There is a quiet crisis playing out across enterprise software relationships. Vendors celebrate renewal rates. Buyers celebrate token consumption. License counts go up, dashboards turn green, and QBRs are full of charts showing utilization climbing toward the right.
Meanwhile, the actual business outcomes that justified the investment in the first place go largely unmeasured, unmanaged, and underpriced.
This is the trap of vanity metrics, and it is deepening fast as AI reshapes how software is built and sold. The shift to consumption-based pricing, particularly in AI-native products, has made it structurally worse. When you charge per token, you create an incentive for “token maxxing,” where the goal becomes maximizing model usage rather than maximizing business impact. Buyers end up overpaying for activity. Vendors leave enormous value on the table. Neither side builds the trust required for a lasting partnership.
There is a better path, and the companies finding it are reorganizing around a concept we call Outcome Maxxing.
What Outcome Maxxing Actually Means
Outcome Maxxing is not a pricing model. It is a posture. It means orienting every dimension of the client relationship, from discovery through delivery through renewal, around demonstrable, realized business value rather than activity or utilization signals.
The pricing models that follow from that posture, whether milestone-based, gain-sharing, or outcome-indexed, are expressions of the underlying commitment. You cannot credibly price on outcomes you have not measured, and you cannot measure outcomes you have not defined with precision at the start of the engagement.
The journey typically moves through three stages:
- Stage 1: Cost and productivity visibility. You establish baselines, instrument workflows, and begin capturing the measures that will eventually feed a value case. At this stage, you are often still on traditional SaaS or consumption pricing. The goal is to build the measurement infrastructure and trust that make the next stages possible.
- Stage 2: Value pricing. Pricing is set in reference to the value delivered, not the cost of the service. You are not yet sharing risk on outcomes, but your price reflects your confidence in delivering them. This requires a shared value model and ongoing realized-value reporting.
- Stage 3: Outcome pricing. Compensation is indexed to measured outcomes, with contractual mechanisms for attribution, control, and governance. This is the destination for mature, high-trust relationships with sufficient measurement infrastructure.
Most enterprise software relationships never leave Stage 1. The goal of this guide is to show you how to move through the sequence deliberately.
The Measurement Problem: Are We Measuring What Matters?
Before we can talk about pricing, we have to talk about measurement, and specifically about a myopia that afflicts most value programs.
The metrics that are easiest to measure tend to be the metrics that are most peripheral to actual business outcomes. This is not a coincidence. It is a structural consequence of how enterprise software is instrumented. Usage logs are easy to capture. Task completion rates are easy to track. Headcount metrics are relatively easy to benchmark. So that is what most value programs measure.
But consider a canonical example from the current AI wave: deploying an AI system to handle customer service contacts that were previously handled by human agents.
The productivity case is straightforward and measurable: fewer agent hours per resolved contact, lower cost per interaction, faster average handle time. A traditional ROI analysis stops here. The headcount savings justify the investment. The QBR slides look compelling.
What often goes unmeasured is what happens downstream.
AI-handled contacts have measurably different customer experience profiles than human-handled contacts, particularly for complex, emotionally charged, or escalation-prone interactions. When customer satisfaction erodes, it rarely shows up in the cost-per-contact report. It shows up, months later, in churn rates, in NPS trends, in revenue-per-account trajectories.
You replaced agents. You saved on headcount. You lost customers and the revenue they represented. The ROI model was technically correct and practically misleading.
This is not an argument against AI-assisted service delivery. It is an argument that measurement programs built around first-order cost metrics are systematically blind to second and third-order effects on revenue and retention.
The question every outcomes program has to answer honestly: are we measuring for predictability and pricing convenience, or are we measuring for the true shape of the value we are delivering?
The Attribution Chain: First, Second, and Higher-Order Impacts
Here is a useful framework for thinking about the measurement challenge. Business impacts from any technology investment tend to fall into a chain with different properties at each level.
- First-order impacts are direct, proximate, and easy to attribute. Cost avoidance, task automation, process cycle-time reduction, transaction throughput improvement. If your solution automates invoice processing, the reduction in manual processing hours is a first-order impact. These are relatively easy to measure, relatively easy to control, and relatively easy to attribute specifically to your solution.
- Second-order impacts require one inferential step. Productivity improvements that free knowledge workers to focus on higher-value activities. Error reduction that decreases downstream rework. Faster cycle times that accelerate revenue recognition. Attribution is harder because multiple factors contribute. Did revenue accelerate because of your solution, or because the sales team hired five new reps? Controls are harder to establish and often require A/B designs or difference-in-difference analysis.
- Higher-order impacts travel further up the value chain and are the most commercially significant. Risk avoidance from better regulatory compliance. Customer lifetime value changes driven by improved experience. Strategic optionality created by new capabilities. Revenue is influenced indirectly through customer success, retention, and expansion.
