AI usage pricing requires metering usage at scale, and billing accurately based on that usage. That created a new infrastructure category: usage-native metering and rating platforms such as Metronome, m3ter, Orb, OpenMeter, and Togai.

Zuora, Kong, Stripe, Salesforce, and Adyen have all moved to acquire that layer that turns product consumption into billable revenue. The meter is getting absorbed into the systems that already own the revenue workflow: payments, CRM, API infrastructure and subscription billing. The pace has accelerated.

Why metering tools exist

Billing AI usage has two hard parts: metering and rating.

Metering is observing and recording raw usage events. Events are API calls, tokens processed, GB transferred or GPU-second consumed. That sounds simple until you are ingesting millions of events per minute in real time across thousands of customers, with full auditability for every one.

Rating takes those counted events and applies your pricing logic to them, converting usage into numbers, e.g.

• Tiered limits that reset monthly

• Overage charges once included balances are drawn down

• Minimum commitments that floor the invoice

• Volume discounts that kick in at thresholds

• Custom contract terms that override standard pricing

The two functions are separated by design. You can meter perfectly and rate wrong. You can have correct pricing logic but incorrectly meter events upstream. B2B contracts make this specifically hard for three reasons:

The scale problem: AI companies’ token consumption has orders of magnitude more events than what SaaS billing systems were designed to handle. Think of metering millions of events per minute vs. thousands of seats per month.

The flexibility problem: Usage-based pricing models iterate constantly. Rob Litterst from PricingSaaS reported on average more than 3 pricing changes per company in 2025, with fast-growing companies making multiple pricing changes per month. Each change needs to flow through the rating engine without an engineering sprint.

The auditability problem: Revenue recognition requires a complete event ledger. Every metered charge needs a traceable lineage from raw event to billable line item. If you cannot defend a charge when a customer disputes an invoice, you will lose trust.

A small set of purpose-built companies emerged to solve this. They became the infrastructure layer for billing AI usage. Now larger platforms started acquiring that capability rather than building it.

The M&A wave

Five acquisitions, five different acquirers, the same underlying logic: a platform that already owns part of the revenue workflow acquires the metering and rating layer it was missing.

Adyen + Orb (announced June 11, 2026)

Adyen already serves customers like Uber and Spotify. Orb’s customers include Perplexity, Replit and Vercel. Orb lets engineering teams define billing metrics directly with SQL queries, simulate pricing changes against historical usage data before deploying in production, and report on usage in real-time. Orb CEO Alvaro Morales on the rationale of bringing both together:

“Standalone billing systems are fundamentally limited because they operate blind to transaction execution … By joining forces with Adyen, we can connect this ingestion layer directly to real-time financial health signals, closing the loop between billing logic and payment success.”

Prior to its acquisition by Adyen, Orb had raised $44M, including $25M from its Series B in September 2024.

Salesforce + m3ter (announced June 8, 2026)

Salesforce already owns the CRM, CPQ, and increasingly the agentic workflow layer. Without a metering engine, they couldn’t close the loop on consumption-based AI monetization. m3ter gives them enterprise-grade rating inside Revenue Cloud.

Salesforce framed it directly: acquiring m3ter to

“bring high-volume mediation, metering and rating capabilities natively to Agentforce Revenue Management, enabling enterprises to launch, track, scale, and bill with the flexible usage and outcome-based pricing models needed for the AI era.”

m3ter had raised $31.5M before its acquisition. Latka estimated the deal cost Salesforce around $140M.

Stripe + Metronome (completed January 2026, reported ~$1B)

Metronome’s customer list reads like an AI infrastructure directory: OpenAI, Anthropic, Confluent, and Nvidia. Stripe CEO Patrick Collison was plain:

”Metered pricing is the native business model for the AI era.”

Metronome had raised $128M, including a $50M Series C just months before the deal. Stripe reportedly paid around $1B, ~2.1x Metronome’s last $470M valuation, per Sacra.

Kong + OpenMeter (acquired September 2025)

Kong is the API gateway and already sits in the traffic path for API calls. Adding OpenMeter means Kong can now translate that traffic directly into billable events. The API gateway becomes part of the monetization system.

“OpenMeter’s capabilities will be integrated into Kong Konnect, enabling usage-based pricing, entitlements, and invoicing for APIs, events, and AI workloads.”

OpenMeter had raised $3M prior to the acquisition.

Zuora + Togai (acquired in May 2024)

Zuora has owned subscription billing for years but was architecturally weak on usage.

“Togai’s low-code builder and developer-friendly interface to quickly configure metering and rating of raw events” filled the gap without Zuora having to rebuild its core engine.

Zuora acquired Togai for $24.3M. Togai had raised $3.1M prior to its acquisition.

The pattern across all five:

Five different players, all move into usage metering and rating:

• Orb → Adyen = Payments + Billing + AI consumption metering

• m3ter → Salesforce = Enteprise CRM + AI consumption metering/rating

• Metronome → Stripe = Payments + Billing + AI consumption metering

• OpenMeter → Kong = API gateway + API usage rating

• Togai → Zuora = Subscription monetization suite + usage rating

Who’s left standing

Not everyone got acquired. Two notable independent players are still building as standalone companies, and their positioning reveals what they’re betting on.

Sequence goes wider than pure metering. It positions itself as a CPQ + billing + usage metering + revenue recognition + receivables automation platform. Their customers include Cognition, Legora and 11x. Their bet is fast-scaling AI companies need one platform for the entire workflow from contract signature to accurate billing. Sequence raised $38M from investors like 645 Ventures and a16z, including $20M from a Series A in December 2025.

Lago is the open-source bet. Customers include Groq and Mistral.ai. It allows AI labs to self-host and control their billing infrastructure. Lago’s differentiation is deployment flexibility: fully managed cloud or on-premises, with full access to the underlying data. For companies with data residency requirements or platform independence as a strategic priority, Lago is the natural alternative to the lock-in that comes with being inside Stripe’s or Adyen’s billing stack. Lago raised $22M from investors like FirstMark, SignalFire and Y Combinator. Its Series A valued it at reportedly $100M.

What this means for pricing strategy

For Pricing teams, metering is becoming a platform capability. Instead of having to buy a separate metering tool, Pricing teams can leverage their existing CRM / Billing / API gateway. That might make it a little easier to secure a budget for metering.

For founders building on AI, billing infrastructure is now a platform alignment choice. The vendor that owns your metering has full visibility into your unit economics: consumption patterns, pricing model structure, which customers have custom contract terms, where you draw overage thresholds. The vendor that owns your metering owns your pricing intelligence. That leverage matters the next time you’re renewing with them.

For the remaining standalone players, Lago and Sequence, a strategic acquisition looks like the most plausible path. The obvious buyers are platforms that already own parts of the revenue workflow but have not yet acquired a usage-native metering and rating specialist: Microsoft, ServiceNow, HubSpot, or larger billing incumbents like Chargebee, BillingPlatform, and keylight. Those billing incumbents already support subscription, usage-based, and hybrid models, but as AI monetization keeps pushing toward event-native metering, real-time rating, and constant pricing iteration, they may still decide that owning deeper metering infrastructure matters.

The window for those acquisitions is closing. Once the category fully consolidates, founders building AI products will have to commit to a wider platform when choosing their metering and rating engine.

Keep Reading

Billing engines from Metronome, Stripe and Charbee made it easier to bill customers based on actual consumption. But just because you can doesn’t mean you should.

Customers have come to hate the unpredictable AI usage bills. They need budget certainty. They want adoption without the risk of runaway spend. And they do not want high usage bills without clear outcomes.

Token consumption varies on top of token pricing

Tokens are still relatively new to customers, and companies are figuring out how many they need and how to control usage. On top of that, token consumption varies by vendor and model, making token prices hard to compare.

This study from March 2026 found that the cheapest AI model on the pricing page is not always the cheapest in production, because token consumption varies among models. Though Gemini 3 Flash’s listed token price is 80% cheaper than GPT-5.4’s, in their test the cost across tasks was 38% higher.

Not knowing how much tokens a particular action or outcome takes for each model makes understanding token usage and total cost hard. Putting the risk of varying token consumption on the customer when models don’t deliver the expected outcome consistently (because they are not deterministic) makes this problem worse.

Customers need predictability

And companies are struggling with unexpected AI usage bills:

Uber CTO Praveen Neppalli Naga acknowledged in April 2026 that he had already spent his AI budget for the entire year, 4 months into the year:

Axios reported a company even spent $500 million in a single month on AI usage. They had not set usage limits, and usage went unexpectedly high.

Cursor faced a backlash after changing its Pro pricing in the summer of 2025. Their CEO Michael Truell had to publish a pricing clarification, acknowledging “we missed the mark” and offer refunds to customers who had a “surprise usage bill.”

OpenAI CEO Sam Altman said in May 2026 “customers are increasingly asking us for certainty on capacity”, and that OpenAI is now selling “discounted tokens for 1-3 year commits”. He sees it as “a big win-win” since it also helps OpenAI plan. He’s right.

AI companies need growth

He hit the nail on the head. AI vendors need predictability to plan too. They need to fund huge capital investments in the billions and justify forward-looking revenue. That’s harder if you don’t have ARR. Anthropic reported in April 2026 that they have surpassed $30 billion “run-rate revenue”, up from $9 billion at the end of 2025, which justifies their $50 billion investment in computing infrastructure.

Until now though, AI companies focused on glorifying token consumption (not recurring commitment) to fuel growth and have been great at this:

OpenAI started giving physical awards in 2025 to companies crossing 10B, 100B and 1T tokens last year. The Tokens of Appreciation program was so successful that OpenAI increased thresholds for 2026 by 10x to 100B, 1T and 10T.

That was the start of token-maxing: employees now use AI tokens to show they leverage AI more than others, that they are the ones that lead innovation.

A Meta employee even created a dashboard so coworkers can compete to become the company’s highest token consumer.

So…

AI vendors want usage to keep growing and are great at promoting it.
Customers need the spend to be predictable.
That tension is why pure pay-as-you-go billing will break down in B2B AI.

Subscriptions don’t have to be per seat

Usage pricing and usage billing are not the same thing. Pricing defines the logic of value, what customers pay for. Billing defines the rhythm of cash, when customers pay.

Companies and CFOs are used to SaaS subscriptions that are priced per seat and usually paid annually or monthly. If the pricing metric changes from users to usage, that does not mean billing has to become pure usage-based billing. For example

• If your pricing metric is usage, contract capacity.

• If your pricing metric is outcomes, contract outcome volume.

• If your pricing metric is credits, sell recurring credit subscriptions.

You can still contract commitments on AI usage, capacity, tokens, credits, outcomes, or resolutions. Then bill that commitment as a subscription.

I wrote more about this distinction in Untangling billing and pricing.

AI is moving away from pure pay-as-you-go

This isn’t just me saying this might happen. The shift is already underway:

AWS sells Bedrock Provisioned Throughput alongside its on-demand pricing since September 2023. AWS has also sold EC2 Reserved Instances since at least 2009.

Microsoft started selling Azure OpenAI Service Provisioned Reservations in late Summer 2023, alongside its hourly no-commitment purchasing.

Google started selling Provisioned Throughput for its Vertex AI in June 2024.

Salesforce introduced Agentic Enterprise License Agreements (AELA) in late 2025, offering flat rate, unlimited usage pricing for its AI agents.

