On this page
These five principles come from different places. Two are findings from experimental psychology, one is a principle of perception from the 1920s, one is a memory effect, and one is a heuristic from a usability consultant. They are often presented together as "laws", which makes them sound more certain than they are.
Each section below gives the principle, a weak and an improved implementation, the practical lesson, and the limits. The limits matter as much as the lesson: they are where a principle stops being evidence and starts being an excuse.
01Fitts's law: make important targets easy to reach
In 1954, Paul Fitts showed that the time needed to point at a target depends on two things: how far away it is and how big it is. Far and small is slow. Close and large is fast. The relationship is logarithmic, so doubling the size of a tiny target helps a lot, while doubling an already large one helps very little. A common form of the model, from Scott MacKenzie's work on human-computer interaction:
time = a + b × log2(distance / width + 1)
Long travel to a tiny target that also sits where nobody finishes a form.
The action sits where the task ends, and its hit area is generous.
On touch screens, size matters even more, because a fingertip is less precise than a cursor. WCAG 2.2 sets a minimum target size of 24 by 24 CSS pixels at level AA, with exceptions, and 44 by 44 at level AAA. Apple's guidelines recommend 44 by 44 points and Material Design 48 by 48 dp. The hit area can be larger than the visible element: a 32px icon button can have a 44px target.
Place frequent actions where the task ends and give them generous hit areas. On desktop, screen edges and corners are easy targets because the pointer stops there.
Fitts models pointing time, not decision time or errors. Making a destructive action big and close makes it easy to hit by mistake, so sometimes friction is the goal. And every large target takes space from something else.
02The Hick-Hyman law: choice complexity affects decision time
William Hick (1952) and Ray Hyman (1953) found that the time to choose between alternatives grows with the logarithm of the number of alternatives, when each one is equally likely and the person knows them well. More options, slower choices, but each extra option adds less time than the one before.
time = a + b × log2(n + 1)
The conditions matter. The law describes choosing among known options. When options are unfamiliar and people must read each one, the time grows roughly in line with the number of items instead. In interfaces, both cases occur, and both point the same way: at a decision point, fewer and clearer options help.
No default, no grouping. Most users want PDF or Excel and still have to scan all twelve.
Three choices with a reason each, a sensible default, and the rest one click away.
At the moment of decision, show the few options most people need, describe them in terms of purpose, and offer a default. Keep the rest available through progressive disclosure.
Splitting one choice into several steps can cost more time than it saves. Experts who know exactly what they want may prefer one long, well-ordered list. Hiding options also hides them from people who would have needed them.
03Jakob's law: respect existing mental models
Jakob Nielsen put it this way in 2000: users spend most of their time on other sites, so they expect yours to work like the ones they already know. It is a heuristic from usability practice, not an experimental law, but it describes something real. Every convention people have learned elsewhere is knowledge you get for free, as long as you do not contradict it.
Distinctive, and every user has to learn it from scratch.
Nothing to learn. Attention goes to the content instead of the controls.
"Other products" includes more than consumer apps. For professional software, the reference products are the tools people use all day: spreadsheets, email, the previous version of your own product. Tax professionals who live in spreadsheets expect keyboard navigation, copy and paste, and sortable columns. Meeting those expectations is often worth more than a fresh visual idea.
Use established patterns for navigation, search, forms and common controls. Spend your originality on the problem only your product solves.
Conventions differ between platforms and audiences, and they change over time. A new pattern is justified when it is clearly better and easy to learn. Copying a competitor also copies its mistakes.
04The law of proximity: spacing communicates relationships
The proximity principle comes from Gestalt psychology, described by Max Wertheimer in 1923: elements that are close together are seen as a group. It is one of the fastest signals in perception. People read groups from spacing before they read any label, which makes it both powerful and easy to get wrong.
Each label sits as close to the field above as to the field below. Which field does "Last name" belong to?
6px between a label and its field, 20px between fields. Help text sits with the field it explains.
Make the space inside a group clearly smaller than the space between groups. Check that every label, hint and error message is closest to the thing it describes.
Proximity can be overridden by stronger cues, such as a shared background or a connecting line. In dense data tables, alignment and rules often do more work than spacing. Watch for accidental grouping too: a "Delete" button placed right next to "Save" reads as a pair of equals.
05The Von Restorff effect: use distinctiveness for what matters
In 1933, Hedwig von Restorff showed that when one item in a list differs from the rest, people are more likely to remember it. Strictly, this is an effect on memory. Designers usually apply it to attention too: an item that differs in colour, size or shape is noticed first. That extension is reasonable, as research on visual search shows that a unique feature can "pop out", but the two claims are not the same.
Five badges in four colours. The one that needs action has to compete.
Other states move to quiet text. The overdue file is the only distinct row, and its label says why.
Pick one element per area that deserves to stand out: the recommended option, the primary action, the item that needs attention. Make everything around it calmer.
If everything is distinct, nothing is. Elements that look like ads are often ignored, a pattern known as banner blindness. Colour alone fails users with colour vision deficiencies (WCAG 1.4.1), and motion used to attract attention can distract or cause discomfort, so respect reduced-motion settings.
06Using these principles in a critique
Principles are good at explaining why a design might work and at generating options. They are weak at proving that it does. A few habits keep them honest:
- Name the condition, not just the law. "This target is small and far from where the task ends" is a design argument. "Fitts's law" on its own is a citation.
- Expect principles to conflict. Hick says fewer choices; Jakob may say the full menu is what users expect. The task decides which wins.
- Check with users. A five-second test, a first-click test or a short usability session will tell you whether the effect you expect shows up.
- Treat the numbers as models. The formulas describe lab tasks under controlled conditions. They predict direction well and exact timings poorly.
- Are frequent actions large enough and close to where the task ends?
- Does each decision point show only the options most people need, with a default?
- Do navigation, search and forms follow the conventions of the platform and the domain?
- Is every label, hint and error closest to the element it describes?
- Is there exactly one distinct element per area, and is it the one that matters?