This is a great observation. It’s the reason so much of what we create fails. Aligning design impact with business results isn’t a straight shot and requires understanding human behavior.
I love this stuff, so I thought it would be interesting to dive into more research. Here are five areas that Glare needs to address:
1. Humans overtrust intuition in complex systems
Daniel Kahneman and Gary Klein have great research on intuition (a core strength of design).
Daniel Kahneman
- Describes intuition as fast, automatic thinking built from pattern recognition. It feels confident, but it only works well when the environment is consistent, practice is repeated, and feedback is immediate.
- He draws a line between high-validity and low-validity environments. UX, product strategy, and markets are low-validity by default. Patterns shift, feedback arrives late, and signals are noisy.
- In low-validity environments, intuition drifts toward overconfidence unless it is corrected by real feedback.
Gary Klein
- Has shown that strong intuition is not about comparing options. Experienced people act quickly by recognizing patterns and mentally simulating outcomes.
- His Recognition-Primed Decision model explains why good decisions often look instinctive. They are built from repeated exposure to real situations and clear feedback.
- The overlap is simple. Intuition works when it is trained. Without feedback, it feels strong but stays unreliable.
- Here is his book on intuition, The Power of Intuition
Basically, intuition becomes reliable only after repeated exposure to valid feedback. UX metrics supply that missing feedback loop.
2. Ego and motivated reasoning distort judgment
Lots of research on “motivated reasoning” shows people favor information that supports their prior beliefs, status, or identity (aka cognitive bias)
This shows up in business as:
- Executives defending sunk costs
- Teams overvaluing their own ideas
- Stakeholders dismissing early warnings as “subjective”
Organizational psychology has shown that externalized evidence reduces defensive reasoning (need to find sources here). Metrics work not because they are perfect, but because they shift debates away from identity and toward shared reference points.
3. Users are predictably inconsistent
Behavioral science repeatedly shows that users:
- Say one thing and do another
- Rely on heuristics
- Resist change even when alternatives are better
Richard Thaler argues that people are not perfectly rational. We are biased, impulsive, and often favor short-term comfort over long-term benefit.
Because of that, single user observations are noisy and easy to misread. Real understanding comes from patterns across many small signals. That is why aggregated UX metrics beat anecdotes, without replacing qualitative insight.
4. Measurement improves intuition over time
There is strong evidence that measured feedback trains intuition.
Psychologist K. Anders Ericsson studied elite performers across domains like medicine, chess, music, and sports. His core finding was simple.
Expert intuition is built through deliberate practice, which always includes:
- Clear goals
- Repeated exposure
- Immediate, measurable feedback
- Correction over time
Which translates to:
- Novices rely on rules
- Experts rely on intuition
- Experts only become experts through tight feedback loops
UX metrics act as a feedback loop.
Over time, teams make better decisions AND they develop better instincts. I hypothesize that UX metrics reveal where intuition actually works, and show where new patterns can emerge.
Without feedback, people repeat mistakes while becoming more confident. With feedback, intuition becomes calibrated.
5. Systems thinking backs this up
In systems, intuition without measurement creates unstable loops. Small biases compound. Errors hide until they are expensive. Hard to see early in the process.
Measurement introduces a type of damping (which designers potentially dislike). However, it does not remove human judgment… this type of constraint stabilizes it.
So… back to your assertion @ben. Love @Helge’s thoughts!
- Business is adversarial
- Humans are biased
- Users are inconsistent
In my experience, UX metrics succeed not because people are rational on teams, but because they are not.