Zendesk has priced AI Agents based on Automated Resolutions since November 2025. Customers can choose to commit to volume at $1.50 per resolution, or pay a higher price of $2 per resolution without commitment.

OpenAI announced it is introducing Guaranteed Capacity on May 19, 2026, saying:

Even if AI pricing stays usage-based for the foreseeable future, AI billing is moving toward contractual commitments and predictable recurring revenue.

B2B contracts are a Win-Win

For customers, commitments help with:

• Budget certainty: CFOs prefer predictable spend over volatile usage bills that add another risk to EBIT.

• Internal planning: Teams can forecast AI spend, allocate budgets, and avoid monthly surprises.

• Adoption pressure: Longer commitments create a commitment effect: “We already bought it. Let’s use it.”

• Commercial leverage: Larger commitments can justify better rates, capacity guarantees, stronger support, or roadmap access.

For AI providers, they help:

• ARR: Usage becomes contracted recurring revenue, instead of uncertain pay-as-you-go consumption revenue.

• Better forecasting: Finance can model revenue, capacity needs, GPU commitments, and gross margin with better confidence.

• Infrastructure funding: Long-term customer commitments can support long-term compute commitments.

• Lower churn risk: Every renewal is a chance to reconsider competitors. Longer terms reduce evaluation frequency. And after 1 to 3 years, status quo bias works in your favor.

So what’s next

Recognize that pricing and billing are two separate decisions.

Usage-based pricing is fine. But don’t bill customers against a metric they can’t predict, at rates that vary by model, for outcomes that aren’t guaranteed. Don’t put all the risk on the customer while you capture all the upside of growing consumption, while eroding the trust you need for a long-term relationship.

Pick the pricing metric that’s right for your business. But let customers contract a commitment against it, and bill that commitment as a predictable subscription. Give them visibility and controls to manage their spend without a mid-year crisis.

That way AI vendors get ARR, and customers get predictability. And AI moves from a line item CFOs fear to a partnership that deepens over 1-3 years.

Subscribe now

Keep Reading

This builds on the classic price waterfall from Managing Price, Gaining Profit, which showed how companies leak revenue from list price, invoice price, to pocket price. It’s a classic and useful, but with AI we need to consider variable costs.

For years, the standard SaaS price waterfall was:

List price → discounts → net price:
Pricing Teams were focused on determining the list price and questioning standard discounts and sales concessions. But with AI, one more step is critical:

Net price → variable cost → contribution margin:
A customer can look attractive at the net price level and still be margin dilutive once you account for AI inference.

The AI Price Waterfall

The AI price waterfall puts variable cost in perspective to the discounts sales is already used to from B2B SaaS and AI pricing.

List Price: The standard published price before any discounts, concessions, fees, or customer-specific adjustments are applied.

• Packaging Discounts increase product adoption or deal size by incentivizing customers to buy multiple products, modules, bundles, or the full suite together.

• Volume Discounts reduce the effective unit price as customers commit to higher seat counts, usage levels, transaction volumes, or enterprise-scale deployments.

• Commitment (Billing & Contracting) Discounts incentivize longer contract terms, annual billing, or prepayment upfront. This article is a deep dive on these.

• Channel Discounts (or fees) compensate intermediaries such as resellers, MSPs, or marketplaces for selling, managing, or distributing your product.

• Discretionary Discounts extend deal-specific discounts selectively to win early adopters, match competitors, or secure strategically important accounts.

= Net Price: The actual realized customer price after all discounts, credits, rebates, channel margins, and negotiated concessions.

Variable Cost of Revenue are the variable costs to deliver your product such as AI inference / usage, cloud hosting, third-party APIs, or data acquisition and processing.

= Contribution Margin: The revenue remaining after subtracting variable cost of revenue, used to assess deal profitability and discount headroom.

Why This Matters More for AI

Notice the waterfall does not stop at the net price. It continues through variable cost of revenue and ends at contribution margin.

Traditional SaaS companies ignored that step, because variable costs were generally low. They had high gross margins and relatively predictable cost to serve. That made list price determination and discount management the two main pricing problems.

AI changed that. Two customers paying the same net price can generate very different margins. One customer is a light user with predictable workflows. Another runs high-volume inference, relies on expensive frontier models, and triggers APIs often. If both customers get the same discount, they look similar in ARR but very different in contribution margin. Your pricing strategy needs to account for that, or you end up subsidizing your heaviest users that get the most value.

A better waterfall shows where price is being given away and where margin is being eroded. Net price is no longer the finish line. For AI products, it is the midpoint. The waterfall helps pricing teams see whether a pricing strategy works after discounts AND usage costs.

Waterfall Template for Google Sheets

I made you a simple Google Sheets template for this: the list-to-net price waterfall template. It has two tabs: one with instructions, the other with the waterfall.

Make a copy and use it as you wish. Use it to compare customer segments, align your leadership on which discounts are creating value versus eroding margin, or to simply compare deals. My goal is to make it easier for you to do your job and show:

List price → discounts → net price → variable cost → contribution margin

If you share it, please just be so kind and mention where you got it from. 🙂

Each team needs incentives that reward revenue quality, retention discipline, value realization and focus. So how do you set these incentives for your teams?

Sales: Prioritize quality revenue, not just more ARR

Most sales incentive plans still overweight one metric: new ARR.

That is simple to manage, but it misses an important point: not all revenue is equally valuable. A deal closed with heavy discounting or poor pricing discipline may help bookings while hurting long-term economics.

BCG argues that pricing metrics should be part of Sales Incentive Plans (SIPs), yet fewer than 10% of SIPs actually include any price-based metrics. That means 90+% still focus on New ARR with no indicator of revenue quality.

Two better metrics you should consider:

Price realization measures how well reps hold price versus list or target price.

• Pro: keeps product cost and margin confidential

• Con: does not directly steer portfolio mix or favor higher margin products

Margin performance measures the profitability of what is sold.

• Pro: directly connects rep behavior to the company bottom line

• Con: exposes margin and cost data broadly, which may change over time

Rep performance can also be evaluated in different ways:

• Absolute performance

• Improvement over time

• Hybrid model combining both

If you want more profitable growth, don’t pay Sales as if every dollar of ARR is equally good. Set targets that align with your strategy and consider price realization or margin performance targets. For most companies, price realization is the easier starting point because it improves discipline without exposing sensitive margin data.

Customer Success: Separate retention from expansion

The same logic applies after the initial sale. For years, many CS teams were measured primarily on NRR, which combines retention and expansion into one number.

Move from NRR to GRR to isolate retention

That worked well in an era when expansion was abundant and growth the overriding goal. The TSIA State of Customer Growth and Renewal 2025 reports that

“The days of unchecked expansion have been replaced by a strategic focus on profitability, forcing companies to rethink how they allocate resources, manage renewals, and drive expansion.”

And GainSight’s Customer Success Index 2025 shows that companies are separating those motions and putting more emphasis on GRR:

• 62% of CS orgs use GRR as their primary metric (up from 45% 2 years ago)

• CS orgs using NRR as their primary metric declined from 59% to 51%

NRR is still valuable, but it can mask underlying retention problems when expansion is strong. GRR makes customer health and renewal performance easier to see. It answers a more disciplined question: are we keeping the revenue we already earned?

Companies are adopting a more disciplined approach to financial management, moving beyond rapid expansion to focus on sustainable, long-term gains.

Companies centralize renewal ownership in CS

Once companies separate retention from expansion, the operating model usually changes too. Renewal ownership becomes more specialized, and companies start asking whether expensive AE capacity should be used for every renewal motion.

They centralize renewal ownership within Customer Success rather than leaving it with higher-cost AEs. According to TSIA, shifting renewal and upsell responsibilities from AEs to specialized CS renewal teams can lead to:

• ~2/3rds lower renewal costs

• 10% higher net renewal rates

• Nearly 10% higher attached upsells

AI reinforces the efficiency play

AI adoption in CS reflects the same shift toward productivity and disciplined growth. The highest adoption use cases of AI in CS focus on efficiency:

• Auto summarization (82%)

• Drafting email/follow-ups (71%)

• Sentiment analysis (54%)

• Churn or risk prediction (39%)

• Next best action recommendations (30%)

These are not “AI for AI’s sake” use cases. AI is being adopted first where it can reduce manual work, improve consistency, and help teams manage retention at scale.

The incentive lesson is that CS teams should not be rewarded only for coverage or activity. You should reward them for scalable retention outcomes: lower renewal cost, faster risk identification, and better GRR.

Product: Value creation trumps feature velocity

Product teams used to be rewarded for shipping: roadmap delivery, launch volume, and shipping velocity.

In the AI era, incentives for Product teams need to shift to reward adoption, usage, retention, expansion and creating value to increase willingness to pay. There is a large gap between what teams build vs. what customers actually use.

The Product-Led Alliance’s The State of Product Analytics Report 2025 shows a similar shift towards value realization:

• Customer feedback is increasingly used to decide on new features (44% → 49%)

• Retention has overtaken revenue as the #1 goal for analytics programs

Role of AI: Define use cases with ROI & set targets

The same principle applies to AI investment. McKinsey’s State of AI in 2025 found that while 88% of companies use AI, only 6% get 5+% EBIT impact. “AI high performers” are more disciplined. They set growth and/or innovation as an objective of AI efforts, define use cases, redesign workflows and align leadership.

They follow 10 best practices across 4 areas:

Use cases and leadership alignment on value:

- Have defined a road map with specific AI initiatives and use cases across priority business domains, aligned with broader AI strategy

- Top leaders understand how AI can create value for the business

Rapid iteration:

- Have an agile product delivery organization or an enterprise-wide agile organization with well-defined agile team delivery processes

- Have an established process for building AI solutions and iteratively improving them (eg, guardrails, approach to development)

- AI efforts process quickly and are adaptive (ie, characterized by quick decision-making and iterative learning

Require Processes and role model:

- Embeds AI solutions into business processes effectively (eg, changing frontline employees’ processes, creating users interfaces)

- Senior leaders are actively engaged in driving AI adoption, including role modeling the use of AI

- Have developed a clear workforce plan (for technology and nontechnology roles) that incorporates the anticipated changes from AI

Frontier tech with human in the loop:

- Technology infrastructure and architecture allow implementation of core AI initiatives using the latest technologies

- Have defined processed to determine how and when model model outputs need human validation to ensure accuracy

Scaling agents are downstream of process redesign. High performers are almost 3x as likely to fundamentally redesign workflows and push for transformative change. They also scale use of agents 3x faster though, and get >5% EBIT impact.

The overall lesson

Incentives drive behavior. Behavior shapes the quality of your growth.

If you pay Sales only on new ARR, you incentivize discounting.
If you measure CS on NRR, early signs of churn can get masked.
If you manage Product by velocity, you incentivize features regardless of value.
If you experiment with AI without clear use cases, you are not going to see returns.

Profitable growth requires that you set incentives that reflect your goals.

Reward Sales for acquiring high-quality revenue.
Reward CS for protecting existing revenue.
Reward Product for prioritizing value creation.
Focus AI investments on measurable use cases with clear ROI.

Then your incentive system rewards profitable growth.

In SaaS, companies defaulted to per user pricing because seats were easy to explain, tracked value well enough, and variable costs were manageable. AI changed that. Costs are real, usage is often concentrated in power users, and outcomes can be attractive but hard to attribute.

In AI, the pricing metric often matters at least as much as the price point. Pricing teams spend a lot of time thinking through which pricing metric is ideal, so I pulled together what the experts say, and synthesized them into one practical framework.