The relationship between measurability and commercial significance is inverse. The highest-value impacts are the hardest to measure and the hardest to attribute.
This creates a structural tension that every outcome pricing model has to navigate. If you price only on what you can measure cleanly, you systematically undervalue your contribution. If you price on higher-order impacts you cannot cleanly attribute, you expose yourself to disputes and buyer skepticism.
The answer is not to pick one extreme. It is to build a measurement architecture that captures the full chain, be explicit with buyers about what is attributed directly versus modeled, and price on a portfolio of indicators rather than a single metric.
The Four Barriers to Outcome-Maxxing : Measurement, Control, Monetization, and Attribution
Moving to outcome pricing requires solving four interrelated challenges. Understanding these four barriers clearly is the prerequisite for designing programs that work.
#1: Measurement
The measurement challenge has two dimensions. The first is instrumentation: do you have access to the data that reflects the outcomes you are claiming? The second is validity: does the metric you are measuring actually represent the outcome that matters, or is it a proxy that can be gamed or that diverges from business reality over time?
Token consumption is the obvious failure mode of validity. It measures activity, not output. Output volume may or may not correlate with business impact. In many AI applications, the highest-value interactions are the ones that require the most reasoning, which may consume more tokens or fewer depending on the architecture.
For outcome pricing to work, measurement programs need to go deeper than activity logs and into the business systems where outcomes actually register: CRM for pipeline and revenue outcomes, HRIS for workforce outcomes, customer data platforms for retention and satisfaction outcomes, financial systems for cost outcomes.
#2: Control
Attribution is only meaningful if there is sufficient control to support a causal claim. This is where many vendor value programs overreach.
The cleanest control situations involve well-bounded workflows with clear inputs and outputs, limited confounders, and the ability to establish baselines before deployment. Robotic process automation of a defined transaction type is a high-control situation. Broad productivity improvements across a knowledge workforce are a low-control situation.
Before designing an outcome pricing program, both parties should be explicit about the counterfactual assumptions embedded in the value model. What would have happened without the solution? How confident are you in that baseline? What other factors are operating simultaneously that might account for part of the observed change?
This is uncomfortable territory for most sales conversations. It is also the territory that separates sustainable outcome-based programs from ones that generate disputes at renewal.
#3: Monetization
Assuming you can measure and attribute outcomes, how do you structure the financial relationship?
The options range from outcome-adjusted fees (base plus variable tied to measured results), to shared savings (vendor takes a percentage of documented cost savings), to gain sharing (both parties benefit from revenue-side outcomes), to at-risk structures where a portion of fees is contingent on outcome achievement.
Each structure has different implications for who bears risk and at what point in the engagement. Buyers typically want to minimize risk exposure before they have confidence in the solution. Vendors typically want to participate in upside they have good reason to believe they will generate.
The practical approach for most relationships is staged: fixed fees in early engagements when measurement infrastructure is not yet established, transitioning to value-adjusted structures as confidence and instrumentation mature.
#4: Attribution
Attribution is the hardest problem and the one most likely to generate relationship friction if not addressed explicitly upfront.
Modern enterprise environments are complex adaptive systems. Your AI platform is one of dozens of tools a knowledge worker uses. Your service is one of many inputs to a customer outcome. Claiming that a specific revenue outcome is attributable to your solution, and should therefore be the basis for your pricing, requires assumptions that are often disputed.
The most durable approach combines methodological transparency, governance structures that give buyers oversight of the measurement process, and agreed-upon attribution models established before outcomes are measured rather than after.
The worst outcome is a vendor presenting an attribution claim to a buyer who had no involvement in designing the methodology. That is a recipe for disputes. The best outcome is a joint measurement council, established at the start of the engagement, that governs how outcomes are tracked and attributed.
What Buyers Need to Get Right
For outcome pricing to work, buyers cannot be passive recipients of vendor value claims. They need to bring four things to the relationship.
- Predictability requirements. Buyers operating on fixed budgets have real constraints on variable cost exposure. An outcome pricing structure that could result in a 3x fee increase if results exceed expectations creates planning problems even if it represents excellent value. Buyers should negotiate caps, collars, and predictability mechanisms that allow them to plan, while still giving vendors meaningful upside participation.
- Governance over measurement. Buyers should insist on joint oversight of any measurement program that feeds into pricing. This is not about distrust. It is about the integrity of the model. Measurement systems controlled entirely by vendors with incentives to show positive outcomes will be viewed with skepticism by finance and procurement, regardless of their technical validity.
- Protection against gaming. When a metric becomes a pricing input, it becomes a target. This is Goodhart’s Law, and it applies as much to outcome pricing as to any other incentive structure. Buyers should think carefully about whether the metrics in an outcome program can be gamed, and what governance mechanisms prevent that.
