Your Customers Aren’t Lying, But They’re Not Telling the Truth Either

The Problem

Every founder gets told the same thing: "Go talk to your customers."

And look, it's decent advice. But here's the catch: people don't always say what they mean. Your customers will tell you what sounds helpful in the moment. They'll suggest features instead of explaining their actual struggles. And honestly? They can't always picture solutions they've never experienced before.

So if you're only relying on what people tell you in conversations, there's a real chance you end up building something nobody actually needs.

The Core Insight

Here's what we’ve learned: talking to customers gives you surface signals. Watching them? That's where the truth is.

What people say and what people actually do are two completely different things.

What to Do Instead

Start by observing behavior. Run usability tests, look at what people are clicking, pay attention to the clunky workarounds they've cobbled together just to get by.

Then validate with experiments. Throw up a landing page, test a fake door feature, build a bare-bones prototype. Whatever helps you measure real interest instead of polite enthusiasm.

Focus on measuring pain, not collecting opinions. The problems that matter are the ones people keep running into, not the things they mention once in passing during an interview.

And mix your methods. Combine what you hear in conversations with what your data shows: analytics, cohorts, user funnels. When different signals point to the same thing, you're onto something real.

Quick Example

Users kept saying: "I need help budgeting my money."

Then you built a budgeting tool and saw 80% abandon it after setting it up once.

What actually worked? A simple alert that fired whenever they were about to overdraft. Turned out they didn't want to budget, they just wanted to avoid embarrassing declines.

Start Here

Pick one tool this week and compare what you learn to your last customer interview:

To watch behavior: Hotjar or Microsoft Clarity (both free) - See where users actually click and get stuck

To run experiments: Typedream or Carrd - Build a landing page in 20 minutes to test real interest

To measure usage: PostHog (free tier) - Track what people actually do in your product

The gap between what the tool shows and what customers told you? That's your blindspot.

The Takeaway

Talk to customers to discover what problems exist. Watch and measure them to figure out if your solution actually works. Real insight lives in that combination.

Next time you're tempted to take what users say at face value, ask yourself: "Have I actually seen them do this?"

If this resonated, share it with a founder friend. More ideas like this drop every week at Idea TBD. :)