Convention over creativity? The ups and downs of UX design | by James Harrison

There’s an inherent friction in the world of UX Design. On the one hand, design has always been considered a creative field, and its practitioners routinely celebrate innovation and ingenuity. On the other, good user experiences tend to lean heavily on well-established interaction patterns, and seem to increasingly revolve around sets of standards. The increasing deference to established conventions raises concerns. As AI elbows its way into the space, capable of executing complex tasks based on sets of instructions, perhaps there’s no need for humans in UX Design at all? The existential question looms large: where does creativity fit in?

In his landmark design book Don’t Make Me Think, Steve Krug writes that “conventions are your friends”. He goes on, suggesting that “one of the best ways to make almost anything easier to grasp in a hurry is to follow the existing conventions–the widely used or standardised design patterns”.¹

This advice appears in any UX Design guide worth its salt. The Nielsen/Norman group, one of the leading UX research consortiums in the world, placed “consistency and standards” on their oft-cited ten usability heuristics, suggesting that well-designed interfaces exploit “platform and industry conventions”.

It’s hard to argue that anyone should do otherwise. Leveraging convention in an interface or experience increases user comprehension and speed, and reduces user errors. Humans have very good procedural memory (remembering how to do a specific thing or set of things), and so reproducing a process exactly in similar contexts brings value, and reduces cognitive load. Accomplish these things in your work, and you’re considered a good UX designer– but is it your work that’s good, or is the convention just good?

The suggestion that most of a UX Designer’s work is simply the correct application and combination of standardised patterns can be troubling in a world where AI agents threaten careers (no doubt a thought that’s crossed the mind of cost-conscious CEOs). From another angle, it might cause you to question your agency and creativity as a designer.

Before you hang up your hoodie, peel the stickers off your laptop, and retire, there are some important things to consider. Who created the conventions in the first place? How does any given pattern establish itself? Most importantly, conventions do change. How (and why) does that happen?

First, let’s examine a bastion of convention; the passenger elevator.

If you’d like to know what a space entirely ruled by convention looks like, step on to an elevator (ironically, elevators in England are frequently called “lifts” — a different language convention). I can say with confidence that if you’re in a passenger elevator built in the last ~50ish years, and that if this isn’t your very first time in one, you probably already know how to use it (let’s set aside exceptions until later). Over just a few short rides, you’ve learned the interface. There’s very little to learn, and very little to remember unprompted, save the floor you’re supposed to be going to. The patterns are ubiquitous, and it’s clear to see that is of incredible value for any one elevator’s user base, made up of regular, occasional, and first -time riders. If you had to learn how a specific elevator worked each time you went somewhere new, you’d likely wonder why this type of thing just couldn’t be standardised to make everyone’s life easier. Fortunately for you, they are.

Passenger elevators aren’t the same by coincidence, or by one elevator company copying what another is doing. The similarities are intentional, required. For example — the indicator for what floor you’re on? That’s subject to the Safety Code for Elevators and Escalators, the 2019 version of which lists five separate guidelines dictating its size, format, and presentation.

Elevators are standardised top-to-bottom for safety and accessibility purposes. They’re used by people from all walks of life, so there’s an increased need to ensure they’re highly usable. There is no– “gulf of execution” between stepping into an elevator, and knowing how to work it. Convention shines here. It’s easy to imagine how powerful applying near universal standards to the things we commonly use (restroom door locks anyone?) in different places for the same reasons. Would that work for digital design as well? Should there be a universal ecommerce checkout to eliminate customer confusion and save time? Apple, Google, Paypal and Shopify all have products in this space, which standardise the payment experiences for users in their orbits.

It’s important to remember UX standards aren’t delivered from above on carved tablets. They’re typically the product of years of iteration and change before they become crystallised into anything resembling a standard or best practice.

But what do we really mean when we say “best practice”? Is it the best way of doing something, or is it simply the most familiar way of doing something?

Usability guru Jakob Nielsen, coined his namesake law of internet user experience (Jakob’s Law), stating that “users spend most of their time on other sites”. What Nielsen is driving at here, is that users are likely coming to you with predefined expectations for how something should work, and that setting those expectations is often beyond your control.

This is true for most of us. But not all. There’s a small handful of companies that control this conversation, simply by owning the vast share of users. They’re those “other sites” Nielsen warned you about.

