Congrats to the OpenRouter team for securing this round of funding.
The 5% surcharge for their pricing model may not be palatable to enterprises. In fact, the OpenRouter team could be a pivotal part of the enterprise GenAI stack if they can allow configurable, pluggable endpoints for routing directly to enterprise vetted endpoints to 1P/3P LLM APIs. A couple of large companies I’ve worked so far kinda have this system in place, albeit the dev and maintenance cost and of setting up such an “LLM gateway” could be significantly reduced with OpenRouter. I feel that this is largely an ignored, forgotten part of operating GenAI apps at scale.
I wish I had a straightforward answer for you. This is an important question to quantify, in order to justify the scope and timeline for building features. As an engineer, I always believed that there’s some version of the “no free lunch theorem” applicable to building product roadmaps (and as an extension identifying pertinent product features). Meaning, you can’t reasonably predict what features are going to work, and how well they’re going to work to translate that to $$$.
If I could take a guess, I would build out a framework to accurately guesstimate the value of a feature as follows. A brute-force approach to build out a free —> paid feature funnel: 1. Launch a beta or a “free” version of whatever feature you’re building. This could be as hand-wavy as conducting focus groups and doing proper UI/UX research to validate there’s “some amount” of interest in your user base. 2. Tighten the scope and refine the feature to make it as seamless as possible. Check the app/usage stats and KPIs to ensure people actually derive value out of your features albeit not actually paying for it now. 3. Start charging for it. And see how many people will actually pay for this well-refined feature.
Do this long enough, your company should have some baseline numbers for all of these steps. %conversions on steps 1 —> 2 —> 3 and therefore, better ways to estimate the “why”.