- Resisting the temptation to overpay for noise. Not every positive business result that occurs during a vendor engagement is attributable to that vendor. Buyers should maintain rigorous attribution standards and resist pressure to credit vendors with outcomes they did not drive. This requires an internal measurement capability, not just reliance on vendor-provided reporting.
- What Vendors Need to Get Right
Vendors moving toward outcome pricing face their own set of challenges.
- Revenue predictability. Pure outcome-based models can create significant revenue volatility, which creates problems for vendor financial planning and investor expectations. The practical solution is a portfolio approach: a mix of base fees and variable components that provides floor-level predictability while preserving upside participation.
- Investment in measurement infrastructure. You cannot run an outcome-based program on spreadsheets and quarterly surveys. Vendors serious about this transition need to invest in the tooling, processes, and talent that make ongoing realized-value reporting credible and efficient. This is a competitive moat: the vendor with the best measurement infrastructure has the most defensible value story.
- Honest scoping of attribution. The fastest way to destroy a promising outcome-based relationship is to claim credit you have not earned. Vendors should be explicit about what they are attributing to their solution, what assumptions underlie that attribution, and where uncertainty is high. Buyers who feel they are getting an honest picture, even when the picture is complex, are far more likely to expand the relationship than buyers who feel they are being sold a simplified story.
- Differentiating value tiers. Not all outcomes are equal, and not all use cases are ready for outcome pricing. Vendors should segment their portfolio: high-control, easily measurable use cases are candidates for outcome pricing early. Complex, high-order-impact use cases may need to stay on value pricing until measurement infrastructure matures.
Making the Transition: A Practical Sequence
Moving from SaaS or consumption pricing toward outcome pricing is not a flip-the-switch event. It is a multi-quarter, often multi-year transition that requires building trust, measurement infrastructure, and governance frameworks in parallel.
Here is how the most successful transitions tend to unfold.
- Quarter 1-2: Baseline establishment. Before claiming outcomes, you need to know where you started. This means instrumenting the relevant business processes, establishing historical baselines, and agreeing with your buyer on what will be measured, how, and by whom. Many value programs skip this step and pay for it at renewal.
- Quarter 3-4: First-order measurement. Begin capturing the clean, high-control metrics: cost avoidance, transaction volumes, process cycle times, error rates. Report these regularly in a realized-value dashboard that buyers can access and interrogate. Build the habit of evidence-based value conversations.
- Year 2: Second-order analysis. With a year of data, you can begin analyzing second-order effects. Are the freed knowledge-worker hours actually being reinvested in higher-value activities, or is the productivity benefit being absorbed without a corresponding business outcome? Are the error reductions translating into downstream rework savings? This is where more sophisticated measurement methods, including A/B analyses and difference-in-difference modeling, become useful.
- Year 2+: Value-adjusted pricing negotiation. Once you have a credible body of evidence, you have the foundation for a pricing conversation grounded in delivered value rather than vendor cost structure or market comps. This does not mean the buyer will immediately agree to pay more. It means you are having a substantively different conversation, one about the shape of the value you are both creating.
- Year 3+: Outcome pricing pilots. For use cases where measurement is mature and attribution is defensible, introduce contractual outcome linkages. Start with upside sharing before introducing downside risk. Build experience with the governance mechanics before making outcome pricing the primary commercial structure.
With AI accelerating everything, perhaps the demand on these timelines get condensed, but change can take more time to execute than we initially anticipate.
The Bottom-Line
The companies that master enterprise software economics over the next decade are the ones building the capabilities today to measure, attribute, and price on outcomes rather than inputs.
This is not just a pricing strategy. It is a product strategy, a customer success strategy, and a trust-building strategy.
Vendors who can demonstrate realized value with rigor earn the right to expand. Buyers who build the internal capability to evaluate outcome claims gain negotiating leverage and avoid the trap of paying for activity they cannot assess.
The move away from SaaS, usage pricing and token-maxxing to Outcome Maxxing is ultimately a shift in what both sides of the table are optimizing for. And it turns out that when buyer and vendor are both optimizing for the same thing, the actual business result, the economics of that partnership tend to look very different from the ones built around utilization dashboards and renewal discounts.
Start with the measurement. The pricing follows.
Additional Insights:
From Token Maxxing to Outcome Maxxing: Salesforce Just Proved the Outcome Economy Is Real
From Features to Outcomes: What Deloitte’s Software Industry Outlook Signals About the Outcome Economy
Buyers and the Outcome Economy: Gartner’s 2026 CFO Priorities for GTM Teams
Let’s discuss the four barriers and how to overcome them to meet Outcome Economy demands and drive Outcome Maxxing with customers: Click here to schedule a consultation with us.