Stop Running Discovery Like a Checkbox

I watched a product team do "discovery" with their biggest customer.

They'd already designed the feature. Built wireframes. Named it. Now they just needed validation.

So they did an interview. Asked leading questions. Got the validation they needed. Shipped the feature.

Six months later, nobody used it. Turns out the customer had said yes to be polite, not because they actually needed it.

The team had done the checkbox. They hadn't done the work.

How Discovery Theater Works

Real discovery is uncomfortable. You ask questions you don't know the answer to. You sit in silence. You listen for what people don't say.

Theater discovery is much easier. You ask questions designed to confirm what you already believe. You keep talking so there's no silence. You rationalize away anything that contradicts you.

Theater discovery looks like:

The Leading Question: "Would you use a feature that let you do X faster?" Everyone says yes. X is obviously useful. You've learned nothing.

The Biased Sample: You interview power users. Of course they love your product. They're not representative. They're the exception.

The Selective Hearing: The customer says "I don't have time to use this." You hear "this is useful but they're busy." You write it down as validation.

The Confirmation Collection: You interview 12 people and listen really hard for the 8 who agreed with you. You remember those. You forget the 4 who didn't.

The Premature Certainty: You interview three people, you're already sure you're right, but you schedule four more interviews to look legitimate. By interview seven, you're annoyed because nobody's confirming your theory.

I worked with a team that interviewed customers before redesigning their onboarding. Twelve interviews. They came back certain: users wanted a step-by-step wizard.

I asked to see the notes. Three-quarters of the way through, every interview went like this:

Interviewer: "Do you think a step-by-step wizard would help?" Customer: "Maybe? I'm not sure." Interviewer: "It would walk you through each step." Customer: "Oh, sure, that could work."

Not a single person had asked for it unprompted. But the interview process had basically answered the question for them.

What Real Discovery Looks Like

Real discovery is:

Open-ended: "Tell me about your workflow." Not "would feature X help?"

Surprising: You learn something that contradicts what you believed. And instead of rationalizing it away, you change your mind.

Generative: You listen for problems you didn't know existed. You might come back with five hypotheses instead of one.

Humble: You sit with customers who know more than you. You learn from them, not validate against them.

Time-consuming: Real discovery takes weeks, not days. Because good data isn't quick.

I watched a team redesign their data export feature. They didn't ask: "Do you want a better export?" They asked: "How do you use the data after you export it?"

That question unlocked an entirely different problem. Users didn't want better exports. They wanted to see the data in real time without exporting. The export was a workaround for something they actually needed.

If they'd done theater discovery, they would've shipped a shinier export. Because that's what they'd decided to build.

Instead, they shipped something completely different. Because they actually discovered.

Why Theater Discovery Wins In Organizations

Theater discovery wins because it's fast and it doesn't threaten anyone.

Real discovery sometimes shows that your idea is wrong. Theater discovery shows that your idea is validated. Theater wins internally.

But theater loses with customers. Because customers eventually figure out that you weren't actually listening.

I worked with a leader who ran discovery the theater way. "Did interviews, got validation, shipping." Always confident. Always quick. Almost always wrong about what customers actually needed.

He had great internal political capital. Terrible product-market fit.

Then we switched how we ran discovery. Longer. Less biased. More willingness to change direction. Slower internally. Much better products.

The org eventually figured out that slow discovery leads to fast shipping. Theater discovery leads to shipping the wrong thing fast, then rebuilding.

How To Know If You're In Theater Mode

Here are the tells:

You did interviews but came back with the exact hypothesis you started with You kept interviewing until you found someone who agreed Most of your interview notes are quotes that validate your position You haven't changed your mind about anything since you started Your interviews took two hours total (real discovery takes hours per person) Your sample is all similar (power users, tech-savvy people, people like you) You're writing up the findings before you've finished interviewing

Any of those? You're running theater.

What To Actually Do

Stop thinking of discovery as validation. Think of it as investigation.

You're not trying to confirm your brilliance. You're trying to find out what's actually true.

Ask questions you don't know the answer to. Sit in discomfort. Change your mind. Let customers teach you things.

Do less discovery, but do it well. Three deep interviews from diverse users beats twelve leading interviews with confirmation bias.

Then build based on what you actually learned, not what you wanted to hear.

That's the difference between shipping features people use and shipping features that validate your internal politics.

Most teams choose politics. That's why most products disappoint.