What's the key difference between Eppo and say, Optimizely?

Optimizely, vwo, abtasty type of tools were built to allow self serve experimentation to marketers, and don’t provide the depth and quality of analysis that most product experiments need. They create separate data planes that usually limit the type of metrics you can use, leading to most tech firms just rolling their own analysis.

Eppo is built for product experiments and data teams, with an emphasis on trustworthy, deep analysis. The initial data/metrics plane is sourced from data warehouses via SQL, and the workflow and stats engine is built so that engineers, PMs, designers can self serve experiments while making statistically sound, scientifically valid decisions.

Let me know if that makes sense, happy to elaborate

Ah, I see. Does Eppo also manage the exposure part of setting up experiments?

At Lyft, we have an internal platform that helps us run and manage experiments. Seems like this is a SaaS equivalent. Do you have an idea of how it compares to other companies established experimentation platforms?

Yup, Eppo will do the randomization/exposure part of the experiments. But we also can be (and are often) paired with existing randomization tools, whether it's LaunchDarkly, Optimizely, or in-house capabilities.

And I'd bet that there is a lot in common with Eppo and your tooling at Lyft. :) The product was inspired by what we built at Airbnb, and what we saw at Netflix, DoorDash, etc. An easy way to understand Eppo is "a SaaS experimentation tool that feels like what you had at Airbnb/Lyft/Dropbox."

This sounds great! Best of luck!