Why You Only See What You Want to See

Teresa Torres argues that product teams often fall for confirmation bias: people notice evidence that supports their ideas and overlook disconfirming signals, even when given the same data. Her example shows how customer interviews and user tests can still mislead if the person running them is attached to an idea. Getting out of the building and talking to customers is not enough on its own. Torres draws on the Heath brothers' book Decisive to suggest concrete countermeasures, which matter because unchecked bias leads to building features nobody uses.

01Key takeaways

• Expect your brain to notice confirming evidence more readily than disconfirming evidence, especially for ideas you love.
• Temper your enthusiasm when seeking feedback so colleagues feel free to disagree.
• Ask people why an idea won't work and explore their ambivalence rather than dismissing it.
• Deliberately consider doing nothing or going the opposite direction before committing.
• Apply the same skepticism to customer interviews, listening for what isn't said as well as what is.

02Key sections

The bias in action

A product lead excitedly pitches an idea, sees lukewarm reactions, and then cherry-picks customer feedback that supports the idea. The feature ships and flops.

Why we only notice what we want

Confirmation bias operates automatically, making us roughly twice as likely to notice supporting evidence. Attachment to an idea amplifies it, as shown when two people watch the same user test and each walks away convinced their design won.

Getting out of the building is not enough

Customer contact is necessary but insufficient, because the interpretation of what customers say is itself biased. Teams must actively counter the bias.

Three mitigation strategies

Make disagreement easy by asking for neutral feedback, ask questions that surface contrary information, and deliberately consider the opposite option before moving on.

Applying this with customers

The same techniques apply in customer conversations, where hidden signals like a competitor contract or unresolved problems can be missed if you stop listening at 'that might work.'

03From the post

“This post is part of a series about making better product decisions. You arrive at the office one morning, excited to tell your team about a new idea you had the night before. They look at you nonchalantly. One of them says, "that might work." You were expecting more. You”

04Frameworks mentioned

Summary and takeaways written by PM Atlas; quotes are short excerpts. © the original author.