What the experts say about a good pricing metric

A lot of smart people in pricing have already described what makes a good pricing metric. They use different language, but they are largely pointing at the same underlying truth: a pricing metric has to work for the customer and for the company at the same time.

Here is what some great pricing leaders said makes an ideal pricing metric:

1. Nagle & Mueller’s classic “The Strategy and Tactics of Pricing” has 5 criteria:
   - Tracks with Difference in Value Across Segments
   - Tracks with Differences in Cost-to-Serve
   - Is Easy to Measure and Enforce
   - Facilitates Favorable Positioning Versus Competition
   - Aligns with How Buyers Experience Value in Use

2. Kyle Poyar gave a SaaS monetization list of criteria in 2023 recommending metrics should be Value-based, Flexible, Scalable, Predictable, Feasible.

3. Simon-Kucher focuses on value and fairness: “The most effective price metrics are the ones that are aligned with the value customers derive from the offering. This means that the price scales alongside customers’ success or their usage of the software. How predictable and measurable these metrics are will depend on the industry your company operates in. At a minimum, however, price metrics should be accepted and considered fair by customers.”

4. Ed Arnold says that “Good pricing strategies align the product’s price metric with either the customer’s usage or value metric.“

5. Amy Konary has 5 criteria with a heavy focus on the customer view:
   simplicity, predictability, transparency, control, and behavior.

6. Mark Stiving names these 5 criteria for a pricing metric: Fairness & trust, Comprehension, Predictability, Growth & expansion, and Profitability.

Each of these focus on something important. Nagle & Mueller and Kyle Poyar lean more toward value capture and feasibility. Simon-Kucher, Ed Arnold, and Amy Konary put more emphasis on customer value alignment, predictability, and acceptance. Mark Stiving comes closest to balancing both sides by making the commercial consequences explicit. What I wanted, though, was one practical framework that separates and brings together the needs of both the customer and the provider.

Bringing the expert frameworks into one

These experts focus on different dimensions of the same underlying challenge. The core pattern across all of these expert frameworks is that a good pricing metric has to do two jobs at once. It must work for the customer, and it must work for the provider.

My synthesis is to collapse them into six practical criteria:

• For the buyer: be intuitive, controllable and predictable

• For the seller: be attributable, scalable and feasible

Customer-facing criteria

1. Intuitive

   • Is easy to understand and meaningful to the buyer persona

   • Reflects how customers perceive the value your product creates

2. Controllable

   • Customers can influence what they pay through their behavior

   • Feels reasonable relative to the perceived value received

3. Predictable

   • Does not vary materially due to inputs outside the customer’s control

   • Customers can forecast future spend within a reasonable range

Company-facing criteria

1. Attributable

   • The metric captures willingness to pay across segments and use cases

   • It helps isolate and monetize the value your product creates

2. Scalable

   • The metric increases as your product creates more value over time

   • You can reasonably control and forecast the cost-to-serve

3. Feasible

   • The metric is measurable within the product

   • You can measure it without depending on customer-reported inputs

Evaluating common AI pricing metrics with this

Once you apply the framework, you see the focus of the common AI pricing metrics:

User metrics

User metrics were a great proxy for value for SaaS providers, and they are easy to explain, control and budget for buyers. That is why they were popular in SaaS, and why they remain popular for AI in B2B, e.g. Canva as well as ChatGPT (the UI version).

However, for AI products, seats don’t track value creation well, and they cause issues for scalability and attributability. When users connect their Canva account with ChatGPT Agent, their Canva account still tracks this as 1 user seat, even though a human and an Agent both use their product. Identity providers are starting to separate agent identities from human ones, which could solve this also for pricing.

Usage metrics

Usage in AI is the best indicator for scaling revenue with cost, protecting margins when variable costs are high. They tend to also be feasible since a product can track its own usage. They are very common in infrastructure already, e.g.:

• Per API call: Google Cloud API Gateway

• Per duration and/or request: AWS Lambda

• Per token: ChatGPT API

However, they are often not great for the customer. They tend to be overly technical (e.g. input + output tokens, API calls) and volatility causes issues for budgeting.

AI credits

To solve for volatile usage metrics, many AI companies allow purchasing AI credits. These are not a pricing metric in itself but rather a fungible resource unit that can be used for consumption later. So they’re a billing mechanism. How much one credit is worth is often not transparent or at least hard to understand, which causes issues to connect back to perceived value for the customer.

Outcome metrics

These tend to be more intuitive and controllable for customers, and also easier to predict because they are close to the customer value, e.g.:

• Per successful payment: Stripe

• Per recovered chargeback: Chargeflow

• Per customer resolution: Intercom Fin AI

The challenge for AI vendors is to define outcomes they can attribute and control the cost-to-serve for. This is where many vendors fail, as we discussed in this article.

Hybrid metrics (Platform fee + usage or outcome)

This is where many AI companies will land, but from a practical standpoint they are also NOT a pricig metric per se. They combine multiple metrics in one pricing architecture as discussed here. The still leaves the choice of which metrics to combine into one architecture unanswered.

So how do you choose your right pricing metric?

The goal is to choose the one (or combination of multiple) that balances what your customers need and what you need. Here is the process that has worked for me:

1) Create a long list of ideas of potential pricing metrics

Start listing all plausible options that align with how the customer receives value:

• Outcomes

• Outputs

• Usage

• Activity

• Users

• Employees

• Workflows

2) Score each metric

Take the ideal pricing metric criteria and rate each one on a simple 5-point scale: High, High-to-medium, Medium, Medium-to-Low, Low.

Don’t overthink this. The point is to force tradeoff discussions.

3) Aggregate the scores

Average the ratings to create an overall view of which metrics look strongest. For example, High for Intuitive + Low for Predictable = Medium overall.

This will never be perfect, but it helps structure your discussions.

4) Model the pricing using real prospects

Consider a hybrid pricing architecture of multiple metrics (e.g. seats + outcomes) if you believe that balances the customer and your company needs best.

Then model out the financials! This is the step novice pricers skip. You need to see what pricing would actually look like for real customers with each potential metric.

5) Pick the metric that works for your customers and you

Choose the metric where the scoring indicates strong customer fit as well as strong company fit, and where the resulting pricing works best for your target customer profile.

The AI pricing metric that works

In AI, pricing power starts with charging with the right unit of value. If you choose the right metric, packaging and price points fall into place more naturally. If you choose the wrong one, even the most elegant pricing architecture will struggle.

That is why pricing teams should spend more time determining the right metric. Customers need to be able to understand it, budget for it and control their spend. Providers need a metric that tracks with value, scales economically, and works for billing.

The best pricing metric works for both sides.

Subscribe now

It has four phases: You start with quantifying value for a buyer to justify their investment. Then you create value for your customer by focusing on their success. Track value and outcomes against clear baselines, and finally articulate that value back to customers and the market.

When these phases work together, monetization becomes an engine that fuels growth where every customer outcome strengthens retention, expansion, and the next deal.

Here are that four phases, and what to focus on during each:

1. Quantify Value

To close a deal, Sales needs to convince a prospect that they get 3x-5x in potential value to justify their investment. Value often isn’t guaranteed, while spend is committed and certain. This is how B2B sales work. To convince a prospect to buy requires them to see they have a problem, and that you can solve the problem for them.

Sales Engineering / Presales quantifies the value a customer gets. To close a deal, they translate product capabilities into economic impact: Revenue lift, risk reduction, cost savings, or productivity gains. The problem is this is often done ad hoc. Value calculators standardize this across customers so reps don’t have to start from scratch for every deal. They ensure the main drivers of value are considered when negotiating deals, reducing discounting, increasing win rates and making sales more efficient.

Sales needs to convince the prospect that the value applies to them. Benchmarks and independent value studies give reps credible reference points, but prospects can question their applicability: Different industry dynamics, varying geographic needs, differences in company size. There are many reasons why they would not get the same ROI. Customer success stories from comparable customers who achieved real ROI help here. A product without success stories won’t survive very long.

That way you move from selling features and capabilities to quantifying and selling outcomes comparable customers have actually achieved.

Common pitfalls to avoid:

• Forgetting about the status quo: The prospect always has the option to change nothing. Change = risk. Always address the status quo in your ROI model, and make the case for change!

• Overly ambitious ‘benchmark’: Just because McKinsey or another reputable source states on average customers achieve something, doesn’t mean your customer believes this is achievable for them.

What works:

• Business case based on a value calculator → standardize how value is quantified

• Benchmarks & independent value stories → give reps credible reference points

• Customer stories → prove that value already happened elsewhere

2. Create Value

Once a customer buys, you need to deliver on what you promised, and time-to-value becomes critical. This is where many companies struggle. A customer bought based on expected business outcomes. If the implementation and adoption is not focused on achieving business outcomes, the buyer easily ends up disappointed.

Customer Success is tasked with making sure value is realized. They build a Success roadmap with the customer, align stakeholders, and accelerate adoption to get the customer to outcomes. But they can only do that if they know what outcomes Sales has sold. In large organizations, the handoff from Sales to CS is often a failure point.

Over time things also change, commercial context gets lost. The customer might have a reorg, or just normal attrition. Your CSM might have moved to a new opportunity, Sales has long moved on to the next pitch. Chances are your signed Order Forms are in some repository, but the pitch decks? Not very likely. Then CS is left guessing what business problem the customer was trying to solve. Make sure you have ONE value repository for all docs around value, be it the value model from Sales Engineering, the pitch deck and business case from Sales, or the Success roadmap from CS.

Common pitfalls to avoid:

• Treating onboarding just as a project: Timely setup and successful training are table stakes, but time-to-value means more, i.e. measurable business outcomes

• Sales and CS have separate siloed repositories: Handover between teams are THE most common failure points I’ve seen over the years in many environments

What works:

• Clear handoff from Sales to CS → preserve the business case and value model

• Success roadmaps → tie product rollout to measurable business outcomes

• ONE value repository → give Sales and CS visibility on value promised over time

3. Track Value

Aligning on outcomes is not enough. To prove value was created, you need to measure improvement against a baseline. If the first time you think about measuring impact is after the implementation is completed; or worse: during a renewal discussion, you are too late. You lost the chance to show evidence.

Customer Success needs to quantify success numerically. Outcomes need to have clear KPIs with targets so the finish line isn’t constantly moving. Stakeholders need to be aligned, especially when the buyer in the organization is not the same as the one overseeing the implementation - think of a BI tool used by Finance, but implemented and rolled out by the CIO org. The CFO signed off on the business case with certain outcomes to justify the ROI. Then the CIO organization manages your rollout. Different stakeholders, different targets.

It’s tempting to skip the baseline measurement when stakeholders ask for progress. You’re accelerating the implementation, training, and adoption to get the customer to value faster. The baseline? Gets skipped. How many staff did the team have when we sold ‘efficiency gains’? How long did the process take before we ‘accelerated revenue’? Improvements cannot be measured without a baseline though! Without it, value realization stays vague. And vague does not renew well.

Common pitfalls to avoid:

• No aligned baseline: Not documenting baseline measurements of your success KPIs, or measuring but not agreeing with the customer often leads to debates later

• Tracking activity instead of outcomes: Implementation progress and adoption are table stakes, but are not the same as outcomes or Return-on-Investment (ROI)

What works:

• KPIs & targets per outcome → define outcomes, quantify each and set a target

• Baseline measurements captured up front → set a reference point of what the KPI was before you started, so you can measure improvement later

4. Articulate Value

Once you created and measured value, articulate it back to the customer. Consistently, on an ongoing basis. If you do not communicate the impact, you are relying on the customer to connect the dots. Value not perceived might as well not exist.

