay you. If the price rises by x and a share c of customers leave, revenue is unchanged when c = x / (1 + x). On gross profit, where a lost customer also removes its costs, the break-even is c = x / (x + gross margin).

Illustration: 1,000 customers paying $500 a month, $6,000,000 in annual revenue. A 10% increase with 3% of customers leaving produces $6,402,000, up $402,000. At 5% churn revenue is $6,270,000. At 9.09% churn it is unchanged, and at 12% it falls to $5,808,000, a loss of $192,000.

Some cautions on reading the table:

• It measures the customers you lose over the pricing change, not your normal churn. Compare against your baseline.
• It ignores second-order effects: lower expansion, downgrades to a cheaper tier, and slower new sales. A price rise can hurt new-customer conversion without showing in churn at all.
• Customers on annual contracts feel the increase only at renewal, so the churn shows up over 12 months, not one.
• Buyers are used to increases: Vertice's SaaS inflation index reports 16.4% for June 2026, though it reflects one procurement platform's customers. Buyers who negotiated a renewal cap, as our negotiation guide describes, are protected from your increase up to that cap, so check your contracts before you announce one.

Practical rules: raise prices at renewal with notice in writing, test on new customers first, tie increases to something the customer can see (a new capability, a plan limit), and give your largest accounts a call before the letter. If the increase falls only on light users of a metered plan, your revenue effect depends on their share of the base, so model it by segment.

Packaging tiers

Good-better-best tiers work when each step up adds something that a specific group of customers needs. Three fences hold up better than the rest:

• Scale limits: users, records, projects, or usage allowances, which grow with the customer's value.
• Control features: single sign-on, audit logs, role permissions, and admin controls, which matter to larger companies and cost little to provide.
• Service levels: support response times and uptime commitments, which large buyers will pay for.

Avoid fencing a feature that decides whether the product works at all for a small customer, or your entry tier will not convert. Check plan mix each quarter: if 80% of accounts sit in your middle tier and nobody buys the top one, the top tier is either misplaced or exists only to make the middle tier look reasonable. That is a legitimate use, but you should know which it is.

The Van Westendorp survey asks four questions about a product's price: too cheap to trust, a bargain, getting expensive, and too expensive. The intersections of the answers define a range of acceptable prices. It relies on what respondents say, which usually differs from what they do, so use it to pick a range to test and confirm it with win rates, discount levels, and churn.

Discount governance

Discounts are a price change made one deal at a time, and without rules they drift down. Illustration: 100 deals a year at a $60,000 list price. At an average discount of 15%, the average contract is $51,000 and the year totals $5,100,000. At 25%, the average is $45,000 and the total $4,500,000, a $600,000 gap.

A deeper discount has to win more deals to pay for itself. Going from 15% to 25% off requires 13.3% more deals to leave revenue unchanged, and 18.2% more if the gross margin is 80%, since each extra deal also carries cost. Moving from 10% to 20% off requires 16.7% more deals on that gross-margin basis.

Rules that keep this manageable:

• Approval levels by discount depth, for example the rep up to 10%, a manager up to 20%, and finance or the head of sales above that.
• Trade discounts for something: a longer term, annual prepayment, a larger commitment, or a reference. Ask for it before you concede the discount.
• Record a reason on each discount and review the average by rep, segment, and quarter.
• Keep the discount on the first-year price only, and make the renewal price explicit so the customer does not face a jump later.

When to leave pricing alone

If you have fewer than a few dozen customers, talk to them before you change anything, since the survey and the break-even table need volume to say much. If churn is rising for product reasons, a price change will not fix it. And if your competitors are moving to usage pricing, do not copy them without checking that your own costs and customers behave the same way.

This guide is for informational purposes only and is not legal, tax, or financial advice. Vendor prices and plan terms are as of September 2026 and change often; the margin and cost figures in the illustrations are assumptions, not benchmarks. Test price changes on real customers before rolling them out.