21 laws of UX explained: a practical guide for modern digital products

What the Laws of UX are.

The Laws of UX are a practical collection of psychological principles and design heuristics that help explain why some interfaces feel intuitive while others create friction. A commonly referenced collection groups 21 principles across heuristics, Gestalt principles, cognitive biases, and broader design principles.

The 21 principles, grouped.

Heuristics: Aesthetic-Usability Effect, Fitts's Law, Goal-Gradient Effect, Hick's Law, Jakob's Law, Miller's Law, Parkinson's Law. Gestalt principles: Law of Common Region, Proximity, Prägnanz, Similarity, Uniform Connectedness. Cognitive effects: Peak-End Rule, Serial Position Effect, Von Restorff Effect, Zeigarnik Effect. Additional principles: Doherty Threshold, Occam's Razor, Pareto Principle, Postel's Law, Tesler's Law.

Why these principles matter in practice.

Fitts's Law reminds designers that target size and distance affect interaction effort. Hick's Law explains why too many choices can slow decisions. Jakob's Law highlights the value of familiar interaction patterns. Gestalt principles explain how users visually group information. But these principles should not become a checklist that replaces research. They are a starting vocabulary, not a substitute for testing with real users.

Applying them to modern products.

For mobile apps, use Fitts's Law to improve touch targets. For SaaS dashboards, use Gestalt grouping and progressive disclosure to manage density. For AI products, combine familiar interaction patterns with clear feedback about uncertainty and system state. The strongest application is contextual: identify the user problem, choose the relevant principle, test the design, and measure the result. UX laws give teams a shared vocabulary—they turn “I don't like this layout” into a more useful conversation about cognitive load, discoverability, hierarchy, and interaction cost.

What these laws are, and what they are not.

UX laws are named patterns drawn from decades of research into human perception and behavior. They are useful because they are shorthand for findings that have been repeatedly replicated, and useless if treated as checklist items to satisfy mechanically. The distinction that matters is descriptive versus prescriptive. A law describing how people behave is a finding you should design around. A law dressed as a rule about what to build is a misapplication. Jakob's law observes that people spend most of their time on other products, which means familiar patterns work better than novel ones. It does not mean never innovate; it means innovation has to earn the cost it imposes on users. The laws also sit at different levels. Several concern perception and cognition, such as Miller's law and Hick's law, which constrain what is possible. Others concern specific interactions, such as Fitts's law, which applies directly to target size and distance. Applying a perception law to a specific component requires an intermediate step that is often skipped, which is where most misapplication comes from. Used well, they shorten a design argument. Rather than debating whether a seven-item menu is better, you can point at the research and reason about the trade-off. That is the real value, not a guarantee of a good result.

What happens when the laws are ignored.

The value of a law becomes obvious through the failure mode it prevents, and these are the ones teams routinely learn the hard way. Ignoring Hick's law produces the twelve-item dropdown nobody can scan, which reliably sends a portion of users to the search field or out of the product entirely. Ignoring Miller's law produces settings pages with forty ungrouped toggles, where nobody can form an accurate mental model and so nobody changes anything. Ignoring Fitts's law produces small targets placed far from the hand, which is felt as sluggishness even when the underlying response time is fine. On mobile it is the difference between a flow that feels immediate and one that feels sluggish. Ignoring Jakob's law produces the novel interaction nobody asked for. The design may be defensible in a portfolio, but every user pays a learning cost that the feature has to earn back, and for anything routine that cost is rarely repaid. The pattern across all of them is that violations feel like design problems at review time and like usability problems in production, which is why they are easy to introduce and hard to diagnose later. The mitigations are unglamorous: test with real users, keep a written rationale, and treat repeated patterns in support tickets as law violations to investigate.