It’s hard to do search without looking at Google, social networking without Meta, e-commerce without Amazon, streaming without Netflix. Apple, with its series of game-changers has defined how we interact with our devices, our music, our charging cords. Conventions are being set by a small group of giants, and if you want users to quickly and easily understand your products, it’s best to adopt them. One proviso is that while some standards have become more or less universal, different industries adopt internal conventions. E-commerce shopping has a very particular and different set of interface conventions than social networking (although the two are becoming increasingly intertwined). Using a successful social networking convention in an ecommerce space might not always translate well, famously demonstrated by Etsy’s brief implementation of an “infinite scroll” pattern on their search listings, that turned out to negatively impact engagement.

Outsized influence aside, these companies are still fallible. Even design standard bearers like Apple don’t always get it right (remember this every time you see someone inadvertently turn on their iPhone flashlight). There are situations when the most established practice isn’t actually the best one, forcing designers to have to weigh the benefits of familiarity against what they feel is an improved experience. Maybe you can design a “better” elevator experience, but with the convention so well established, you’d be doing users a disservice in breaking with it.

Computer scientist Jaron Lanier calls this process “lock-in,” wherein any given technology becomes so widely adopted and familiar, it becomes nearly impossible to change.² Consider the ubiquitous QWERTY keyboard layout. Despite studies that suggest that QWERTY is both slower and less ergonomic for typing than some alternatives (Colemak or Dvorak layouts, for example), the standard persists.

It’s common for a UX designer to have to justify their design decisions via competitive analysis (examining popular patterns employed by industry equivalents), or by reference to a best practice authority, such as The Baymard Institute, Nielsen/Norman Group, the ACM, and others. For better or for worse, this perpetuates convention and lock-in.

Standards, once formed, are notoriously difficult to change, but they still do. How?

There are, in my opinion, a number of processes through which conventions and standards change.

Adding to the conversation:
UX design is a conversation with other designers. Something I’ve designed might end up in your competitive analyses, or maybe referenced as meeting (or failing to meet) a best practice (not my work of course!). Over time designers refine solutions, or adapt them for new circumstances. Good ideas spread, and designs change. This process is gradual, low-risk, and tends towards refining conventions rather than blowing them up. Even the standards for elevators are updated periodically: for example, in 1945, the emergency stop button and telephone were added to the code. You can’t help but think of this process as idyllic. Designers collaborating, conversing through history in some well-intentioned dialectic, all for the betterment of design, nay– humanity. This isn’t always the case. Dark patterns proliferate too. But this gentle evolution isn’t the only way conventions change.

Going rogue:
Quite the opposite to the above, there are instances where the decision gets made to intentionally reconsider, or outright defy existing conventions. But this hinges largely on influence in terms of its success. If I decide to defy best practices for reasons I think are well-founded on my personal portfolio site, it’s likely that a. No one cares and b. Nothing changes. Contrast that with say, Jony Ive’s work during his tenure at Apple, and the sheer scale and visibility of that work empowers change. That also means that people in positions of power can influence the spread of practices and design decisions that aren’t best (here the Tesla Cybertruck comes prominently to mind). Powerful entities can easily control the conversation, and ultimately the convention, and that can have immense effects. Perhaps we’d all be driving electric cars by now.

New spaces, new patterns
If you have your heart set on defining conventions in your design career, then you should seek out places where conventions don’t exist, or haven’t yet crystallised. Companies that make a splash in new spaces tend to define the standards for users. Consider the “card swipe” pattern that Tinder brought to prominence in the world of dating apps, a convention that’s not just been adopted by many apps in the dating space, but by pop culture writ large where “swipe right/left” act as synecdoche for approvals and denials. As a result the UX pattern has outgrown the dating space, and is leveraged widely in apps for a host of applications, including a recent update to the messaging app Slack.

Abstraction
When a given interface convention becomes so well known by such a large audience, it’s not uncommon to see changes occur that abstract its elements to its simplest terms. Humans do this all the time with language. In English, the phrase “what’s up with ____ (insert subject here),” becomes “what’s up,” and maybe even abstracts further to “‘sup,” depending on the formality of the occasion. Most fluent speakers can interpret this with ease. This is often true with affordance in interface design. When designing roads, you could likely remove the text from the red octagonal “STOP sign”, and users could still interpret its meaning effectively. In the digital world, many users have come to understand that discrete groups of items, such as cards or list items, can be clicked or tapped, without the need for additional affordance. Abstraction can only occur over time, when a significant number of users understand the convention so well, that they no longer need queues. In this case you might say that the convention itself doesn’t actually change, simply its expression, although abstracted patterns can be then applied to new situations, without a necessary link to the original, and become their own “simulated” convention.

