Dark Patterns: Interfaces Designed So You'll Get It Wrong

Dark patterns are the reason you have paid for something you did not want, kept a subscription you meant to cancel, and tapped “Accept all” on a consent banner you never read. I want to be careful about the accusation, because most bad interfaces are not sinister — they are just badly made, built in a hurry by tired people under a deadline. But there is a particular kind of confusion that deserves a harder look: the interface that is muddled in exactly one direction, the direction that makes the company money, and perfectly lucid everywhere else. That is not incompetence. That is a design decision, taken deliberately, by someone who was paid to take it.

The test that separates sloppiness from strategy

Here is the check I use, and it costs nothing to apply. When an interface confuses you, ask which way the confusion runs. Is the app equally clumsy at taking your money and at giving it back? Is the sign-up as slow and multi-screened as the cancellation? Is the button that costs you money as pale and small as the one that saves you money?

If the friction is spread evenly, you are looking at a team that has not yet got good at its job. If the friction sits only on the path that costs the company revenue — while the revenue-generating path glides along in one thumb-tap — you are looking at a choice. Nobody accidentally builds a checkout that works flawlessly and a cancellation flow that requires a phone call during office hours. Somebody drew that. Somebody approved it. Somebody watched the numbers go up afterwards and called it a win.

This asymmetry is the whole thing. It is what turns a design question into an ethical one, and it is the reason I think the subject belongs in any honest account of how technology gets captured — because capture rarely announces itself. It shows up as an extra checkbox, a greyed-out word, a countdown timer that resets when you reload the page.

Nobody accidentally builds a checkout that works flawlessly and a cancellation flow that requires a phone call during office hours.

Where the name came from

The term belongs to Harry Brignull, a British user-experience researcher who began cataloguing these tricks around 2010 and gave them a name that stuck. He built a public archive of examples so that designers could recognise the moves and consumers could point at them. Over the past few years he has pushed for a plainer label — deceptive design, or deceptive patterns — partly because the older phrase can sound like a dark art rather than what it usually is: ordinary commercial deception, rendered in pixels.

I have some sympathy with the change. “Dark pattern” has a certain glamour to it, and glamour is the last thing this deserves. A pre-ticked donation box is not sorcery. It is a man in a shop slipping an extra item into your bag while you are counting out change. The plainer we describe it, the harder it is to defend. But the older name is the one most people know, so both stay useful: one gets you found, the other gets you understood.

A field guide to the moves

Once you know the shapes, you start seeing them everywhere, which is both useful and slightly exhausting. The common ones:

• Drip pricing, or hidden costs. A price is advertised, and then convenience fees, handling charges, platform fees and insurance appear one screen at a time, so the final figure only exists at the last step — when you have already invested ten minutes and do not want to start again somewhere else.
• The roach motel. Easy to check in, hard to check out. Subscribing takes one tap; unsubscribing takes a hunt through settings, a retention offer, a survey, and sometimes a phone call. The design promises an exit and then hides the door.
• Confirmshaming. The decline button is written to make you feel small. “No thanks, I don’t care about saving money.” “I’ll stay unprotected.” The copy is not information; it is a nudge delivered as mild humiliation.
• Pre-ticked consent. A box you never touched, already ticked, agreeing to marketing, to data sharing, to a small donation, to insurance. Your silence is recorded as a yes.
• Basket sneaking. Items you did not select arriving in your cart — a warranty, a tip, a service add-on — so that the default is purchase and the effort is refusal.
• Disguised ads. Promotions styled as search results, recommendations, editorial articles or user reviews, so that the interface’s own voice is quietly for sale.
• False urgency and fake scarcity. “Only 2 rooms left at this price.” “Offer ends in 04:59.” Sometimes true. Often a counter that resets, a stock figure that is not a stock figure, a deadline that renews forever. Panic is a very reliable way to stop someone comparing prices.
• Nagging. The same request, repeated until refusing costs more energy than agreeing. Notifications, pop-ups, “rate us” prompts, reminders to install the app you deliberately did not install.
• Trick wording. Double negatives and deliberately vague labels, so that the fast, instinctive answer is the one the company prefers. “Untick to not opt out of receiving offers.”
• Bait and switch. You are drawn in by one outcome and delivered another — the advertised item is suddenly unavailable at the final step, and a costlier substitute is waiting.
• Consent dialogues. The purest example of the asymmetry test. “Accept all” is one large tap. Refusing means “Manage preferences”, then a list of several hundred “legitimate interest” partners, each with its own toggle, across several screens. Both routes end in a legally valid consent record. Only one of them respects your time.

