011 min
What breaks without consistency and standards
Picture a team that builds a project-tracking tool. On one screen a team member clicks Remove, on another Delete, and on a third a trash icon. All three do the same thing: they take a task off the board. The team member has no way to know that. They stop and ask whether Remove only hides the task, and whether Delete destroys it for everyone.
The same pause happens again on other screens. In one table a plus sign opens a task's details, and in another place a plus sign adds a new task. Each choice is small. Together they make a tool that feels unpredictable, so people check before they act, and the checking takes attention away from their work.
Designers call this failure inconsistency. The tool does not follow its own rules, and it does not follow the habits the person learned elsewhere. Where did the idea of a rule against it come from?
021 min
Where consistency and standards comes from
Designers wrote about the problem long before it had a heuristic's name. Ben Shneiderman put "strive for consistency" first among his rules for dialogue design in his 1987 book Designing the User Interface, and called it the rule designers break most often.
In 1990 Rolf Molich and Jakob Nielsen published nine heuristics, and the fourth was named Be Consistent. Their description added two ideas. One system action should, when it makes sense, always be done with one user action. And consistency also means coordination between separate systems that the same people use.
In 1994 Nielsen tested the idea against evidence. He rated 101 heuristics, taken from several published sets, against 249 usability problems collected from earlier projects. He then ran a factor analysis, a statistical method that groups ratings that rise and fall together. Consistency and standards came out as a factor of its own, and it kept that place in the set Nielsen published afterwards.
Nielsen Norman Group, the consultancy Nielsen co-founded, publishes the current wording:
Users should not have to wonder whether different words, situations, or actions mean the same thing. Follow platform and industry conventions.
The first half is about sameness inside a product. The second half is about sameness with other products. How do the two halves work together?
032 min
How consistency and standards works
Consistency and standards has two halves, and each one asks for a different kind of sameness.
Internal consistency
Internal consistency is sameness inside one product, or inside a family of products from one company. The same word names the same action, the same icon does the same job, and the main button stays in the same place. In the project tool, it means one name for taking a task off the board, used on every screen.
External consistency
External consistency is sameness with products the person has used before. A convention is a habit that many products share: a shopping-cart icon for items you plan to buy, a magnifying glass for search, a logo that returns to the home page. Nielsen Norman Group ties this half to Jakob's Law, which says people spend most of their time on other sites and expect yours to work in the same way. Conventions are not laws. They work because most people have already met them.
Why knowledge transfers
People learn a product by building a mental model, a working picture of how it behaves. A product that matches the picture asks for little attention. A product that does not match forces the person to rebuild the picture first. The attention spent on the rebuild is cognitive load: mental effort that does not go toward the task.
Evidence from a 114-person experiment
John Satzinger and Lorne Olfman tested this in 1998. They asked 114 people to learn two prototype teamwork programs in a laboratory. For some people the two programs used different action language syntax, which means the commands were worded and ordered in different ways. For others the two programs matched. Satzinger and Olfman found that people needed more actions to finish tasks when the two programs asked for actions in different ways.
A surprise: a different look helped accuracy
The study found something different for appearance. In the same study, people whose two programs looked different were more accurate than people whose programs looked alike. The authors suggest that a distinct look helps a learner keep two programs apart in their head. This is one laboratory study with prototypes, so it does not prove that matching looks are harmful. It does show that sameness should be chosen for a reason.
Consistency matters most when the task is hard
Earlier studies of consistency gave mixed results. In a 2010 thesis at Clemson University, Jeremy Mendel suspected that task difficulty explained the mix. In his experiment, 48 people searched for answers to 24 questions across four websites. Mendel found that the effect of consistency depended on how hard the task was. A simple task can make an inconsistent design look harmless, and a hard task shows the cost.
042 min
Consistency and standards in real products
Real products show both halves of consistency and standards, and they show what each half looks like when it fails.
Where consistency holds
Kate Kaplan of Nielsen Norman Group described Microsoft Power BI in 2021. Power BI keeps the plus sign for one job, adding something new. The main button for creating an item sits in the upper-right corner of each workspace and is drawn in the same way. A person who learns it once in one workspace can find it in the next.
Microsoft Office works in a similar way across products. Word, Excel and PowerPoint all use a ribbon, a toolbar of tabs along the top, and a person formats text in the same way in all three.
