Why Most UX Research Gets You Nowhere

I’ve worked with over 30+ different startups now. I’ve watched so many designers do user interviews. And I’ve noticed one thing:

Your users are lying to you.

They don’t even know they’re doing it.

It’s not on purpose. They want to help. But UX research as we know it is broken. It gets you polite lies, not hard truths.

Here’s what I saw last week:

Three months later? Less than 3% of users used the feature more than once.

What the f*ck happened?

The Psychology Behind User Lies

Let’s understand why users lie without meaning to. Three big forces are at work:

1. Social Desirability Bias

Users want to be nice. You show excitement about a feature? They’ll match it. They don’t want to hurt your feelings.

2. The Optimism Gap

We all think we’ll be better tomorrow. “I’ll check my metrics daily!” is just like “I’ll hit the gym every morning!”

Users picture their perfect future self. Not their real, busy, tired self.

3. The Visualization Problem

When users think about your feature, they see the best case. They have time. They have energy. They care. Reality isn’t like that.

The Guiding Question Trap

Most UX researchers use questions that lead to fake insights:

• “Would you use this feature?”
• “Is this valuable to you?”
• “How often would you check this dashboard?”

These questions suck because they:

1. Point users to your solution
2. Make “yes” too easy
3. Tell you nothing about real problems
4. Get polite lies, not truth

What Users CAN Tell You (And What They Can’t)

Users CAN tell you:

• What pisses them off right now
• How they solve problems today
• Where they spend their time and money
• What hacks they’ve built

Users CANNOT tell you:

• What features they’ll actually use
• How often they’ll really use your product
• What solution would work best
• How much they’d pay (without context)

Better UX Research: The Detective Framework

Don’t ask what users want. Be a detective solving a case:

1. Watch Behavior, Not Opinions

Instead of: “Would you like better metric tracking?”

Try: “Show me how you made your last business decision.”

Instead of: “Would this dashboard be useful?”

Try: “Walk me through your morning when you first check your business.”

2. Find The Evidence

Good research shows what users do, not what they say:

• What tabs are open in their browser?
• What tools do they pay for?
• What notes are on their desk?
• What spreadsheets have they made?

3. Follow The Pain

Pain beats wishful thinking:

• When do they sigh or get frustrated?
• What wastes the most time?
• When do they say “it’s fine” or “we make it work”?
• What do they bitch about to teammates?

Case Study: The Dashboard Nobody Used

A fintech startup spent 4 months building an “investor dashboard.” Users said they wanted it. They were excited about it.

After launch? Nothing. Nada. Less than 2% logged in more than once.

When we dug deeper with the detective framework, we found:

• Users only cared about one number (their total return)
• They wanted this sent by text, not in a dashboard
• They only cared when the number changed by a lot

So what did we do? We killed the dashboard and built a simple alert system. Engagement jumped to 68%.

The Science of Real User Behavior

Research on how users really behave shows:

• Past behavior predicts future behavior 3x better than what people say they’ll do [Source: Journal of Applied Social Psychology - Attitude-Behavior Relations]
• Open-ended questions get way more insights than yes/no questions [Source: Nielsen Norman Group - Open-Ended vs. Closed Questions]

Three Questions That Predict Real User Behavior

Want to know if users will actually use your feature? Ask these:

1. “What are you doing right now to solve this problem?”
   • If they say “nothing,” they don’t care enough
2. “Show me the last time this happened. What did you do?”
   • Look for real problem-solving
3. “What happens if you can’t solve this problem?”
   • If the answer isn’t painful, they won’t use your solution

How to Start Using This Today

Here’s your action plan:

1. Fix your interview script
   • Cut all yes/no questions
   • Remove any mention of your solution
   • Or copy the script I use (Notion Doc)
2. Watch first, ask later
   • Start by seeing how users work now
   • Note where they get pissed, create hacks, or waste time
3. Test for commitment, not interest
   • “Would you switch from your current tool today?”
   • “If this cost $X, would you pay for it right now?”

Final Thoughts

UX research isn’t about validating your ideas. It’s about finding problems worth solving.

Remember: What users say ≠ What users do

Next time a user loves your feature idea, be skeptical. They aren’t lying on purpose. They really think they’ll use it. But humans suck at predicting what we’ll actually do.