Customer Success is usually responsible to protect ARR, sometimes to drive expansion. According to Gainsight’s Customer Success Index 2025, 62% of CS organizations utilize GRR as their primary metric, so they’re isolating expansion and focus purely on avoiding churn. Communicating value repeatedly, whether in QBRs or executive reviews, is the critical role CS plays to proactively address churn risk.

And for smaller customers that don’t have a CSM, value dashboards need to fulfill the same role. They need to focus on business outcomes, KPIs and progress against those. No Executive renews to increase weekly active users or other adoption metrics.

Marketing also needs to document proof points from one customer as inputs for the next prospects. Success stories with quantified outcomes, before-and-after metrics, and Executive’s quotes all strengthen the funnel and help close the next deal. That is how the Monetization Flywheel comes to spin. Value delivered to one customer becomes the proof point for the next one.

Common pitfalls to avoid:

• Adoption reporting: adoption metrics are easy via telemetry, and standard across customers. Not tying them back to outcomes invites questions on why to renew

• Treating customer proof points as marketing assets: Marketing usually collects “success stories”, but Sales and Presales also need them further down the funnel

What works:

• QBRs and executive reviews → translate product impact into business language

• Value dashboards → make progress visible for smaller customers

• Capturing success stories → turn customer outcomes into future proof points

The Monetization Flywheel that works

The Monetization Flywheel only works when value is treated as a continuous commercial discipline, not a one-time sale or pricing exercise. Companies that continuously quantify, create, track, and articulate value systematically are better positioned to reduce discounting, improve renewal rates and increase expansion.

In that sense, monetization is not just about price setting and price getting, which most pricing teams focus on. Implementing the Monetization Flywheel requires a lot more cross-functional collaboration, but the effort pays off to compound growth.

1) Naming conventions stayed similar to SaaS

Many SaaS companies had 4 tiers (with few exceptions like Salesforce having more).

“Free” or “Starter” was the entry tier for many SaaS companies. Though ”Starter” is not common for AI companies, “Free” remains the default entry tier name for more than two thirds of AI native companies such as ChatGPT Free and Claude Free.

“Pro” remains a very common name for a paid tier. In our dataset, over 80% of AI-native companies have a tier named “Pro” (e.g. Cursor Pro, Perplexity Pro).

“Business” persists as the most common name for the tier that has some limited business features. 50% of companies in our data set had a tier that contained “Business”, such as ChatGPT Business, Grok Business and n8n Business.

“Enterprise” remains the most common top tier name, e.g., Lovable Enterprise, Replit Enterprise. Companies still require SSO, admin controls, and centralized billing. It’s actually not easy to find a company that does not have an “Enterprise” tier.

2) Packaging added power user tier(s): Ultra/Max

AI-native companies also have generally an additional middle tier. This tier is aimed at power users (developers, analysts, researchers) and priced in the range of $100-$200 per month. The name varies with “Ultra” and “Max” being the most common at the moment, e.g. Cursor Ultra ($200), Claude Max 5x and Max 20x ($100 and $200).

In B2B SaaS, tiers were mainly differentiated by features and company size. “Land and Expand” sales motions often started with one team, expanding enterprise-wide later. AI adoption often starts with a few individuals experimenting too, but power users have high costs. The additional power user tier(s) capture higher prices earlier in the adoption cycle.

To maximize price capture in B2C, some AI companies also introduced a new lower-priced tier to monetize the many light users not willing to upgrade to Pro. ChatGPT Go ($8, ad-supported) promises 10x the usage of their Free plan, and Gemini AI Plus ($7.99) has twice the credits of Free. Both were launched globally in January 2026.

3) Differentiators shifted: features → usage

While in SaaS, tiers were mainly differentiated by features and company size, AI-native companies like Claude differentiate tier much more around “More usage” and “Priority access at high traffic times”. The reason is simple: AI products incur real compute costs, and power users can cause a significant cost far beyond what they pay, even with the new Ultra/Max tiers.

So usage must be controlled, and companies have two levers for this:

1. Throttle usage to indirectly ensure value scaling. Claude has these limits:

   1. Session limits: A limit for each session, which resets every 5 hours.

   2. Weekly limits: Weekly limits across all models + some model specific ones.

2. Meter usage directly: scale pricing with usage within a tier (often via credits), e.g.:

   1. Replit Pro pricing has a drop down to select between $100 and $4,000 (40x)

   2. Lovable Pro pricing allows selecting between 100 to 10k credits (100x)

Your operator takeaway

Even though AI changed the economics, the tiers naming did not radically change. AI-native companies kept what buyers were already familiar with (“Free”, “Pro”,
”Business”, “Enterprise”).

They evolved the tier structure and names to reflect their economic reality by adding one or two “Ultra” / “Max” tier(s) for power users, differentiating tiers with usage throttling, and allowing to purchase metered usage (up to 100x) within a tier.

I touched these three levers before in my post on Discounting as a Growth Lever. This is the deeper benchmark cut focused on billing and contracting discounts.

Across the pricing pages I reviewed, billing and contracting discounts show up in three forms.

1) Multi-year vs annual: Churn reduction

Multi-year commitments are usually still paid annually. The customer typically does not prepay 3 years up front. The main benefit is churn reduction through fewer renewal moments.

A 3-year commitment means the customer questions the purchase 67% less often. Once every 3 years instead annually. A 2-year commitment still reduces the decision points by half.

One observation before the numbers. Very few B2B companies publish multi-year discounts on pricing pages. That makes the few that do publish them unusually valuable. If you work in pricing, these data points are gold.

In my dataset, the median for multi-year vs 1-year discount is 20%.

3 examples:

• Codeanywhere Premium: 13% ($23 per month on 2-year vs $20 on 1-year)

• TaxDome Pro: 10% ($900 per year on 3-year vs $1,000 on 1-year)

• SEO PowerSuite Professional: 32% ($239 per year on 3-year vs $349 on 1-year)

2) Annual vs monthly: Cash flow optimization

In most B2B products, “annual” typically means annual contracting plus annual prepayment. This is a working-capital lever as much as it is a commercial one.

Annual prepay improves cash conversion and lowers billing friction. On top of getting the cash up front, your Finance team has to issue and reconcile 92% less invoices.

In my dataset, the median discount for annual prepayment is 17%.

3 examples:

• Cursor Pro: 20% ($16 p.m. billed annually vs $20 monthly)

• Replit Core: 20% ($20 p.m. billed annually vs $25 monthly)

• GitHub Copilot Pro: 17% ($100 annually vs $10 monthly)

3) Monthly vs Paygo: Behavioral economics

Monthly subscriptions (vs annual commitments) are much less predictable and create a lot of uncertainty for AI companies like Lovable. Their Growth leader Elena Verna recently said this:

Monthly subscriptions (vs paygo) don’t change that, but they use the same behavioral economics pattern that makes gym memberships so successful. People overestimate future usage. Paygo forces you to feel the pain of EVERY single use. A subscription removes that friction, makes continued usage feel like a gain (since you’ve already prepaid), and charges even if you use it less.

In my dataset, the median discount here is 21%.

3 examples:
• Lovable: 17% ($25 for Pro including 100 credits vs $15 for 50 top-up credits)

• Luma AI: 20% ($20.99 for Plus including 10k credits vs $4 for 1,200 credits)

• ElevenLabs: 27% ($22 for Creator including 100k credits vs $0.30 for 1k credits)

Stacking discounts

Buyers (and you Account Executives) stack these billing and contracting discounts. Multi-year plus annual prepay plus commit instead of paygo.

Using the percentile table, the implied 3-year vs paygo discount lands at 48% at the median. Top quartile is even 58%.

Most companies don’t publicly offer all three levers, but many B2B companies already discount annual and multi-year commitments. If you layer AI credit pricing on top, make sure you understand unit economics well, and define stacking rules clearly.

To capture your fair share of the value you create, you need to:

1. Explain value in a way a buyer can quantify quickly.

2. Personalize it so the buyer see this is my situation.

3. Validate it with evidence the buyer trusts.

Build a simple chain: value calculation → relevance → proof. Here are the three key tools to do it, with practical guidelines and real-world examples:

• ROI calculators

• Customer success stories

• Independent value studies

1) ROI calculators

You need to articulate value with an ROI calculator. A good calculator teaches the value model, then shows the math with the buyer’s inputs, so they can see why buy.

A great example is ServiceNow’s Value Calculator: it

• has 4 inputs to tailor value to a prospective buyer

• Articulates value broken down into 3 buckets

Another one is Crowdstrike’s Cybersecurity Consolidation Value Calculator:

• Has 5 inputs to tailor value to the specific buyer

• Articulates value in 4 buckets

The best ROI calculators follow these practices:

• Focus on 3-4 business outcomes (e.g. cost saved, risk avoided, revenue gained)

• Design the formula (e.g. savings = incidents * time/incident * % reduction)

• Limit to 3-5 inputs (e.g. # incidents, time/incident, team size)

• Use defensible default assumptions (e.g. industry ranges) for inputs

• Pre-fill with typical values so the buyer can edit instead of starting blank

Common pitfalls to avoid

• Hidden assumptions / unrealistic savings → ROI inflated, not credible

• Too many outputs → Lack of focus

• Too many inputs → The prospect moves on

2) Success Stories with KPIs

Make the value feel specific with customer stories that show achieved KPIs. Case studies are not just proof of value, they make value relevant. The buyer is asking: “Is this my world?”

Your story library should:

• Lead with KPIs and time to value
  (e.g. “Reduced onboarding from 14 days to 5 days in 90 days.”)

• Let prospects filter to match their situation

A great example are Snowflake’s 342 stories that can be filtered by industry, product categories, department and region.

Each success story then shows the top 2 KPIs upfront. On the left, they show metadata conveniently: the industry, location and product used. Here is an example:

Many leaders do this well (e.g. Salesforce 231 success stories, Adobe’s 304 success stories) and allow filtering by product and industry.

Buyers are not going to scroll through dozens of stories, so filters and metadata matter! Think about how value varies, and consider these ones in this order:

• Industry

• Business size

• Region

• Product or use case

• Buyer department or persona

• Tech stack or integration context

Common pitfalls to avoid

• Soft outcomes only → Lack of KPIs signals ROI is not tracked

• KPIs not business-critical in industry → Hard to establish why it matters

• Stories with no KPIs → Buyers want to know if they get the ROI you articulated

3) Independent Value Studies

The final step is to build trust through independent value verification. Buyers discount vendor claims. Third-party validation reduces that discount rate. It also gives internal champions something they can cite without looking biased.

Here are three companies offering them:

Forrester’s Total Economic Impact (TEI) studies, e.g. ServiceNow, Zoom

IDC Business Value (BV) white papers, e.g. Microsoft Dynamics 365 for Manufacturers, AWS

Nucleus Return on Investment (ROI) studies, e.g. SFDC at Pearson, Oracle at SNFC

Zoom’s TEI study shows the benefits over time by type, calculates ROI and payback period. The most important part is it is independent and Forrester states how many customers they interviewed, and what type of customers these are (see example).

