Brian Chesky on customer feedback: understand a few people deeply

Follow the customer’s journey

In the Stanford conversation Designing a 10-star Experience, Chesky recalls advice from Paul Graham about a small group of people caring deeply about a product. He then describes studying one person’s journey, improving it, and continuing to iterate before trying to repeat the experience for more people.

The distinction is between designing a meaningful experience and solving the separate problem of scaling it. The talk does not establish that a personalized service is automatically economical or that every request should become a feature.

Recording [1]

Keep the attribution attached to the story

In the startup lecture on culture, Chesky again credits Graham for the advice about a small group of enthusiastic users. This is a useful example of why speaker attribution matters: Chesky is explaining advice he received and how he understood it. Describing it as an original Chesky quotation would lose part of the record.

That lecture also includes Alfred Lin. A passage from the same video is not necessarily Chesky speaking. The linked timestamp is in Chesky’s account of early users, rather than the opening culture presentation.

Recording [2]

Turn feedback into a concrete investigation

The following checklist is Timbre’s editorial application. It is intended to make customer conversations more specific, not to suggest that either speaker prescribed this exact process.

• 01Choose a customer whose situation you can understand in detail.
• 02Map what happens before, during, and after they use your service.
• 03Ask where they hesitated, improvised, or needed help.
• 04Change one part of the experience and check whether it solved the problem.
• 05Separate what would improve the experience from what would make it repeatable and sustainable.

Use feedback without promising to build everything

A customer’s account is evidence about an experience. It is not automatically a specification for the solution. Bring the user’s goal, the observed friction, and your operating constraints into a follow-up question. Then compare any generated suggestion with the source discussion and with evidence from your own customers.