A thought about leverage, entry points, and experiments
A loose note on Uber, experiments vs pricing, and why the front door matters more than the spreadsheet
I was watching the Dara x Nilay interview and there was this one bit I can’t shake.
Nilay asks him something like:
If another app wanted to let people book Ubers through them, what would you charge?
And instead of the usual MBA answer — take-rates, margins, CAC math, etc. — Dara basically goes:
start with zero, try it, see if it actually brings new people in or just reroutes the same demand, and then decide what it’s worth.
The order matters there.
Don’t decide value before you know if value exists.
Anyway, someone then raised a much messier question:
What if the “other app” is iOS or Android?
What if ride-hailing becomes a system-level thing — like “Siri, get me a ride” — where you never open Uber at all?
And that’s where the whole thing breaks open. Because you can experiment with another app. You can renegotiate later. But if the OS becomes the front door, Uber’s leverage disappears. No loyalty surface. No upsell moment. No tap.
And that’s what I’m sitting with.
As someone who works in finance/strategy, I’m used to thinking in unit economics, take rate deltas, cannibalization curves, contribution profit — all the nice quantifiable stuff. But none of that matters if you lose the point of entry. You can’t price what you no longer control access to.
So my brain ends up here:
• Try things before modeling them to death.
• Price once you know what’s real.
• Be careful who owns intent.
Not a grand lesson, just something I’m turning over.
Maybe the real “strategy work” isn’t about getting the model right — it’s about making sure you still get to have a model in the first place.