User-Centered Performance Detector
Spot the rituals where a team signals customer obsession without any intent to change the decision.
- Difficulty
- Easy
- Time to result
- ~weeks to results
- Steps
- 4
- Confidence
- 96%
Antin's coined term for customer-obsession that is symbolic rather than learning-oriented: work done to signal to each other how user-centered we are, not because anyone intends to make a different decision. The framework names the explicit tells (late-stage 'quick validation' studies, executive listening sessions, focus groups run for the exec's benefit) and the implicit ones (confirmation bias, ego, intuition-protection), and offers the antidote mantra: we don't validate, we falsify.
Origin
Term invented by Judd Antin ('it's fun to make up terms') and popularized in his 2023 essay. He explicitly names its cultural sources — IDEO's design thinking and Bezos's 'customer obsession' — as the doctrines that turned genuine practice into ritual.
Core principles
- 01Customer obsession that is symbolic rather than focused on learning is performance, not practice.
- 02A study requested at the end of a product process is a box-check, not a question.
- 03We don't validate, we falsify — the goal is to be shown wrong, fast, and specifically.
- 04Executive listening sessions are ~97% performance: well-intentioned, not learning-oriented.
- 05It takes two sides — researchers, PMs, designers, and the org structure all sustain the performance.
How to run it
- 1
Run the timing test on every research request
Ask when in the product process the request arrived. If it arrives at the end — 'can you just run a quick user study to validate our assumptions' — the decision is already made and shipping is already scheduled. That is performance.
Pro tip The word 'validate' in a request is the single loudest tell.
Watch out The requester is usually sincere. Performance is not the same as cynicism, which is why it survives.
- 2
Audit the implicit rituals
Catalogue the recurring user-centered activities — exec focus groups, listening sessions, dogfooding shares, voice-of-customer readouts — and for each, ask what decision it has changed in the last two quarters. Anything with no decision changed is a signalling ritual.
Watch out Do not delete these rituals outright — reframe them so the participant is exposed to disconfirming evidence rather than confirming anecdotes.
- 3
Replace 'validate' with 'falsify' in the team's vocabulary
State the ask as: 'do research that shows we were off base, tell me exactly how and why, in a way that lets me fix it quickly.' Make being wrong early the explicit, rewarded outcome.
Pro tip Leaders must model it first — the person with the most to lose from being wrong sets the ceiling on the whole team's honesty.
Watch out Many PMs and designers do not want to be wrong. This step will surface who those people are.
- 4
Move the research upstream so falsification is still cheap
Performance persists because research arrives too late for a different decision to be possible. Insert the researcher at the framing stage so being wrong is affordable.
In the wild
A PM approaches a researcher after design is locked and asks for a quick study to validate assumptions. Antin's read: it is too late to matter, the PM has no interest in being wrong, and what is actually wanted is a checked box.
→ The study confirms what the team already believed, nothing changes, and executives later conclude research is not impactful.
An engineering team wanted to use post-hides as the signal of post quality. Antin found hiding was power-law distributed, brought a heavy hider into a one-way-mirror think-aloud session, and insisted the directors and product people sit behind the glass. The participant hid every story she finished — she was chasing inbox zero on an infinite feed.
→ An n-of-1 burst the room's bubble; the team abandoned the naive hides-as-quality signal and built a better one. The point was not statistical significance — it was that the decision-makers were in the room.
Common mistakes
Assuming you personally don't do it
Antin's direct challenge: if you're thinking 'I don't do that,' think about it for a second. The behaviour is extremely common and well-intentioned, which is exactly why it is invisible from the inside.
Blaming researchers for the performance
The performance is a property of the system — PMs who want cover, designers following process orthodoxy, execs who don't know what research is for. Fixing it requires all sides plus a structural change, not a better researcher.
Is it for you?
Best for
Product leaders and researchers at companies that loudly profess customer obsession but ship on gut anyway
Not ideal for
Very early startups where the founder is legitimately the user and decisions must move faster than evidence can
From the transcript
“customer Obsession or user centered practice that is symbolic rather than focused on learning”
“every time a PM comes to a researcher at the end of a product process and says can you just run a quick user study”
“this is 97% performance it's well-intentioned but it isn't focused on learning”
“mantras was we don't validate we falsify right we are looking to be wrong”
From the episode
The UX research reckoning is here
Judd Antin (Airbnb, Meta)