Where it breaks: the plus sign
Kaplan also reported on a project-management app made for DevOps teams, the software teams that build and run their own systems. Its plus sign added new items in one place and expanded a table row in another. A daily user asked whether they could add a task there, then read the hover text and learned the sign only expanded the row. Kaplan drew a clear conclusion: people who use a product every day are confused by inconsistency too, not only newcomers.
Where it breaks: two environments in one product
In 2012 Jakob Nielsen reported a test of Windows 8 with 12 experienced PC users, on regular computers and on Surface RT tablets. Windows 8 put two environments on one device: a Start screen made for tablets and a desktop made for PCs. Nielsen found that the two environments worked differently, so the experience felt inconsistent, and people had to remember which environment held which feature.
Where it breaks: a button that moves
The check-in form at Sutter Health, a health-care provider, ran over several screens. A Finish Later button sat to the right of the main action on the first two screens and moved to the left on the third. That change can lead people to press it by accident and then find they are not checked in for their appointment.
Each case repeats the mistake the team made: a sign or a position that means different things in different places. These examples show inconsistency causing harm. Is more consistency always better?
053 min
When consistency and standards hurt users
More consistency is not always better, and one of the best-known objections is older than the heuristic. In 1989 Jonathan Grudin published The Case Against User Interface Consistency in Communications of the ACM. He did not argue for random design. He argued that consistency is a poor guide unless it serves the user's work.
Three meanings of one word
Grudin said people use the word in three senses. A design can match itself. It can match other interfaces the person knows. And it can match things in the world outside computers, such as arrow keys laid out like the points of a compass. He called this third kind analogic consistency. The three senses can disagree, so an instruction to make a design consistent does not say which one to follow.
Nobody agrees what it means
Grudin noted that a researcher wrote in 1981 that it was not clear what consistency means. In 1988 a two-day workshop of 15 experts could not produce a definition. That does not make the heuristic useless. It means a team needs to say which kind of consistency it wants, for which screens.
The cursor-key example
Grudin asked 24 interface professionals which cursor-key layout novices would rate highest, and all of them chose the star. The star matches the points of a compass. But studies Grudin cites found the inverted T more usable. In an inverted T, one finger can cover both the up key and the down key, while two other fingers cover left and right. The layout that looks neat was not the layout that worked best for a hand.
Learning and use can disagree
Most experiments on consistency measured how fast people learn. Grudin pointed out that people buy software to use it, not to learn it. If a consistent design helps learning but slows people who already know the product, and most of the users are those people, consistency works against the design.
Place things by task, not by type
Grudin used a kitchen to make the point. Table knives and steak knives share a drawer, carving knives stay in a block, silver knives sit with the good china, and a putty knife is in the garage. That is inconsistent, and it is the right arrangement, because each knife is stored where it is used. Interface controls, he said, also belong where the task needs them.
When breaking a convention is justified
Nielsen Norman Group gives a test for departing from a convention: do it only when the task needs it or when it clearly improves efficiency. Its example is the save icon. The usual symbol is a floppy disk, but many people have never held one, so a different symbol may be the better choice. Consistency can also be taken too far. Nielsen Norman Group points to an update of Google's Mail, Calendar and Drive icons that gave them the same colors. At the small sizes common on phones, they are hard to tell apart.
062 min
Consistency and standards in accessibility rules
The Web Content Accessibility Guidelines (WCAG), the W3C's technical standard for accessible web content, turn part of this heuristic into testable rules. A success criterion is one rule in that standard, and its level (A, AA or AAA) says how strict it is. Three criteria in WCAG 2.2 concern sameness.
Consistent Navigation (SC 3.2.3, Level AA)
Navigation that repeats on several pages must keep the same relative order each time, unless the user changes it. The W3C explains the reason. People with low vision who use screen magnification see a small part of the screen at a time, so they rely on the position of repeated content to find it again. People who are blind and move through a page in sequence rely on a stable order too.
Consistent Identification (SC 3.2.4, Level AA)
Components with the same function across a set of pages must be identified in the same way. If a search button is labelled Search on one page, a different label on another page breaks the rule. The criterion covers repeated components only. It does not demand identical labels for unrelated components that happen to do similar things.
Consistent Help (SC 3.2.6, Level A)
Help options, such as contact details or a self-help link, must stay in the same order relative to the rest of the page when they repeat across pages. A person who needs help is often already frustrated, and a help link that moves between pages adds to the problem.