Technological change
When new technology emerges, existing conventions can help users make the transition. Digital interfaces have long appropriated their analog forbearers to encourage adoption, a technique known as skeuomorphism. Eventually, users become familiar with the new interface, and the skeuomorphic elements can be abandoned. Consider the move from paged to pageless digital documents, or the slow abandonment of the landline headset “telephone” icon, or the floppy disk “save” icons, both which have lost their relevance as technology changes. Technological transitions often provide both the opportunity and the capability to introduce or alter a pattern. Touchscreens opened the door for gestural interface patterns that weren’t effective in older formats. In a similar example, the “dial” — generally the primary pattern for adjusting the volume of physical devices– gets replaced in the digital age with more screen-friendly controls, like the “slider”.

Technological changes are often at the root of changing conventions, even if it may seem otherwise. In the 1920s, when four provinces in Canada switched from left to right hand driving, this likely had more to do with the proliferation of the automobile and the advent of highways, than some sense of national unity.

Was the move to the “flat design” style of the late 2000s an aesthetic choice, or simply a product of the proliferation of small scale graphics on touchscreens? When skeuomorphic elements re-emerged on these devices some years later, it could be argued that this was simply because the screens had become more sophisticated, and could better express the style.

Even elevators are not immune to technology-inflected convention changes, as I discovered one day when arriving at a building I’d never been to. Late as usual, I dashed into an open elevator, only to find that there were no buttons on the doorside wall. Here I’d obeyed a convention, only to find that it had changed. This elevator was in fact part of an elevator bank system, where multiple elevators are controlled centrally to maximise passenger flow and efficiency (an enormous benefit for large buildings). Had I not rushed past it, I might have noticed the console ahead of the elevators where I was supposed to select my floor. Since then, I’ve been witness to others having made the same error, perhaps a lesson in the price of changing conventions, especially those that are very well established.

All this talk about UX’s reliance on convention might have you feeling like you’re trapped in that elevator. Truly, if you want to do work free from conventions and standards, UX Design is not the place for you (maybe try poetry?). But that doesn’t mean there aren’t outlets where you can apply creativity in strategic ways. Architects are constrained by endless building codes, and yet we still see innovators like Frank Gehry or Zaha Hadid create new forms within them.

UX Design is not devoid of creativity, it’s merely constrained to time, space, and place. You have to carefully select moments to deviate and experiment (hint: no one who has lost their credit card wants to engage with surprise and delight). And you can take advantage of nexus points like technological change, or novel interfaces, to play. UX Research is a valuable tool for measuring what’s successful.

In About Face, one of Alan Cooper’s (many) design principles is that “significant change must be significantly better”³, and designers should take this as an important caveat. The tradeoff for breaking convention needs to be worth the friction this creates, both for the number of users it affects, and for the time required to adapt (In testing, it took users about 52 hours to reach equivalent typing speeds with a non-QWERTY keyboard, after which their speeds greatly increased. While that’s not a lot of time, it’s enough over the span of the world’s users for computer companies to resist making the change).

The ideal circumstance for innovation in UX patterns is one where the interface patterns haven’t yet crystallised, or when new patterns can be introduced as auxiliary, and users don’t have to depend on adapting to them to complete their job to be done, or spaces where users aren’t time- and conversion-based, such as an interactive exhibit in a Museum.

While I might be proven wrong eventually, knowing when to apply an existing pattern, which pattern, and whether or not to deviate for a chance at better results is still a calculus that requires a human brain. Generative AI agents are likely to play a role in the near future in terms of rapid iteration and testing of variations in these patterns, to see how they can be optimised, but they can only (now) follow and recompose what exists. Human UX designers can dream up improvements, or borrow them from tangential places, and by applying them cautiously, can contribute to the general evolution of UX conventions, in step with the changing pace of their users. It’s a process that requires a little patience, something that getting stuck in an elevator from time to time can help you work on.

1. Steve Krug, Don’t make me think, revisited: a common sense approach to Web usability (Third ed.) New Riders, 2013. Page 29.

2. Jaron Lanier, You are not a gadget: a manifesto. Knopf Doubleday Publishing Group, 2011.

3. Alan Cooper, Robert Reimann, David Cronin, Christopher Noessel, About Face: The Essentials of Interaction Design (Fourth Ed.). Wiley, 2014. Page 297.

All images used with permission.