Bad UX is anything in a product that makes people work harder than they should to do what they came for. They get lost, make mistakes the design invited, wait without knowing why, or give up. It is not the same as ugly: a product can look polished and still work badly.
Below are ten bad UX examples that turn up in almost every UX audit, each with the UX law that explains why it hurts and how to fix it. Most of them break one of Nielsen's 10 usability heuristics, so a heuristic evaluation is the quickest way to find them in your own product. Fixing one, and writing up why, also makes a strong UX case study. AI features add mistakes of their own, covered in AI product design, and many of the layout ones come down to the Gestalt principles.
A cookie banner, an offer to install the app and a newsletter sign-up, all before the page has loaded. The person came to read or buy something, and their first job is now to close three boxes. Many leave instead.
A cookie banner, an app prompt and a discount pop-up stacked on arrival: three decisions before the person has seen what they came for.
Why it hurts: every interruption costs attention, and it costs most when it arrives before people have any reason to care. That is Interruption Cost. The fix: show the content first, ask for one thing at a time, and ask after the person has done something that makes the request relevant, such as finishing an article.
A top navigation that lists every section the company has, in the company's own words. People scan it, do not see the word they are looking for, and reach for search or the back button.
A navigation bar with every section at the same level, so nothing stands out and people have to read every item to find theirs.
Why it hurts: more options take longer to choose from (Hick's Law), and labels that follow the org chart do not match how people think about their task (Information Architecture). The fix: group pages around what people come to do, keep the top level short, and name items in the words people use. Hiding everything behind a menu icon is not the fix either: Nielsen Norman Group found that hidden navigation is used less and found less, especially on desktop.
Icons and links squeezed close together on a phone, so people hit the wrong one, or tap three times before one lands. It is worst for anyone with large fingers, a shaky hand or a phone held in one hand on a bus.
Action icons only 24 pixels wide with 2 pixels between them: easy to miss, and easy to hit the neighbour instead.
Why it hurts: small, close targets are slow and error-prone to hit, which is Fitts's Law. The fix: make every tappable area at least 44 by 44 pixels, even if the icon inside is smaller, and leave space between neighbours. Minimum Tap Target has the numbers.
Pale grey text looks calm and modern in a design file on a bright monitor. On a phone in sunlight, or for anyone with weaker eyesight, it is hard or impossible to read.
Sample text at four contrast ratios: the pale grey and the yellow fail, the dark grey and the blue pass. The pale grey looks fine in a design tool and fails real people.
Why it hurts: text people cannot read is text that does not exist. The fix: give body text a contrast ratio of at least 4.5 to 1 against its background, the WCAG AA level, and check it with a contrast tool rather than by eye. See Accessible Contrast.
A phone field that only accepts one format, a postcode field that fails if you add a space, a card number that must be typed without gaps. The input is correct, but the form refuses it and makes the person guess what it wants.
Why it hurts: the system is making the person do work it could do itself. The fix: accept the formats people actually use, tidy the value behind the scenes and show it back in one clean format. That is Postel's Law, and its page shows the strict field next to the forgiving one.
"Invalid input." "Something went wrong." "Error 400." The message tells people they failed but not what happened or what to do next, and sometimes it clears the whole form as well.
Why it hurts: people cannot fix what they cannot understand. The fix: say what went wrong in plain words, say how to fix it, put the message next to the field it is about, and keep everything the person already typed. Errors, empty screens and loading are all states that need designing, which is Design Every State.
A button that does nothing visible when pressed, so people press it again and pay twice. Or the opposite: a spinner that flashes for a fraction of a second on every small action, so the screen flickers.
A saving spinner that appears for about a tenth of a second: too short to read, long enough to make the screen flicker.
Why it hurts: people need to know their action registered. The Doherty Threshold puts the point where a response stops feeling instant at around 400 milliseconds. The fix: respond to every press straight away, even if only by changing the button, skip the spinner for actions that finish quickly, and show clear progress for ones that do not.
Text sizes picked one screen at a time: 15, 17, 18, 21 and 22 pixels, all slightly different and none clearly more important. People cannot tell a heading from a label, so they cannot skim.
Text sizes chosen screen by screen blur together; a small set of sizes from one scale makes the hierarchy obvious at a glance.
Why it hurts: hierarchy only works when the steps between sizes are big enough to notice. The fix: choose a small set of sizes from one ratio and use only those. Modular Type Scale shows how.
The cart moved to the bottom-left behind a custom icon. The logo that does not link home. A search field disguised as a button. Each looks like a fresh idea in a design review and costs every visitor a search.
Why it hurts: people spend most of their time in other products and expect yours to work the same way. That is Jakob's Law, whose page shows the shop with the cart in both places. The fix: keep common things where people expect them, and be original where it makes your product better.
Some bad UX is deliberate. A subscription that takes one tap to start and a phone call to cancel. A pre-ticked box that adds insurance to the basket. A decline button that reads "No thanks, I don't like saving money." Designers call these dark patterns or deceptive patterns.
Why it hurts: they may lift a number this quarter, but they cost trust, bring complaints and refunds, and in many places they attract regulators. The fix: make leaving as easy as joining, never pre-select something people pay for, and write choices that respect both answers.
Done together and written up, those steps are a UX audit.
Spotting these mistakes quickly is a skill, and it grows with practice. Every challenge in UX Quest puts a real screen in front of you, asks what to change, and explains the UX law behind the answer. Start with Interruption Cost, Minimum Tap Target and Design Every State. Each law page says whether its challenge is free or part of Pro.
Three laws behind the mistakes above. Open one to see the screen and the challenge.
Bad UX is anything in a product that makes people work harder than they should to reach their goal: confusing navigation, forms that reject valid input, missing feedback, unreadable text or tricks that work against them. It is about how a product works, not only how it looks.
Pop-ups that pile up on arrival, a menu with every page in it, buttons too small to tap, light grey text, forms that reject valid input, vague error messages, missing feedback, inconsistent text sizes, moved conventions and dark patterns.
Designing for how the company is organised instead of what people want to do, ignoring states such as errors and empty screens, small tap targets, low contrast, and breaking patterns people already know.
Not always. A pop-up that interrupts people before they have seen the page usually is. One that appears at a relevant moment, asks for one thing and is easy to close can be fine.
They hide navigation, and Nielsen Norman Group found that hidden navigation is used less and found less than visible navigation, especially on desktop. On a phone a menu icon is often a reasonable trade-off, as long as the most important items stay visible.
Find it first: walk through the product, check it against usability heuristics, look at where analytics show drop-offs and watch a few users. Then fix the problems that block the main task first.