LLenny's Podcast
← All frameworks
StrategySachin Kansal

The End-User Feedback Spectrum

Deliberately gather user feedback across the full quantitative-to-qualitative range, not just the end you're comfortable with.

Difficulty
Easy
Time to result
~ongoing to results
Steps
4
Confidence
92%

A mental model that arranges all methods of understanding users on one axis: from pure quantitative (dashboards where every user is a number) at one extreme, through surveys and focus groups, to pure qualitative (one-on-one and firsthand experience) at the other. The insight is that no single point on the spectrum is sufficient — numbers show behavior and friction but never the visceral emotion, and emotion is a motivator numbers can't supply.

Origin

Sachin Kansal's articulated framework for why Uber invests across all feedback methods rather than defaulting to Uber's historically data-driven culture.

Core principles

  • 01You need to operate across the entire spectrum, not camp at one end
  • 02Quantitative data tells the story of what users do; it never shows what they feel in the moment
  • 03Emotion carries an intensity and motivating force that data cannot
  • 04The goal is a culture where emotion and numbers coexist, not compete ('it's not an or, it's an and')

How to run it

  1. 1

    Anchor the quantitative end

    Start with the dashboards where every user is an MAU or DAU — collective behavior across large numbers reveals real patterns and is genuinely valuable.

    Watch out If understanding users becomes purely a mathematical exercise, it is not enough — you'll miss what they actually feel.

  2. 2

    Move inward with surveys

    Send surveys to thousands of users to quantitatively understand preferences (do they like A, B, or C) at scale.

  3. 3

    Run focus groups and roundtables

    Put ~10 users in a room for an hour of discussion. Kansal does at least one or two driver roundtables (and rider roundtables) every time he visits a Brazil, India, or UK office.

    Pro tip Do these on-location in your major markets so you hear region-specific realities directly.

  4. 4

    Go one-to-one and firsthand

    Have direct one-on-one conversations, and ultimately experience the product yourself — riding, driving, delivering — to capture the visceral reaction that only firsthand use produces.

    Pro tip A daily rider commute is a free one-on-one interview plus a chance to watch how the driver actually uses the app.

    Watch out Until you get behind the wheel yourself, you miss the visceral reaction you get when something actually happens in context.

In the wild

Watching a driver leave the app for a text

As a passenger interviewing his driver, Kansal watched a text message come in, the driver leave the driver app to address it, then return — an 'aha moment' about real usage he'd never see in a dashboard.

Firsthand observation surfaced an interaction-flow problem invisible to quantitative analysis.

The human touch the office ignores

Firsthand driving revealed that when a passenger enters the car, the center of the driver's universe is the human interaction with the passenger, not the app Uber obsesses over. A friendly passenger lifts the mood; a rude one is demotivating.

A whole category of driver experience (the human interaction) that no metric captured became a recognized design consideration.

Common mistakes

Living only at the quantitative extreme

Treating every user as a number and optimizing funnels alone reduces understanding to a mathematical exercise that never reveals the in-the-moment emotion, which Kansal argues is the strongest motivator for building better products.

Assuming one archetype represents all users

Kansal warns that even in a single country there are different demographics and socioeconomic statuses; understanding 'one archetype' of user is not the same as understanding your actual user base.

Is it for you?

Best for

Product leaders deciding how to invest limited research time and wanting to balance a data-heavy culture with qualitative depth

Not ideal for

Teams that need a fast single-signal decision and can't afford to run multiple feedback methods

From the transcript

There's the extreme quantitative end of the spectrum and there's the extreme qualitative end of the spectrum. And I like to believe that you need…

12:00

every user is an MAU or a DAU

12:00

maybe I have a room of 10 drivers and I'm asking them it's a focus group

12:30

until I get behind the wheel or I am a passenger or I am using Uber Eats, what I miss is the visceral reaction that…

13:30

I think it's not an or it's an and

20:30

From the episode

Why Uber’s CPO delivers food on weekends

Sachin Kansal