Understand Users by Watching Failures in the Funnel
Watch real users, and interview the ones who failed in your funnel — they know what you need to fix.
- Difficulty
- Easy
- Time to result
- ~weeks to results
- Steps
- 4
- Confidence
- 90%
Product owners are usually innovators or early adopters and can't intuit how everyone else behaves, so they must observe. Levine's method: watch new users, ask 'why' when they don't do what you expect, and deliberately interview the people who dropped out of the funnel rather than the successful users at the bottom.
Origin
Uri Levine, from the 'Understanding Users' chapter. Illustrated by Steve Wozniak's reaction to Levine using the iPhone volume button to shoot a selfie.
Core principles
- 01There is no right or wrong way for people to use a product — different people use it differently.
- 02If you build only for yourself and don't know others differ, you build the wrong product.
- 03The 'why' behind unexpected behavior is what makes the next version successful.
- 04To improve, study the people who failed in the funnel, not the ones who succeeded.
How to run it
- 1
Watch new users
Simply observe new users and see what they actually do with the product.
Pro tip Look for surprising uses — they reveal that people aren't like you.
- 2
Ask why when behavior surprises you
When users don't do what you expected, ask why; that 'why' drives the next version.
Watch out Assuming users are like you leads you to build for an audience that doesn't exist.
- 3
Map the funnel of use
Lay out the stages: download/visit → register → first use → get value → come back. Locate where people drop out.
- 4
Interview the failures, not the successes
Instead of talking to satisfied users at the bottom, interview those who didn't register, didn't use, or didn't return — they know what you need to know.
Pro tip Target people far down the funnel who still bounced, so you learn real friction rather than pure non-fit.
Watch out Don't waste time on people who were never a fit for the product.
In the wild
Levine shot a selfie with Steve Wozniak using the iPhone volume button; Wozniak said 'finally... someone using it the way that I meant it to be.'
→ Demonstrated there's no single right way to use a product — people diverge, and builders must observe it.
Levine notes that if he polled a large audience, ~70% watch the Waze screen while ~20–30% follow audio prompts — and each group had no idea the other existed.
→ Shows builders must watch real usage because users don't realize (or report) how differently they behave.
Common mistakes
Interviewing only successful users
Happy users at the bottom of the funnel can't tell you why others dropped out — the churned users hold the improvement insight.
Building for yourself
If you're an innovator/early adopter and assume everyone is like you, you build the wrong product for the majority.
Is it for you?
Best for
Product owners and founders iterating toward product-market fit who need to reduce funnel drop-off.
Not ideal for
Chasing feedback from people who are simply not in the target market and will never be a fit.
From the transcript
“watch new users simp simply watch users and see what they're doing and number two if they are not doing what you expect them to…”
“in order to improve we need to speak with those that fail”
“there is no right or wrong there are different people that are using different products in a different ways”
From the episode
Lessons from a two-time unicorn builder, 50-time startup advisor, and 20-time company board member
Uri Levine (co-founder of Waze)