Menu Search
Jump to the content X X
Smashing Conf San Francisco

We use ad-blockers as well, you know. We gotta keep those servers running though. Did you know that we publish useful books and run friendly conferences — crafted for pros like yourself? E.g. upcoming SmashingConf San Francisco, dedicated to smart front-end techniques and design patterns.

Is The Internet Killing Creativity?

The internet is a wonderful place (mostly). An unprecedented revolution in communication, it continues to empower more people to publish and share their knowledge than any other phenomenon in history. It is a limitless playground of ideas and unbridled creativity.

Or is it?

It All Looks The Same Link

In 2014, Elliot Jay Stocks declared that designers have stopped dreaming1. That we’ve stopped being creative. That every site looks the same. A crazy notion considering the magnitude of tools and resources we have at our disposal. But Elliot’s been right, and he’s not alone either.

In 2015, Noah Stokes chimed in, telling us web design is losing its soul2, no less. According to Noah, RWD patterns have become habitual, we’ve become stuck in them, and we’re struggling to break out and be different. Subsequently, at the height of summer last year, Sergio Nouvel declared web design dead3, triggering a minor uproar among the people who cared.

Web design is dead. In the ground. Finished. “Frameworks and templates have us covered”, Sergio said. Our patterns are mature, and “trying to get creative at this point will probably be pointless or even harmful.” He mentions The Grid4 as a pioneer in automation and artificial intelligence, essentially taking over the role of web designers.

“[The Grid] analyzes your content to detect the best layouts, colors, fonts, and extra imagery for your site.”

– Sergio Nouvel

Take that in for a moment. The Grid – a piece of software – analyzes your content and makes a website out of it. Boom. No designer needed, no fiddling around in SquareSpace, no agency fees. Websites at the push of a button.

Judgement Day is upon us. The robots are coming. The very tools we’ve built to make our jobs easier are turning on us. And why wouldn’t they?

All your web design is belong to us!5
“All your web design is belong to us!” (View large version6)

Think about it. We’ve focused all our efforts on making sure we get the tech right. That we follow the right UI patterns. That our sites perform well. For the best part of our short history, we’ve focused on creating a recipe for making websites. So let me ask you this: what’s the point of web designers if there’s a recipe?

Our future is bleak, people. We’ve laid the foundation for our own extinction. In a few years time we’re out of a job.

Unless we do something about it.

A History Lesson Link

Whatever you call yourself – whether you’re a web designer, a visual designer, a UX guru or a ‘creative’ (*shudder*) – we’re all part of a grand movement, a discipline that spans centuries. We all share a common ancestor: the graphic designer.

Our medium is different, of course. We communicate with pixels and bandwidth, not ink and paper. But we can’t ignore our past. Web design is just an extension of graphic design – they are the same thing. We share the same history. And in that history lies the key to our survival.

For the sake of argument, let’s say modern graphic design was born in the 15th century, with the invention of the printing press. At this early stage, layouts were straightforward, largely consisting of justified columns and drop caps.

This pattern largely persisted, although towards the 20th century, under the influence of art nouveau, designs became more ornamental and flourishes were added across the board.

Both these styles were challenged in the early 20th century by the modernist movement, which advocated asymmetrical layouts, simplicity and more complex grid structures.

A little later, we got mavericks like David Carson, who literally threw out all the rules and transformed the page into a playground for creativity.

An illustrated history of graphic design7
An illustrated history of graphic design. (View large version8)

I hope you can excuse the imprecise nature of this summary of events, the details of which have little to do with my point, which is that graphic design has come of age. The discipline has arrived at a place where rules have been defined, but we also have the maturity to know how – and when – to break them.

Web design, on the other hand, is infantile.

We’ve almost moved past centre alignment as the default and we’ve swapped ornamental skeuomorphism and shiny buttons for what we annoyingly call flat design. (Even more annoyingly, we keep referring to it as a trend. For the record, the absence of embellishment is not a trend, it is simply graphic design without the bullshit. But I’m digressing.)

We’ve started playing with different layouts, asymmetrical grids and making the most of different viewports. We’ve finally got a good handle on typography, we have blend modes and animation, we have Retina screens. We’re ready to experiment.

Right now is easily the most exciting time ever to be a web designer.

So why do all websites look the same?9 Have we simply stopped caring? Are web designers incapable of being creative? No. I don’t think so. But I do think we’re facing formidable obstacles to new ideas. And the first of these obstacles is the C-word.

