Changing your business model is not failure. Refusing to change it when the evidence demands it — that is the failure.
I have changed my business model twice in fourteen months. The first time felt like a public admission of defeat. The second time felt like a natural consequence of learning. I am hoping that by the time I need to change it a third time — and I expect I will — it will feel like standard operating procedure.
I want to write about both changes in detail because I think the actual texture of business model iteration is much messier and more instructive than the clean pivot narrative that founders usually tell after the fact.
Change one — from ecosystem sales to direct founder sales
My original commercial vision was to sell to startup ecosystems rather than to individual founders. Accelerators, incubators, venture studios, university entrepreneurship programmes — organisations that work with cohorts of early-stage founders and need tools that support structured validation processes.
The strategic logic was sound on paper. A single contract with a regional accelerator could bring in thirty to fifty active users at once. The CAC would be low relative to the lifetime value. I would not be grinding out one-by-one acquisition. I would be doing institutional sales.
The reality was that institutional sales cycles are measured in quarters, not weeks. The decision-makers I was dealing with had budget cycles, procurement processes, committee approvals, and competing priorities. The fastest deal I closed in the institutional channel took eleven weeks from first conversation to signed agreement. During that time I burned through a meaningful percentage of my runway.
More importantly, the feedback loop from institutional customers was slow and distorted. When an accelerator buys a tool for its cohort, the accelerator staff are the buyers but the founders are the users. The feedback I was getting from the buyers was about programme administration and reporting. The feedback from the actual users — the founders — was buried two layers deep and filtered through coordinators who had their own interpretive lens.
I needed direct contact with the people using the product. I needed short feedback loops. I needed to understand, in real time, which parts of the product were creating value and which were creating friction. The institutional channel was the wrong environment to build that understanding.
So I stopped pursuing institutional deals and switched entirely to direct-to-founder sales. Smaller transactions. Faster decisions. Completely unmediated user feedback. The unit economics looked worse on paper. In practice they were better because the product improved so much faster.
Change two — from flat fee to usage-based
I have already written in detail about the pricing model evolution. The short version for continuity here: the flat fee model that I moved to after abandoning the subscription model solved the wrong problem. It removed the recurring commitment that founders resisted, but it introduced a different friction around project initiation that was subtly distorting which analyses founders were willing to run.
The move to credits was driven by a specific observation: the founders who were getting the most value from the product were treating it as a thinking partner for decisions — reaching for it whenever a significant business question needed structured examination. The founders who were getting less value were treating it as a service provider — going to it when they had a defined deliverable in mind. The pricing model was reinforcing the service-provider mental model because it charged per defined project.
Credits changed the framing. You are not purchasing a project — you are investing in the quality of a decision. That framing shift sounds small. It changed the usage pattern significantly.
What both changes had in common
Looking back from where I sit now, both business model changes were triggered by the same underlying failure mode: I had been designing the commercial model for the customer I wished I had rather than the customer I actually had.
The institutional customer was simpler to model, more prestigious to reference, and easier to imagine scaling through enterprise sales motions. The direct founder customer was noisier, more price-sensitive, harder to reach, and more demanding. But the direct founder customer was real and the institutional customer, at my stage, was a projection.
Every time I get the business model wrong, I know now to ask: whose convenience am I actually optimising for? The answer is usually mine. And the fix is usually to go back and ask the actual customer what would make the exchange feel completely fair to them.