UCD methodsEstablishedDiscovery

Guerrilla Usability Testing Still Works in 2026

Published 4 September 2026 · Source dated 28 August 2026

With so much noise about AI speeding up design, lightweight usability testing remains one of the most reliable ways to catch real problems early. It's an established method, not a new trick, and it still beats guessing.

Amid all the AI-workflow chatter, the fundamentals research method — sitting five real users in front of a prototype and watching where they stumble — hasn't changed. NN/g's ongoing guidance on usability testing reinforces that small, frequent tests (5 users, task-based, think-aloud) catch the majority of usability issues without needing a dedicated research team or big budget. For junior designers this matters because it's a skill you can practise immediately, with no tooling dependency and no need to wait for AI features to mature. It's also a strong portfolio signal: showing you tested with real people, not just shipped a polished mockup. This is established practice — decades of evidence back it, it's just easy to skip when deadlines are tight.

Coach

run one 5-user or even 2-user guerrilla test on your current project this week and write down what broke.

Today's digest

Surf this topic: UCD methods (opens in a new tab)

Source: nngroup.com (opens in a new tab) ·