Every item on that list works on the same fuel: our attention is finite and our patience is cheaper than our money. The interface is designed around the knowledge that you will not read, will not scroll, will not toggle two hundred switches at 11pm. That is the same economic logic that runs everything else on your phone, and I have written separately about the attention economy and about the real cost of free apps — deceptive design is simply that logic arriving at the moment of payment.

The regulatory turn

For a long time the only defence was individual vigilance, which is a polite way of saying no defence at all. That has begun to change, and India is one of the places where it has changed most explicitly.

In November 2023, the Central Consumer Protection Authority — the regulator set up under the Consumer Protection Act, 2019 — notified the Guidelines for Prevention and Regulation of Dark Patterns. The guidelines do something unusually direct: they define dark patterns as design practices that mislead or trick users into an action they did not intend, hold that such practices subvert consumer autonomy and amount to unfair trade practice or misleading advertisement, and then list thirteen named patterns that are prohibited outright. The list reads almost exactly like the field guide above: false urgency, basket sneaking, confirm shaming, forced action, subscription trap, interface interference, bait and switch, drip pricing, disguised advertisement, nagging, trick question, SaaS billing, and rogue malware. They apply to platforms offering goods or services in India, including foreign ones, as well as to advertisers and sellers.

Since then the regulator has pushed further — asking e-commerce platforms to audit themselves for these patterns and declare that they are clean, and, on the public record in 2026, issuing notices and monetary penalties to specific platforms found using them. I would hedge on how much of this has bitten yet. Independent studies through 2026 have kept finding the same patterns live on major apps, self-declaration is a weak instrument, and enforcement on a handful of companies is not the same as a changed industry. But the principle is now written down, and that matters more than any single order.

India is not alone. The European Union’s consent rules and platform regulation have been used against manipulative interface design, and the United States’ consumer regulator has brought actions over hidden charges and hard-to-cancel subscriptions. Different laws, same finding: an interface engineered to produce a decision the user would not otherwise make is a commercial practice, and commercial practices can be unlawful.

This is not clever marketing. It is unfair trade practice with a design brief attached.

Why naming it is the whole defence

I do not think most of us are stupid about these things. I think we are busy. Deceptive design does not defeat your intelligence — it defeats your bandwidth. It takes a decision you would get right with thirty seconds of attention and puts it somewhere you only have three.

Which is why naming the pattern is so unreasonably effective. The moment you can say “that is a countdown timer that resets” or “that is a roach motel” or “that is confirmshaming”, the trick stops working on you personally, because you are no longer reacting to the interface — you are evaluating it. Recognition creates the half-second of distance the design was built to deny you.

It works collectively too. A regulator cannot prohibit “sneaky websites”; it can prohibit drip pricing and pre-ticked consent, because those have definitions, and definitions can be enforced. A company can deny being manipulative forever; it has a much harder time denying the specific pattern sitting on its own checkout page, in a screenshot, next to the clause it breaches.

So the practical advice is dull and effective. Slow down at the last screen, where the hidden costs live. Read the small pale text, because it is pale for a reason. Before subscribing to anything, find the cancellation route first — if you cannot find it in a minute, that is your answer. Treat a countdown as evidence of nothing. And when you catch one, say so out loud: complain, report it to the consumer authority, tell the people around you. Individual care is a floor, not a solution — the fuller argument for what shifts this at the level of rules and markets is in what you can actually do.

The people who build these interfaces are not villains in a story. They are employees with targets, working in organisations where a two per cent lift in conversions is a promotion and nobody is measured on whether users understood what they agreed to. That is precisely why this cannot be left to individual conscience or individual vigilance. The answer is to make the pattern nameable, make it illegal, and make it expensive — so that the cheapest interface to build is also the honest one.