Independent value studies help you defend assumptions in your ROI calculator by establishing a credible value range for a category and solution type, verified by a trusted third party. It arms your sellers and champions with language their finance and procurement teams accept.

How to best use third-party studies:

• Position the study as a benchmark (not a promise), a note its limitations

• State what must be true for similar outcomes (baseline, adoption, scope)

• Use it alongside your ROI calculator + success stories to increase credibility

Common pitfalls to avoid

• Not acknowledging ranging outcomes → Perception it might be cherry picked

• Citing ROI with no context → Prospect may not believe it applies to them

The system that works

The three components reinforce each other:

• ROI calculator: “Here is the value model with your inputs.”

• Filtered KPI stories: “Here are customers like you who achieved it.”

• Independent verification: “Here is a third party validating the results are real.”

If you’re just starting

1. Build your calculator first and test with your next pitch (with buyer’s inputs)

2. Turn early wins into KPI stories once value is realized

3. Commission an independent study only after you have traction (they’re expensive)

Use this guide to design your pricing architecture to support predictable revenue and profitable growth in your reality:

• Is your customer’s demand seasonal?

• Do “power users” consume multiples of everyone else?

• Is usage hard to forecast for customers, and your marginal costs are high?

• Do you incur meaningful costs even for small accounts?

• Do your enterprise customers need 24x7 expert support and strict SLAs?

• Do your customers want budget predictability?

Recurring Subscription Components

1) Committed Volumes (Users, Usage or Outcomes)

A recurring contract with committed volumes in a chosen metric. The defining feature is committed recurring revenue, not the unit of measure. Recurring credit bundles can fit here if they are contracted and revenue-recognizable under ASC 606.

Examples:

• Lovable Pro is $25 for 100 credits per month

• Zendesk AI Agents are $1.50 per committed automated resolution

2) Platform Fee

A fixed recurring fee that captures baseline value and helps cover non-scaling costs. There are 3 common variants:

• 2-part tariff: platform fee does not include any usage

• 3-part tariff: platform fee includes base allowance, overages are billed

• All inclusive: unlimited use, protected by rate limits or a fair use policy

Examples:

• NextMatter charges 7,500 EUR per month, in addition to per user charges

• Basecamp Business is $299 per month all-inclusive

3) Support & Success

Paid support subscriptions layered on top of licenses. These are typically priced as a % of product spend and tied to access, SLAs, and expert resources.

Examples:

• Microsoft Unified Enterprise Support is ~7.5-10% of product spend, min $50k

• Salesforce Premier Success Plan is 30% of license fees

One-time Revenue Components

4) Up-front Services

One-time services for onboarding, integration, testing and training.

Examples:

• Factor A/E charges a one-time implementation fee plus subscriptions

• Link2Feed requires a one-time implementation fee on top of its subscription

5) Usage-based Consumption

Pay-as-you-go with no upfront volume commitment. These are common for variable or AI usage and value-added services.

Examples:

• OpenAI API calls are priced per 1M tokens and usage is billed monthly

• Amazon Bedrock Flows: $0.035 for 1k node transitions, metered & billed monthly

6) Migration Services

One-time services to extract, transform, or move data, or fees for data egress or assisted migration.

Examples:

• MongoDB Atlas charges $0.125 per GB exported plus transfer fees

• AWS Snowball Edge export jobs are priced per device fee plus shipping

Closing Thought

Let’s go back to our questions from the beginning. Here is how you can design you pricing architecture so revenue is predictable and your business grows profitably:

• Is your customer’s demand seasonal? → Use subscriptions with annual caps

• Do “power users” consume multiples of everyone else? Offer pooled allowances

• Is usage hard to forecast for customers, and your marginal costs are high? → Offer committed credits with mid-term top-ups (pre-priced incremental add-ons)

• Do you incur meaningful costs even for small accounts? → Add a platform fee to cover non-scaling cost

• Do your enterprise customers need 24x7 expert support and strict SLAs? → Consider an Enterprise Support priced as a % of ARR with a minimum

• Do your customers want budget predictability? → Price committed volumes significantly lower than pay-as-you-go rates

Which tactics companies choose to turn volatile AI usage into more stable, recurring revenue comes down to answering two questions:

1. Who bears the risk? Do you shift volatility risk to the customer, or absorb it?

2. How do you control it? With contractual terms or in-product levers?

Drawing this as a 2x2 matrix shows four combinations, and all 4 are gaining traction:

1) Committed capacity at a Premium
(vendor-friendly, commercial)

In SaaS, committed volume meant discounts. In AI, committed capacity can cost more because it comes with reserved performance.

OpenAI Scale Tier pricing offers higher reliability and better latency. Under standard pay-as-you-go API pricing, 1M input tokens for GPT-5.2 cost $1.75. With Scale Tier, customers purchase 25,000 “tokens per minute” capacity for $105 per day. In a best-case utilization scenario (100% utilization for all 1,440 minutes of a day), this is a 67% premium over PAYG API pricing.

Amazon Bedrock’s Provisioned Throughput similarly lets customers provision a higher level of throughput at a fixed cost. Longer commitments reduce price, but the core idea is the same: Pay for reserved capacity, not just consumption.

2) Auto-upgrades as default
(vendor-friendly, in product)

Auto-upgrades when credits are exhausted lead to an immediate tier jump. It prevents disruption, but locking the customer into a higher commitment by default can cause backlash.

Sleekflow’s AI Agent Flow Builder has auto-scaling enabled by default. It automatically upgrades the subscription to the next pricing tier when monthly flow enrollment credits are reached. It is framed as preventing disruption, and the upgraded tier remains moving forward.

HubSpot made auto-upgrade the default for HubSpot credits in December 2025. Once purchased credits are exhausted, accounts are automatically upgraded to a higher Credits pack for the rest of a commitment term. Enterprise customers can opt out of auto-upgrades to pay for usage overages instead.

3) AI credit pooling & controls
(buyer-friendly, in product)

Shared credit pools smooth variance across users. Heavy users draw from the same entitlement as light users. Admins get visibility, guardrails and governance.

Figma announced it will allow admins to purchase a shared pool of additional credits starting March 11, 2026. Approved users can draw from the pool when they exceed AI credits included in their seat, and admins can track team-wide credit usage in their dashboard.

Palo Alto Networks includes pooled flexibility and burst protection for its Next-Generation Firewall via pooling credits across deployments and clouds, and averaging usage over a month.

4) Burst protection against usage spikes
(buyer-friendly, commercial)

Burst protection waives charges for short spikes. It typically works by excluding peak days, ignoring extreme daily overages, or capping billable overages, usually with limits on frequency or duration.

Honeycomb’s burst protection does not count events on days where usage exceeds daily event targets by more than 2x. This can be triggered up to 3 times in a calendar month.

Login VSI offers 1 month of burst protection per year. Customers can scale up to 2x contracted usage at no extra cost.

Jetpack Stats excludes the 2 highest traffic days each month. If the customer exceeds the limit in three consecutive months, they require an upgrade to a higher tier.

5) Backdating upgrades
(two flavors, commercial)

Backdating upgrades retroactively applies a higher tier so the customer pays for a higher commitment instead of one-time overage fees.

Customer-requested backdating (forgiveness): Acquia Cloud Platform offers a 30% capacity buffer before overages are invoiced. Customers can backdate the effective date of an upgrade to avoid one-time overage fees above the buffer. They also offer full overage protection for the first 3 months of production launch.

Audit-driven retro tiering (enforcement): Redress Compliance reports that SAP SaaS audits check Fieldglass usage and sometimes push customers into a higher subscription tier retroactively.

Closing Thought

Volatility is a real problem for B2B AI buyers. If you want to reduce tension and churn risk, stop treating surprise usage bills like a punitive tax.

Use these tactics based on your comfort to take on risk:

• Committed capacity when reliability is the value, price it like an SLA.

• Auto upgrades to prevent disruption, but beware of buyer pushback if it feels like a trap.

• Pool credits to smoothen seat-level variance, only when paired with admin controls.

• Burst protection to separate true growth from temporary spikes.

• Backdating to reframe overages into a win-win commitment.

Thanks for reading, and see you in the next one.

Funding builds options. It keeps architectures flexible and narratives open.
M&A hardens decisions, collapses layers, and locks in the control layers where leverage compounds, switching costs rise, and architectural choices become durable.

Those control layers are infrastructure, identity, data, and experience.

1) Vertically integrated AI infrastructure

(Simulation → silicon → networking → compute)

Deals:

• Synopsys acquired Ansys (AI simulation and chip design) for $35B in July 2025.

• Hewlett Packard Enterprise acquired Juniper Networks (AI-native data center networking) for $14B in July 2025.

• CoreWeave acquired Core Scientific (High-density AI compute) for $7B in July 2025.

What is converging:
AI workloads are forcing tight coupling across layers that were historically modular. Simulation, chip design, networking, power, and compute are collapsing into a unified layer. Scale is becoming the competitive moat.

Why it matters:
This signals the end of best-of-breed infrastructure for frontier AI. The winners own the entire infrastructure stack, from silicon to compute.

2) Agent identity & trust

(Human identities → agent identities)

Deals:

• Alphabet acquired Wiz (Cloud and AI-native security posture management) for $32B in July 2025.

• Palo Alto Networks acquired CyberArk (Identity for machine and agent access) for $25B in July 2025.

What is converging:
Identity security, cloud posture, and runtime protection are merging into agent-aware trust layers. The perimeter is no longer users or workloads. It is autonomous AI agents acting independently.

Why it matters:
This is the creation of control infrastructure for non-human identities as agents inside enterprises require new controls.

3) End-to-end data

(Raw data → labeled data → real-time inference)

Deals:

• Meta Platforms invested $14B in Scale AI (Data labeling for AI applications) in April 2025.

• IBM acquired Confluent (Real-time data streaming for AI systems) for $10B in May 2025.

• Salesforce acquired Informatica (Enterprise AI data governance) for $8B in June 2025.

What is converging:
Data ingestion, labeling, governance, and activation are collapsing into AI-ready data layer. The distinction between operational data and big data is disappearing.

Why it matters:
Owning the data lifecycle becomes more valuable than owning models. Models churn. Data moats compound.

4) AI experience and interactions

(Dashboards → continuous, ambient AI)

Deals:

• Qualtrics acquired Press Ganey Forsta (AI experience measurement) for $8B in June 2025.

• OpenAI acquired io Products (Always-on AI device) for $6B in September 2025.

What is converging:
Experience measurement, feedback loops, and interaction surfaces are becoming continuous and embedded. Hardware, software, and inference are merging into always-on experience systems.

Why it matters:
The experience is shifting from UX to continuous presence. AI is no longer something you open as an app. It is always on and with you.

5) The July 2025 deal cluster & the meta-convergence

More than half of the top 10 AI deals in 2025 closed in July alone. I had posted about the spree at that time here. The July cluster was structural, because by mid-year three things had changed:

Cost for compute and inference were better understood. Stanford’s 2025 AI Index Report from April 2025 found hardware costs declined ~30% annually, and energy efficiency improved by ~40% each year.

Confidence in enterprise demand for AI showed and Nvidia’s stock price increased 70+% from low of $94 in April 2025 into new all time highs of $160+ for the first time.

Buyers had H1 performance visibility and board-level approval to commit capital at scale. AI wasn’t experimental any longer.

That made future cost for real use cases predictably cheaper, increased confidence that growing demand is real, and allowed large transactions to act swiftly.