072 min
Applying consistency and standards
Back in the project tool, the team can fix its problems with the moves below. They apply to most products.
Pick one name for each action.
List every label for an action in the product, choose one, and replace the others. Remove and Delete become one word.
Give each symbol one job.
If the plus sign adds, it should not also expand. Write the rule where every designer can see it.
Build a design system.
A design system is a shared set of components and rules, a single place that says how things look and behave. Atlassian lists each component of its product family, with examples and usage guidelines, so new features work like old ones.
Read the platform's own guidelines
before inventing a pattern. Apple, Google and Microsoft each publish rules for their platforms, and users expect those rules to hold.
Review designs together.
Nielsen Norman Group advises regular design reviews with people from several teams, plus a consistency review of all designs about once a year, because inconsistency builds up even when a standard exists.
Run a heuristic evaluation.
In this method, reviewers walk through the screens and check them against the heuristics. Inconsistent labels and moving buttons are easy to find this way.
End with the decision a team faces most often: should this screen do something different from the convention? Ask whether the task needs it. If it does, test the change with users. If it does not, follow the convention.
081 min
How to spot missing consistency and standards
You can check for inconsistency without a lab. Look for these:
- Several labels for one action. Search the product's text for each action and count the different words used.
- One icon with two jobs. Collect every use of an icon in a spreadsheet and see whether the jobs match.
- A control that changes position. Compare the same control across the screens of one flow, as in the Finish Later example.
- Hesitation in recordings. In session recordings, look for a pause before an action that has different names on different screens.
- Support questions about sameness. Look for tickets that ask whether two features are the same or how one differs from another.
- Different formats for the same data. Dates, phone numbers or prices written in different ways on different pages.
- A change of voice. Compare the tone of the home page with the tone of error messages. Nielsen Norman Group shows a hotel site whose home page is formal and inviting while its error message headline is robotic and unhelpful.
091 min
Consistency and standards vs. nearby concepts
Several ideas are related to this heuristic. They differ in what each one asks the designer to match.
| Concept | What it asks for | How it differs |
|---|---|---|
| DesignJakob's Law | ||
| DesignRecognition rather than recall | ||
| DesignDesign system | ||
| Not in the library yetUniformity |
?6 questions
Questions people ask
Does consistency and standards mean every screen must look the same?
Can a product be consistent inside and still fail people?
Does the heuristic cover writing, or only controls?
Does the idea apply outside software?
Does a small team need a design system for this?
What is the difference between a convention and a standard?
§12 sources
Sources for consistency and standards
Krause, R. (2021). Maintain Consistency and Adhere to Standards (Usability Heuristic #4). Nielsen Norman Group.
Nielsen, J. (1994, reviewed 2024). 10 Usability Heuristics for User Interface Design. Nielsen Norman Group.
Molich, R. Molich and Nielsen’s heuristics (1990). A summary by the co-author that reproduces the wording of the heuristics from Molich and Nielsen (1990), Improving a human-computer dialogue, Communications of the ACM 33(3).
Nielsen, J. (1994). Enhancing the explanatory power of usability heuristics. CHI ’94 proceedings, pages 152–158 (a copy of the paper).
Show all 12 sourcesShow fewer sources
Grudin, J. (1989). The case against user interface consistency. Communications of the ACM 32(10), 1164–1173 (a copy of the paper).
Satzinger, J. W. and Olfman, L. (1998). User interface consistency across end-user applications: the effects on mental models. Journal of Management Information Systems 14(4), 167–193 (repository record with the abstract).
Mendel, J. (2010). The effect of interface consistency and cognitive load on user performance in an information search task. Master’s thesis, Clemson University (repository record with the abstract).
Kaplan, K. (2021). 10 Usability Heuristics Applied to Complex Applications. Nielsen Norman Group.
Nielsen, J. (2012). Windows 8 — Disappointing Usability for Both Novice and Power Users. Nielsen Norman Group.
W3C. Understanding Success Criterion 3.2.3: Consistent Navigation (WCAG 2.2).
W3C. Understanding Success Criterion 3.2.4: Consistent Identification (WCAG 2.2).
W3C. Understanding Success Criterion 3.2.6: Consistent Help (WCAG 2.2).
