To truly understand customers, go beyond what they say (unreliable) to observe what they do and why. Triangulate across four levels: what they say, think/feel, do, and why they do it. Observe real workflows, notice subtle cues (hesitations, mouse movements), diagnose rather than validate assumptions, and build genuine user relationships to uncover root causes
7 things you may not know about web accessibility
Automated tools catch only 30-40% of issues — human testing is essential. "Fully accessible" is a myth because user needs often conflict (e.g., dyslexic vs. autistic users). Everyone is situationally disabled sometimes, and accessible content benefits all users. Be skeptical of absolute answers — accessibility requires context and empathy
From UXR to Coaching
A UX researcher discovered that core skills like active listening, non-leading questions, and behavioral observation are shared by both UXR and coaching. Her key realization: people are often blocked not by bad design but by deeper human issues. Coaching simply shifts the focus from improving a product to helping the person directly
The Problem With Watching Session Recordings Manually
Manually watching session recordings doesn’t scale — teams collect more data than they can analyze, creating "analysis debt." Raw recordings provide evidence, not insight, and manual review is slow and inconsistent. AI can detect friction patterns (hesitation, dead clicks) and prioritize meaningful sessions, letting humans focus on interpretation instead of watching hours of video
🎥 NNG: Convenience Sampling - Avoid Bias When Recruiting Study Participants
The participants you recruit for your study matter. Convenience sampling is fast and common in UX research. Learn how to do it effectively and avoid bias in your studies
AI: The waiting problem in AI products
AI products ignore a known HCI principle from 1982: the Doherty Threshold (responses under 400ms keep users in flow). Most AI chats take seconds, agents take minutes, yet provide almost no feedback — just a spinner. Users cope by switching tabs or refreshing. Long operations need progress indicators, time estimates, OS notifications, and logs — all existing conventions. The waiting problem is a design problem, not a technology problem
@uxdigest