Closing Thought

The 2025 M&A wave showed where power actually compounds. Point solutions get absorbed. Control layers become the long-term profit centers.

If you are building or investing in AI, the critical question is which layer you own, and how durable that position is as the stack continues to consolidate.

This article will equip you with a practical playbook how to price PS right. You’ll also get benchmarks to explain it from ten public SaaS companies and the SPI PS Maturity Benchmark.

1. SaaS companies usually make 2-9% of revenue from PS

Across our dataset of 10 of the largest SaaS companies cluster around 2-9% of revenue from PS. This signals that their product can scale without heavy customization.

This band is the sweet spot where:

• Implementation stays repeatable

• Product (not people) drive scale

• Customization is not the default

This is the reason SaaS works - the product (not people) drive ARR.

2. SaaS companies use PS along with S&M as a growth engine

A striking correlation shows up when you compare PS margin with growth rates.

The fastest-growing companies in the dataset invest 40%+ of revenue into Sales & Marketing, and price PS at or below cost:

• Snowflake: 30% ARR growth YoY, -36% PS margin

• ServiceNow: 23% ARR growth YoY, -2% PS margin

• HubSpot: 21% ARR growth YoY, 3% PS margin

These companies use PS along with Sales & Marketing as a growth accelerator.

Affordable priced PS:

• Reduces perceived switching cost

• Accelerates time-to-value for customers

• Unlocks upsell opportunities earlier

Most major SaaS players from workflow platforms to CRMs treat PS like a growth accelerator rather than a profit lever.

3. Pricing PS too low undermines scalability

There’s a tipping point where underpricing PS starts becoming damaging.

The Rule of 40 is a metric to ensure you’re on a path to profitable scale. We’ve calculated it here as:

The companies that price PS far below cost also have the weakest “Growth + Operating Margin” results. DocuSign, Okta, Snowflake, UiPack all fall way short of the Rule of 40, and significantly underprice PS (-19% to -73% margin).

Excessively negative PS margins are a red flag because they highlight problems:

• Low utilization / ballooning bench time

• Skilled niche labor at high cost

• Inefficient PS delivery model

4. The PS sweet spot is between -10% and 10%

Profitably growing SaaS companies meet or are just below the SaaS Rule of 40, e.g.

• Adobe: 11% PS margin → 12% ARR growth + 31% operating margin = 43%

• ServiceNow: -2% PS margin → 23% ARR growth + 12% operating margin = 35%

• Salesforce: -10% PS margin → 10% ARR growth + 19% operating margin = 29%

A side note: The behemoth Microsoft stands out by achieving the rule of 40 even with declining ARR. Its 66% PS margin and heavy PS footprint (77% of revenue) helped it achieve its 46% operating margin.

5. The charge-rate playbook: how to engineer PS economics right

Even if your PS strategy is margin-neutral, you still need a rigorous charge-rate model, or you can drift into unintentional losses.

Every best-in-class PS organization uses a version of the following framework:

Step 1: Set the target margin (strategic, not arbitrary)

Think of margin as a design choice:

• -10% to 10% for standard PS to fuel adoption and SaaS growth

• 30+% for optional value-added services. Consider pricing these on value, not hours

How to optimize: Don’t use the same target margin all your PS work, vary it based on the service. Understand the value of premium services that create value beyond your SaaS or AI offering (e.g. insights from data to optimize your customer’s business), and monetize these at a higher margin.

Step 2: Understand your cost to deliver

A PS org requires different skills and resources to deliver on the mix of work effectively. Understand your cost to deliver based on these three factors:

• Skill mix & staff pyramid

• Onshore / offshore mix

• Fully-loaded salaries (base, bonus, benefits, taxes, tooling, overhead)

How to optimize: Understand what skill mix you require for the work you’re doing. Then be strategic about what to deliver in each geography. Some customers require reviews and collaboration on site, others are okay with remote work across timezones.

Step 3: Model utilization with real-world friction

How much time your team spends on billable work varies by your size, geography and PS maturity. The 2025 SPI PS Maturity Benchmark found that PS orgs within SaaS / Software companies usually achieve ~60-70% utilization. The rest goes towards:

• Onboarding & training (a new hire in productive only after ~2-3 months)

• Non-billable sales support & administrative work

• Internal coordination & downtime between projects

How to optimize: Accelerate time-to-productivity for new hires through better onboarding, improve PS org’s visibility into upcoming demand by integrating sales pipeline visibility, and deploy PS automation and AI tools to effectively forecast and allocate resources proactively.

Step 4: Putting it together: the charge-rate formula

For example: if you aim for a 10% PS margin, your cost is 300k, and your utilization is 70%, you should charge 1.1 x $300k / 70% = $471k.

This ensures your PS pricing supports adoption and sustainable growth.

Final Reflection

PS is a growth accelerator, not a profit engine. Large SaaS players found the sweet spot for PS margins to be between -10% and 10% to power sustainable growth.

To deliver value at reasonable prices, optimize your staffing and delivery model, and increase utilization through faster onboarding and better tooling. That way your PS work can fuel profitable growth by accelerating product adoption and unlocking expansion opportunities faster.

This post gives you a structured way to evaluate discounts so they protect value instead of eroding it. You’ll get:

1. A clean taxonomy of discount types and when to use them + real examples

2. A quantitative test for whether a discount creates or destroys profit

3. A practical playbook to align discounting with your growth strategy

1) The Three Types of Discounts: Billing & Contracting, Lifecycle and Segment

Billing & Contracting Discounts

Annual prepayment discounts:

Exchange flexibility for predictability. The customer commits for a year and prepays. You get better cash flow, lower churn risk, and higher forecast accuracy.

Examples: Lovable Business costs $42 per user per month with annual billing or $50 as monthly, Cursor Teams is $32 with annual billing, or $40 monthly. A 16-20% annual discount in exchange for annual up-front billing is common in SaaS as well as in AI.

Multi-year commitment discounts:

Trade flexibility for longer term revenue durability (RPO) and planning certainty, while payments remain monthly or annually. In AI and cloud, commitment-based discounting is the shift from on-demand usage to committed-use savings plans.

Examples: Microsoft Azure and AWS Compute Instant “Savings Plans” offer 60+% off when for committing to a steady amount of compute spend per hour over 3 years.

Product Lifecycle Discounts

Early adopter discounts:

Reward customers who adopt new capabilities before they are fully mature. That might be an explicit “preview” price or a lower introductory price that increases once the product stabilizes.

Example: Jamesserra.com reported that Microsoft introduced Query Acceleration for Azure Data Lake Storage at a 50% discount relative to the intended standard rate.

Multi product cross sell discounts:

Expand customers by adopting more of your platform through a temporary incentive. The economics rely on increased ARPU and deeper lock in, at the expense of margin per product for a time.

Example: HubSpot Starter Customer Platform bundle is “The Starter edition of each product” and instead of discounting individual Hubs permanently limited to “up to 40% off HubSpot’s Starter Customer Platform for the first year”.

Legacy migration discounts or credits

Ease the transition from legacy products to new ones to simplify the portfolio. This accelerates the adoption of modern architectures and avoids relying solely on “end of life” phase-out.

Example: Atlassian cloud transformation incentives help migrate customers from Data Center or self-managed environments to Atlassian Cloud. These include credits for unused maintenance contracts and dual licensing periods.

Segment Discounts

Segment discounts are deliberate pricing policies for specific customer groups, not one off negotiation levers. They can be extremely aggressive without eroding your enterprise reference price if you keep the segmentation crisp.

Education and student discounts

Fuel adoption in the education segment, so students are familiar with the toolset when they start the workforce - and drive or even demand adoption of the tools they know in the workplace.

Example: Adobe Creative Cloud Pro for students and teachers offers Photoshop, Illustrator and Premiere to students and teachers at a 70% discount for 1 year on Cyber Monday.

Startup discounts

Support early-state companies with limited budgets but high growth potential, and count on your product being sticky. The discount is really a customer acquisition and ecosystem play for future, higher value usage.

Example: Notion for Startups offers eligible startups up to six months free on the Business plan with Notion AI included.

Non profit discounts

Subsidize socially desirable customers, who tend to be very price-sensitive due to limited funds.

Examples: Google offers Workspace for Nonprofits for $0, and discounts Business and Standard offerings by 70+% vs standard pricing. Microsoft 365 Nonprofits offers 75% off for E3 and 60% for E5 licenses compared to commercial licenses.

Military discounts

Salute service members and ease their transition into the commercial sector, but with strict limits such as one license only to avoid abuse and resale.

Example: Adobe Veterans Advantage provides a 15% discount on Creative Cloud Pro for eligible military members and veterans through programs like Veterans Advantage

Frontline worker discounts

Many B2B tools are built for desktop users, and require a monitor to unlock their full value. Workers on shop floors, in hospital rooms or emergency response vehicles may get less value and demand a cheaper option.

Example: Microsoft 365 Frontline Workers F3 license is $8 per user per month (76% less than an E3), marketed as purpose-built (some apps are not included) and limited to “employees whose primary responsibilities involve direct interaction with customers, patients, or the public, or who support essential services.” This effectively fences the lower priced package.

2) Effective Discounting Requires Math and Clarity

Breakeven Sales Depend on the Contribution Margin

Every discount cuts your unit margin. To increase the bottom line, you need enough incremental volume to offset it. The key parameter is contribution margin: revenue minus variable costs as a percentage of revenue.

• If your contribution margin is 80% and you discount 10%, you need a 14% increase in volume to get to the same profit, the “breakeven sales”

• If your contribution margin is only 20%, cutting prices by 10% is very different. Variable costs take up most of the revenue already. A seemingly small discount wipes out most profit, and you’d need to double volumes to break even.

This chart makes the “breakeven sales” tangible:

• Look up your contribution margin on the x axis.

• Find the price discount you’re considering on the y axis.

• Find the volume uplift required to keep profit constant (it is color coded).

This gives you a quick visual check:

• At low margins, even modest discounts require aggressive growth to break even.

• At high margins, small, tightly controlled discounts can be economically acceptable if they drive retention, expansion, or adoption of strategic AI features.

Subscribe now

Clarity Preserves the Reference Price Point

Customers internalize a “reference price” that they view as normal. If you discount regularly and without clear guardrails, the reference price drifts down to the discounted level. That makes full price feel inflated and invites pushback.

Clarity is your countermeasure. Treat discounts as conditional, not a default:

• Conditioned on contract structure: e.g. annual pre-pay, multi year commitment

• Conditioned on product lifecycle: e.g. temporary early adopter discount

• Conditioned on specific segments: startups, education, nonprofits, etc.

When you draw that line clearly, you protect the integrity of your list price. Enterprise customers who do not meet the conditions pay the price without feeling that they are “missing” a perpetual sale.

3) The Practical Discounting Playbook

A disciplined discounting playbook prevents revenue teams from improvising on every deal and keeps discounts aligned with strategy and margin.

Step 1. Pick which discounts you offer, and which you don’t

Start by mapping discount types to business goals:

• Billing and contracting discounts accelerate cash, reduce churn, and improve predictability.

• Product lifecycle discounts drive adoption of new AI features or cloud platforms, support migration off legacy platforms.

• Segment discounts deliberately grow segments that matter for brand, volume, or long term opportunity.

Decide which of these you want to use and which you explicitly do not. That clarity keeps ad hoc discount ideas from creeping in.

