# Consumption-based pricing and changing value roadmaps ## Current Form ### Agentic software agents will change our relationship to software A fundamental part of how we commercialize software is our ability to understand *why* any given software is valuable to a set of users, and then clearly articulate why it's valuable. Over the years, we've seen a few big shifts in how humans interface with technology, and with each shift, how we perceive value has shifted. I've always loved music--it was a gift I got from my dad, a musician. The go-to-gift every Christmas was a Best Buy gift card. Whenever I got paid, a trip to stroll the Best Buy aisles was always on the agenda. Cruising through the catalog, looking for "For fans of" stickers (themselves a form of Uniframing) to try and discover my next album. It was a simple time. Cultivating a love of music was characterized by the exploration and curation of physical media. Of course this was largely upended by the widespread digitization of media, and streaming. Or was it? The strong and continued adoption of vinyl collections, and now CD collections suggests otherwise actually. Consider the growing pushback toward "algorithmically-driven" suggestions and the recursive nature of those recommendation engines. This is the point I'm making. In a pre-streaming era, the physical object (CD, Vinyl) was the primary value storage mechanism in your relationship to music. If you wanted the tunes, you needed the object (or maybe like me you would set your VCR to record TRL so you could have the latest Marilyn Manson track since your parents wouldn't let you buy it). What we think of as streaming services now (a classic example cloud-based software-as-a-service models) doesn't sell you the objects (songs, albums)--they sell you access to all of the songs! Manson on demand. Selling access has been the primary blueprint of SaaS for the last 15 years, more or less following an "Is, Means, Does" framework. The is/means/does framework is an example of a three-tiered value roadmap that serves as the foundation for how to build a message hierarchy. In the era of selling access, this has been something a long the lines of: here's a software that does this thing, that means now you can X, and this is what it unlocks for you/your organization. > Spotify is the largest online music streaming service. You can listen to any artist, on demand, so the music of the world is at your fingertips. ### All of this illustrates a paradigm shift for knowledge work We're navigating a massive paradigm shift in knowledge work right now, and that's no more apparent than in the SaaS industry. I'm not going to debate the morality of agentic AI or whether or not the purpose of AI is to replace people. My position here is simple: as a technology, it's not going back into the box. And despite what most marketing materials would have you believe, many companies are looking to agentic AI to *do jobs for them.* Fundamentally, whether you want to paint agentic AI as a tool to "lower the floor and raise the ceiling," it's still *doing jobs.* The modality in which we relate to technology is changing, and with it, the way that we perceive the value of said technology is evolving. Where before there was a business model built on access, there's now a business model built on output. In the same way that value storage moved from the physical to access to the digital, we're now shifting from *access* to the *successful completion of work*. Unless you're building foundational models (you're not), this will fundamentally change how consumers value your product. It's no longer about what your software enables for them directly, but what your agents will do for them. The products that will win in an AI-era are already changing the foundations of their marketing system: 1) value roadmaps based on agents working for users 2) pricing and packaging built on consumption and successful completions 3) positioning and messaging beyond the rote "AI for X" platitudes 4) building new models for forecasting variable consumption and revenue ## Spark I've been thinking about this idea a lot over the last few quarters, but it came from some osmosis from [[Kyle Poyar]]'s work on the rapid evolution of SaaS pricing in the AI-environment. The shift to outcome and/or success-based pricing is interesting, and represents a fundamental shift in essential value storage for software. ## Tensions or Unknowns This still feels like it needs to grow, conceptually true, but needs more tangible and practical connection to the work we're all doing, especially in [[product marketing]]. ## Next Evolution Thinking on this. ## ChangeLog [[2025-04-16-Wednesday]] - Updated body structure and frontmatter