# Continuous broad A/B optimization versus hypothesis-driven testing only.

**Practitioner:** Brian Chesky
**Source:** Lenny's Podcast — Brian Chesky episode
**Timestamp:** 18:26
**Move:** 5 of 8

## The tension

Continuous broad A/B optimization versus hypothesis-driven testing only.

## Where they landed

Extreme hypothesis-only — no blind A/B testing.

## What they accepted to lose

Loss of continuous optimization velocity. Growth team forfeits a familiar lever for incremental conversion gains.

## Verbatim

> if we do an AB test, there has to be a hypothesis. If we don't have a hypothesis and A is better than B, then we're stuck with B.

— Brian Chesky, Lenny's Podcast, 18:26

## Load-bearing phrase

*if A is better than B, then we're stuck with B*

## What it pushes back against

The standard growth practice of broad continuous A/B testing.

---

**Source attribution:** This call is drawn from Lenny Rachitsky's public podcast archive at [LennysData.com](https://lennysdata.com), released April 2026 with an explicit invitation to builders.

**Verify on votu.app:** [https://votu.app/move/chesky-5](https://votu.app/move/chesky-5)

**License:** Verbatim quotes are © the original speaker; attribution and timestamp guaranteed accurate to the source transcript.