Step 2. Consider how discounts stack

Customers often qualify for multiple discounts at once. Like a startup nonprofit adopting a new AI add-on on a three year contract. Without guardrails, they could accumulate annual, multi year, startup, nonprofit, and early adopter discounts into something economically absurd.

You need rules:

• Which discounts are mutually exclusive

• Whether some categories cap the total allowed discount

Your Sales team needs to know these guardrails to grow your business effectively.

Step 3. Model scenarios to ensure profitable growth

Finally, link discounting policy back to contribution margin and breakeven volume:

• Take your contribution margin by product or bundle

• Overlay planned discounts and compute the required volume lift to break even

• Stress test AI workloads where cost can fluctuate (e.g. tokens, API usage)

This is where usage-based AI models require more rigor than traditional seat-based SaaS. Your discount may make sense at current infrastructure prices and model efficiency, but break down if customers use your product in more intensive ways.

4) Final Thought

Discounting is not inherently good or bad. It’s a tool. Used well, it accelerates growth, opens doors to high growth segments, and secures profitability. Used poorly, it erodes margins, trains buyers to wait for concessions, and muddies your positioning.

The discipline is simple to state and hard to execute.

Decide why you discount.
Decide where you discount.
Decide how far you are willing to go at a given contribution margin.

If you do that work, discounting becomes part of your monetization strategy instead of a last minute lever in a desperate quarter end negotiation.

Let’s unpack why this happens and how leading SaaS teams are solving it.

The Three Core Challenges of Outcome-Based Pricing

Andreessen Horowitz identified the three biggest barriers holding companies back:

1. The lack of clear, measurable outcomes

2. Unpredictable or unscalable cost

3. Difficulty attributing value to a specific tool or model

Each of these hides deeper operational challenges.

Challenge 1: Lack of Clear, Measurable Outcomes

This challenge stems from two root causes:

Ambiguous definitions

Different customer functions often expect different outcomes. Take marketing AI tools like Iterable for example. Marketing may focus on curated moments, while Sales wants conversion uplift. Without a shared definition of what matters, outcomes become subjective.

Loosely defined outcomes lead to debates about what is counted as “success”, even when teams agree on the intent. Unclear edge cases make billing ambiguous, create room for disputes, and those damage customer relationships.

Inaccurate or inaccessible measurement

When companies can’t provide real-time visibility in product and customers only see consumption data after manual reviews and adjustments, trust erodes quickly. Customers begin questioning the accuracy of telemetry, and invoice disputes follow.

Connecting billing measurements with in-product analytics can be complex, especially when legacy billing systems limit what can be surfaced or shared. Newer monetization infrastructure platforms like Metronome can solve this, but adopting them often requires replacing those legacy systems that many SaaS leaders rely on.

Challenge 2: Unpredictable or Unscalable Cost

When your costs fluctuate more than your revenue, outcome-based pricing becomes a financial risk.

Billing in arrears

If customers are charged only after outcomes occur, surprise bills are inevitable. The Cursor backlash in July 2024 after their price model change showed how quickly this can erode trust. AI-native startups are especially vulnerable, as we explored in “Untangling Billing and Pricing”.

Uncontrollable cost drivers

If you cannot control how many attempts it takes to achieve a successful outcome, your unit economics collapse. That’s why companies like OpenAI price APIs by tokens, not results, and why Anthropic’s Claude enforces strict usage caps.

Challenge 3: Difficulty Attributing Value

The toughest problem is attribution: which product, model, or team actually created the outcome?

Unattributable outcomes

If a supply chain platform depends on a Transportation Manager to act to avoid a stockout, who really earned that success? Charging for outcomes outside your control invites disputes.

Human-in-the-loop

Autonomous AI Agents promised end-to-end outcomes, but vendors recommend human oversight (see Scale AI). Until AI systems can act fully autonomously, and vendors are willing to assume accountability, charging for the ultimate business outcome won’t be possible.

Outcome-based pricing can quickly shift from a growth engine to a liability if these traps aren’t managed. Let’s look at how to design it right.

The Playbook: Get Outcome-Based Pricing Right

Step 1: Identify the Direct Outcome Your Product Creates

Price the outcome your product actually delivers, not the business value a customer can ultimately create with your product. Attribution becomes questionable when your product doesn’t make decisions or someone else needs to act to achieve the outcome. Attribution clarity builds pricing credibility.

Step 2: Pick the Right Pricing Architecture

Not all outcomes are binary. Match your model to your outcome type:

• Fixed price per outcome → for binary events (e.g. “identify verified”)

• % of achieved outcome → for performance-based gains (e.g. revenue uplift)

• Refunds or guarantees → to share risk (e.g. Fin AI’s $1M dollar guarantee)

• Hybrid models → combine fixed and variable components

(For more, see “100+ Companies Charge for Outcomes Across 15 Industries. Why Aren’t You?”)

Step 3: Define Outcomes Clearly and Document Exceptions

Ask “What if...?” early. How are outcomes counted? What if the outcomes change after being achieved? Document all these in your pricing terms or FAQ.

Fin AI, for instance, specifies exactly what a “resolved conversation” means, including edge cases like a chat ending right after a greeting, or a customer revisiting a resolved case weeks later.

Define outcomes precisely. Ambiguity is your enemy.

Step 4: Consider a Platform Fee or Contract Minimum

Outcome-only pricing can make small accounts unprofitable. Consider adding a platform access fee to cover fixed costs or a minimum contract value to reduce CAC pressure.

This ensures you don’t lose money on customers with volatile results.

Step 5: Layer on Flexibility for an Upcharge

Encourage customers to forecast and commit to a baseline number of outcomes to get subscription-like predictability (See “Untangling Billing and Pricing” for more.)

Offer flexible commitment options, but price that optionality. It’s not uncommon to charge 33%-50% more for pay-as-you-go billing in arrears. For example,

• DataDog charges 50% more for on-demand events ($3) vs billed annually ($2)

• ZenDesk’s AI is 33% more expensive with pay-as-you-go ($2) vs committed ($1.50)

Step 6: Offer Customers a Choice

If your customers aren’t ready for full outcome-based pricing, offer an alternative as Decagon does, letting customers choose between outcome- and usage-based models.

Customers value choice. Offering the choice of outcome-based pricing signals confidence in our ability to achieve the outcome, and your willingness to share the risk with your customer.

Decision Tree: When Not to Price by Outcomes

Outcome-based pricing only works when you can control cost and predict how it scales. If you can’t, it quickly becomes a liability.

Most cost spirals come from variability: fluctuating success rates, unpredictable AI token use, long-running calls in voice AI, etc etc.
When that happens, price on what you can control:

• Cost scale with users → use per-user pricing with usage limits

• Cost varies by task → use AI credits to tie price to compute (e.g. Adobe charges more credits for high res video generation)

• Pure infrastructure → align price directly to usage (e.g. OpenAI API). But…

Beware The Downward Spiral of AI Cost Decline! Stanford’s 2025 AI Index Report found hardware costs declined ~30% annually, and energy efficiency improved by ~40% each year. If your pricing is tied directly to these costs, your margins will erode as customers expect your prices follow suit. That’s why pricing must link to customer value, not raw compute cost.

Closing Thought

Outcome-based pricing works when you own the outcome, measure it precisely and control the cost to achieve it. When you can’t, a hybrid structure with usage-limits, or usage-based pricing keeps your business stable while you mature your telemetry and attribution.

Companies that win with outcome-based pricing design for measurability up front.

Outcome-based pricing isn’t theoretical anymore.

Buyer Preference: Andreessen Horowitz found that in summer 2025, usage-based pricing (e.g. by token, API call) has overtaken seat-based pricing already as the preferred way to purchase AI (39% vs 21%). That’s a big change from SaaS pricing that was predominantly per user. What’s even more surprising is that 15% of CIOs actually preferred outcome-based pricing (e.g. per lead, resolution, task completion), and another 23% a hybrid (a mix of seat/usage/outcome pricing).

Adoption: The reality check from Ibbakka found in September 2025 that only 3% of companies charge directly for outcomes so far. Andreessen Horowitz identified the 3 biggest challenges with outcome-based pricing:

• Lack of clear measurable outcomes

• Unpredictable or unscalable cost

• Difficulty attributing value to a specific tool or model

Subscribe now

Despite these challenges, many companies across a broad range of industries have successfully adopted outcome-based models. Across 100+ companies from our 2025 database that use outcome-based pricing:

• ~48% charge a fixed price per outcome (e.g., identity verified, action completed)

• ~25% charge a % of the achieved outcome (e.g., revenue uplift, cost savings)

• ~17% offer refunds or outcome guarantees, tying compensation to risk

• ~10% leverage a hybrid of fixed and variable outcomes to balance components

Pricing consistently aligns with the nature of the outcome in each industry. Across industries, pricing models track the measurability of the outcome: binary results favor fixed fees, variable results favor revenue shares, and high-risk outcomes demand guarantees.

Binary Outcomes

In some industries, outcomes are binary - it was either achieved or not. The outcome doesn’t range qualitatively, there is no half way done. So a flat fee per outcome is optimal.

A flat price per outcome has become a prominent pricing model in quite a few industry use cases:

In the Automation & Support space

• Customer Service charges for tickets resolved automatically (Zendesk, Decagon)

• Workflow Automation charges for workflows or steps completed (n8n, Relay)

• Cybersecurity charges for valid vulnerability reported (Intigriti, HackerOne) or security tests completed (Synack)

• Image Editing charges for images edited or retouched (Imagen, Luxedit)

In the Sales & Marketing space

• Performance Advertising charges cost per click or cost per action (Google Ads, Outbrain)

• SDR AI charges for meetings held or tasks completed (ChaseLabs, Funnl)

• Software Marketplace Lead Gen charges for leads or qualified demo requests (SoftwareAdvice, GetApp)

In the Industrial and Utilities space

• Aerospace & Defense charges for engine flight hour (RollsRoyce, GE) or cargo delivered successfully (SpaceX)

• Carbon Credits charges per metric ton CO₂ removed (Climeworks, Indigo Ag)

• Compressed Air charges per compressed air consumed (AtlasCopco, Kaeser)

• Fleet Management charges per mile driven (Michelin Fleet)

• Managed Print charges per page printed (HP, Xerox)

In the Fintech and Legal spaces

• Know Your Customer (KYC) charges per verified identity (Idenfy)

• Legal Doc Summaries charge per deliverable summary (CaseMark)

• Legal Demand Drafts charge per demand package or medical chronology (EvenUp)

​​Why it works: Because the outcome is binary (achieved or not), a flat fee creates simplicity and predictability for both sides, turning performance into a measurable unit of delivery. This model offers predictable revenue recognition and low billing friction, which is ideal when the cost to deliver is stable and outcomes are binary.

Range of Outcomes

In other industries, it is not just the question of whether the outcome was achieved, but outcomes range a lot. A personal injury lawyer did achieve an outcome if he gets a $100k settlement, but another might have gotten a $10M outcome. In these cases, a % of the outcome is the best way to align pricing with value. This model pushes risk to the vendor and aligns incentives to maximize the potential outcome. It works only if attribution is airtight.

