--- title: "What’s wrong with outcome-based pricing models?" pubDate: 2024-12-16 description: "How they can hold customers back and work against them." author: name: 'Bartek Kuban' image: '/blog-assets/authors/bartlomiej_kuban.webp' image: url: '/blog-assets/posts/outcomePricing_bg.png' alt: "Background image for What’s wrong with outcome-based pricing models?" tags: ['insights'] --- Outcome-based pricing is one of those ideas that looks very promising, but can get really messy when you’re trying to implement it. Especially with enterprises. ## Road to billing headaches is paved with good intentions The central idea behind outcome-based pricing is aligning incentives between vendors and customers. In theory, the vendor only gets paid if they deliver value, which should motivate them to perform at their best. Of course, it very quickly brings us to what the aforementioned “value” means. Let’s take the most popular enterprise AI use case — customer support — and the seemingly straightforward notion of paying for “resolved” customer tickets. We all have experience talking to support representatives and chatbots, and we know that our “leaving the conversation without seeking additional help” very often can’t be counted as a resolution. **In many outcome-based pricing models, every situation where there’s doubt about whether the issue was resolved is decided in favor of the AI vendor.** If a customer leaves the support chat out of frustration with poor responses, the AI vendor may still treat it as a resolved case and charge accordingly. And even if the issue was indeed resolved, was it worth the fixed $0.99 price- per-resolution? ## Good for one, bad for another It gets even more complicated when we add what _successful_ resolution means. What one company considers a successful resolution might feel incomplete or inadequate to another. For some, speed is the key metric — give a quick, efficient answer and cross it off. But some brands treat customer experience as a brand differentiator. For them, the stakes are higher, as CX is a critical element of their value proposition. When brands double down on CX quality, their expectations for “successful” quadruple. In theory, these questions can be addressed contractually, but in practice, no agreement is airtight. Edge cases arise constantly, forcing vendors and customers back to the negotiating table. Hardly a seamless relationship. ## Who gets the credit? And then there’s the matter of attribution. If a sale closes or a meeting is booked after an AI agent interacts with a lead, who gets the praise? The AI agent, the sales team, or the product itself? Attribution disputes are inevitable in outcome-based models, particularly in complex enterprise environments where multiple factors drive results. When the invoice comes at the end of the month, what appeared to be a simple definition, suddenly becomes a subject of intense back-and-forth between the client and the account team. Results-based pricing models can work for products where the cost of the tool itself only scales with the customer’s business growth, like Stripe’s transaction fees. In these cases, the variable cost feels justified because it scales in proportion to the customer’s revenue, essentially converting a cost into a net-positive effect on the bottom line. But many solutions, particularly in AI, don’t have such clear-cut value propositions yet that align with customer’s revenue or savings. Large companies don’t want their costs tied to unpredictable and hard-to-measure outcomes. In our experience, they would rather have budget certainty. While outcome-based pricing is sold as fair and flexible, its unpredictability often leaves CFOs uneasy. ## Fear of the unexpected The chargeable volume is one thing, but also every new use case or system improvement requires a renegotiation of terms. A company that improves its support process and, as a result, reduces ticket complexity might not see the financial benefit unless it renegotiates the vendor’s pricing structure. This point is very important to me personally, as it relates to innovation. I believe the ultimate goal is to make AI a fundamental part of every company’s operations while providing the certainty needed to confidently put it in front of customers. The fear of unexpected costs can lead to the opposite result — defaulting to tried-and-tested legacy solutions and hindering progress. It can make companies view AI as a cost to minimize rather than a tool to drive radical innovation. When pitching to enterprises, you’re not only competing with companies like you, but also with the enterprise’s internal technology resources. Stakeholders’ thought process often boils down then to a simple dilemma: “buy” vs “build.” Enterprises evaluate whether purchasing a solution offers enough value to justify its cost compared to the costs of building and operating an equivalent solution in-house. Lack of transparency and predictability doesn’t play to anyone’s advantage. Other pricing models can _actually_ align incentives by encouraging customers to do what we all need the most — use AI solutions as much as possible, generate more conversations, insights, and value without fear of escalating costs. Outcome-based pricing might be trendy now, but trends don’t always translate to sustainable business models. And unsustainable business models can’t serve as a foundation for stronger, longer-lasting partnerships. At the end of the day, what matters most to customers is what they get for their money. And no matter how creative the pricing model is, it’ll always have to face reality by answering a simple question from your customer: “So, does this mean I’ll pay less?”