The use cases where a percentage of the achieved outcome is a prominent pricing model include:

In the Energy space

• Energy companies charge for % of energy savings achieved (Siemens, Budderfly)

• Lighting-as-a-Service providers charge for light generated (Signify, Stouch)

In the Fintech space

• Payment companies charge a % of successful transactions (Stripe, Flycode)

• IT & Cloud Cost Optimization providers charge a % of verified or achieved savings (Zesty, ProsperOps)

In the Legal space

• Personal Injury lawyers charge for a % of the achieved award (Morgan & Morgan, Brown&Crouppen)

• Airline Flight Claims companies charges for a % of the successful compensation (AirAdvisor, AirHelp)

Why it works: When results vary in magnitude, tying price to a percentage of the achieved outcome directly links value capture to value creation and scales incentives proportionally. This model scales naturally with customer value, aligning incentives and sharing upside when outcomes vary in magnitude.

Don’t define outcomes you can’t attribute.

Quality of Outcomes

In some industries or use cases, the outcome is expected, but a negative outcome can have dire consequences. For example, a water facility providing clean water 99.99% of the hours in a year is not sufficient given that would imply unpotable water for one hour every year. One hour of undrinkable water equals a year of lost trust.

In these cases, having a disproportionate refund (i.e., a penalty) might be required to get customers to get comfortable with outcome-based pricing.

Some examples of where a refund or outcome guarantee is prominent are:

• Education guarantees a job (Springboard) or a minimum income after graduating (BloomTech)

• Water-as-a-Service refunds for low water quality (Veolia, Xylem)

• Pharma provides refunds or rebates for negative clinical outcomes (Amgen, Merck)

Why it works: Where failure has outsized consequences, refunds or penalties build buyer trust and de-risk adoption, signaling confidence in the product’s ability to deliver. This model builds trust in high-stakes scenarios, de-risking adoption and signaling confidence through accountability. It only works when you can commit with high confidence.

Hybrid Outcomes

Some industries also blend fixed and variable elements to balance predictability with upside. Examples of these include:

The same company blends fixed and variable outcomes:

• Affiliate Marketing / Partner Programs pay oftentimes both, a % commission and a fixed fee for specific actions (Amazon Associates, Shopify Partner)

• FinTech chargebacks can be a flat fee for prevented chargebacks and a % of the amount for recovered chargebacks (Chargeflow)

Companies differentiate from competitors by using a different outcome-metric:

• Insurance companies vary from charging per miles driven (Metromile) to a % based on driving behavior (Root Insurance)

• Recruiting firms charge a fixed fee per placed candidate (Marloes) or a % of first year salary (Nigel Frank)

• Healthcare varies from charging a flat fee for signups from at-risk members (Omada Health) to % of cost savings achieved (Hinge Health) to % incremental revenue from (SmarterDx)

Why it works: Blending fixed and variable components lets companies balance predictability with performance upside or position themselves as better aligned with the outcomes their customers care about. The fixed fee secures recurring revenue, while the variable layer links earnings to measurable outcomes. This turns pricing into a competitive differentiator rather than just a contract term.

Final Thought

Outcome-based pricing is already working for many companies. The ones mastering it aren’t chasing the hype, they’ve made outcomes their unit of value. They’re charging for outcomes that matter to their customer base.

Winners operationalize value measurement. They manage risk (AI cost), customer alignment (outcome attribution), and cash flow (billing) as integrated parts of monetization design - we explored the latter in Untangling billing and pricing. They build systems that measure, attribute, and monetize value delivery with precision.

The timing of collection can either fuel or constrain growth. In an era of AI unpredictability and tightening cash cycles, understanding that difference can mean the gap between scaling sustainably and running out of runway.

Untangling the two reveals powerful levers for improving both profitability and liquidity.

1. The Evolution of Monetization Models

Every wave of technology has redefined how we charge for value:

• Perpetual software licenses: When PCs became mainstream with Windows 95, you had to buy a license up front. That license allowed you to use the software in perpetuity on your computer.

• SaaS from 1999 onwards changed the game by moving from a one-time payment to charging an ongoing subscription. Salesforce and others turned ownership into access, exchanging upfront capital for recurring predictability.

• Cloud infrastructure in 2006 went further and charged actual usage like AWS compute by second, not by seat, and billed after the customer used it.

• Platforms from 2010 onwards began charging for urgency and context. An Uber ride could cost more or less depending on how long you would be willing to wait. People willing to pay extra to fast-track, would pay more.

• AI platforms now experiment with outcome-based billing, credits tied to tokens, task completions, edited images … and a number of other price metrics.

If you zoom out, every major tech model can be plotted along two axes: how customers buy and what companies price.

2. How You Buy vs. What You Price

Think of it as two different systems running in parallel:

Billing defines when cash moves, and impacts cash flow, not the P&L.

Pricing defines what you charge for.

Let’s look at each of them separately, and the levers you can pull to improve cash flow AND value capture.

3. Billing as a Cash Flow Lever

Billing is the engine behind your operating cash flow. It determines how quickly you turn booked revenue into usable capital.

If you compare the main contracting patterns:

• Perpetual: Cash arrives before delivery. You recognize revenue over time but already hold the cash.

• Recurring: You collect cash while delivering service. The balance between billing cadence and contract term determines how smooth your cash cycle is.

• Usage-based (sometimes also called Consumption-based): You deliver first and bill later. Cash follows consumption, which can lag by months.

These aren’t accounting details. They define your ability to invest and grow. And there are a number of levers you can pull to optimize capital efficiency:

• Contract Period: Consider extending contracts from one to three years, and you lock in predictable RPO. You may even reduce churn with fewer renewals. Some of your customers may even prefer the predictability.

• Billing Cadence: Move from monthly to annual billing and you improve working capital (as long as you bill up front). Most SaaS companies offer an annual payment discount, but some skip monthly billing altogether - see case below.

• Billing Timing: Bill up front! I can’t say this enough in the current AI environment. You can totally charge for usage (price metric), but have the pay for the committed volume up front (billing model).

• Payment Terms: Move from Net-60 to Net-30 and free up one full month of operating cashflow. Or require payments by credit card for smaller amounts.

A three-year contract billed annually up-front means you have working capital to fund operations and invest in growth. It also gives you predictability (RPO), which investors value highly. The same deal billed monthly delays cash by up to eleven months. None of this changes ARR or your P&L, but the effects compound on your cash flow statement and can hinder your ability to invest and grow.

The best operators obsess over billing structure as much as pricing design. You can grow cash faster without touching your price model or price point by simply changing how you collect. Let’s look at two brief case studies:

Case: PagerDuty

PagerDuty is a perfect example of how billing terms design drives efficiency:

• Contract and invoice annually

• Collect in advance of delivering the service, and

• Fees are due within 30 days from invoice date (net-30)

• Payments are non-cancellable and non-refundable

The result is strong operating cash flow even in low-growth periods. Over the last two years they increased their cash position by 10+x:

• Revenue grew from $371M in FY23 to $468M in FY25, a CAGR of ~12%

• Operating cash flow jumped from $17M to $180M, a CAGR of ~225+%!

The quote from their SEC filing says it all: “Our cash flows from operations are primarily derived from the unearned portion of customer billings.”

Billing mechanics, not pricing, funded that growth.

Case: Atlassian

Atlassian offers Enterprise customers of Jira and Confluence only annual billing, while they let smaller companies pay monthly. When it comes to Payment terms, they are strictest with the small companies though:

• Net-30 only for purchases $50k+, and only after “a one-time credit check”

• Net-14 for purchases between $1k and $50k

• Orders less than $1k must be paid up front

By enforcing annual up-front commitments for large deals, and strict payment terms for smaller ones, Atlassian locks in predictability and optimizes monthly cash meticulously. This led to impressive growth in cashflow (as well as revenue):

• Revenue grew from $2.9B in FY23 to $4.9B in FY25, a CAGR of ~30%

• Operating cash flow grew from $868M to $1,460, also ~30% CAGR

• Stock price grew from $169 at FY23 start to $207 FY25 end, ~11% CAGR

These two companies prove that billing isn’t a back-office function. It boosts RPO and operating cash flow, which in turn boosts valuation multiples / stock prices. In a market where AI usage is unpredictable, that level of billing discipline gives them a structural advantage, a billing moat.

4. Pricing as a Value Capture Lever

Pricing is about measuring and monetizing value.
While billing determines when you collect, pricing defines what you count.

If you compare the type of metrics companies use to price products, you can generally group them in 3 categories:

• Per User: Microsoft and Salesforce both charge per user

• Per Usage: AWS charges EC3 by time used, Gemini AI top up credit packs are units that can be used fungibly for their Whisk and Flow products

• Per Outcome: Uber introduced surge pricing where value is based on urgency and context, ZenDesk charges its AI agent per Automated Resolution

From a value-based pricing perspective, you always want to use the price metric that best balances these 5 criteria:

• Tracks with Difference in Value Across Segments

• Tracks with Differences in Cost-to-Serve

• Is Easy to Measure and Enforce

• Facilitates Favorable Positioning Versus Competition

• Aligns with How Buyers Experience Value in Use

Those criteria are not new or even from me. They’re from Nagle & Mueller’s book “The Strategy and Tactics of Pricing”.

In SaaS, the dominant model was (and still is) a price per seat and a Good-Better-Best packaging to create an upsell extension path.

That’s changed with AI. Token counts resemble usage metrics to align with cost. The challenge with usage-based pricing is that customers pay more when the AI isn’t giving a good response right away. It misaligns incentives when you pay more for having to try a prompt 20 times to get to the outcome you want.

The real opportunity lies in pricing higher for better effectiveness. A model that completes a task accurately on the first attempt is worth more than one that needs many iterations. That’s the bridge between usage-based and outcome-based pricing. But there are challenges to address like attributability and measurability.

Your pricing determines how much of the value you create you capture. The further you move towards an outcome-based price metric, the more confidence a customer can have that they only pay for what they want.

5. The Combined Framework

When you map billing and pricing as a 3x3 matrix, you see most combinations of the billing models & price metrics are indeed used in today’s world, though some are less common than others (e.g. per user pricing with usage-based billing).

Salesforce changed how customers paid, not what they paid for. AWS changed both, from per-user to per-usage and from prepaid to postpaid. AI blurs the boundaries with some charging recurring for outcomes (Zendesk) and others allowing one-time purchase of credit packs (Gemini).

I hope this framework will help you navigate these volatile times in AI pricing strategy.

Closing Thought

Billing fuels growth through cash discipline. Customers and investors value predictability. Recurring contracts with upfront cash collection reduce risk, fund innovation, and signal maturity. That way billing can become your moat in AI monetization.

Pricing fuels profitability through value capture. Focus on defining the right price metric before fine-tuning the price level. Changing the metric requires more effort than a price increase. Choose the right price metric early that aligns with value in your customer’s eyes, and you usually have a winner.

Use this framework to dissect billing from pricing and steer pricing discussions into what matters: predictable profit that optimizes value capture and operating cash flow.

Every other week, I’ll unpack what drives profitable growth across SaaS and AI native companies. Expect deep dives into pricing and packaging, outcome- and usage-based pricing models, and real-world examples from public and private companies.

The insights I share are grounded in ~10 years in strategy consulting (McKinsey, Booz&Co / Strategy&) and another 8 years leading go-to-market, growth and monetization teams (PwC New Ventures, Okta, project44). My focus: connecting value creation with value capture, and building “flywheels” that scale revenue profitably.

If you’re into growth, pricing, or the future of AI business models, you’ll feel right at home here.

👉 Subscribe now so you don’t miss the first issue.

Then follow me on LinkedIn for more real-world pricing insights and frameworks.

And if you’re a founder or executive, pass it along to your team.

Subscribe now