The C-Word: How Content Changed Design Link

Around 10 years ago, I started my career in a small publishing agency called White Light Media10. I was a magazine designer.

And in the world of magazines, content is taken seriously. Half my colleagues were editors. Not only did this mean that all the content was written before I designed the final layouts, I could also work with the editors to mold headlines and paragraphs to fit.

Our process looked pretty much like this:

Magazine design process: content, design, then content and design together11
Magazine design process: content, design, then content and design together. (View large version12)

We were a team: content and design, together.

On the web, of course, content is an entirely different beast. The invention of the CMS has changed everything. On the one hand, it allows clients around the world to edit their own content and continuously update their sites without external intervention. That is a great thing.

On the other hand, however, the CMS has, in many, many cases, moved content from the start of the process to the end of the process.

Web design process: design, build, then content13
Web design process: design, build, then content. (View large version14)

What happens when we remove ourselves from content?

Well, first of all, editorial design is no longer bespoke. We no longer match meaningful headlines with complimentary images. We no longer tailor layouts to suit a particular feature.

Instead we design systems, empty templates waiting for content to fill them. We break sites down into component parts and create style guides. We worship atomic design15. We focus on the organization of our design assets, not the meaning of them.

We start our processes, not by coming up with ideas and concepts, but by creating element collages16: collections of buttons and paragraphs and feature boxes. We design them so that our clients can sign off on the look and feel. We obsess over style, ignoring the substance in the process. We fall in love with sexy.

Why? Because we no longer know the context of our designs. And when there is no context (read: content), all that’s left is visual effects and polish. We’ve created a new paradigm. And that paradigm is based on the fallacy that graphic design equals the sum of its parts. That as long as we neatly organize, theme and polish each element of a website, we’re doing it right.


Reclaiming Content Link

If we are to raise our collective creative bar, we need to reclaim content as part of our process. And I don’t simply mean sending an email to the clients and asking them to pass on a Word document. I mean integrate it. Make it ours.

We need to do content and design at the same time. Because the basis of graphic design is image plus text. Not just text. Not just images. Both together.

Of course, lots of graphic design is text only – that in itself is a design decision – but in most cases, photography or illustration improves the experience. Likewise, images can speak for themselves, but in most cases text improves the meaning.

Imagine you’re designing a site for a legal firm. They want something that says that they’re different from the rest, that they are honest, and that they have courage where others lack it. At the kickoff meeting, the CEO hands you this picture for the home page:

The handshake of death17
The handshake of death. (View large version18)

Now, if your text is Lorem ipsum, you’re pretty screwed. Using Lorem ipsum is the same as not having text at all. And if you don’t have text, you don’t really have graphic design.

This is where we should unleash our inner copywriter. You may think you can’t write, or “I’m dyslexic.” That’s fine. We’re not writing an essay here; we’re just using words to sketch out an idea. And we all have ideas. By writing an idea into the picture, we can change the context:

Copywriting adds meaning to mundane images19
Copywriting adds meaning to mundane images. (View large version20)

Immediately, the meaning has changed. You’ve just made the user think. (Take that, Steve Krug21!) The most boring image on earth is suddenly intriguing.

Conversely, changing images has a profound effect on the same piece of text. So if your starting point is “synergy” (the text equivalent of the handshake picture), you can enhance – or change – the meaning of it by choosing the right image. Of course, the bigger the disparity between the picture and the text, the more memorable it gets.

Synergy: the interaction or cooperation of two or more organizations, substances, or other agents to produce a combined effect greater than the sum of their separate effects22
Synergy: the interaction or cooperation of two or more organizations, substances or other agents to produce a combined effect greater than the sum of their separate effects. (View large version23)

Let’s look at some real-life examples. The first example is the work and passion of my colleague Steve Brown24. His core idea is the fusion of two minds, the coming together of a mentor and a mentee. It’s not quite the above two elephants, but it makes so much more sense than your generic stock image.

Mentor and mentee at re-create.com25
Mentor and mentee at (View large version26)

Another example is from an about page we did for activpayroll27, a global payroll company based in Aberdeen. Instead of simply writing “About us” and waiting for the client to choose a stock photo, we create synergy (cue mental image of elephants procreating) by linking the image with the content, which talks about the DNA of the organization.

It's part of our DNA' – activpayroll.com28
‘It’s part of our DNA’ – (View large version29)

Last, but not least, the history page for Cruise Loch Ness (yup, they do cruises on Loch Ness). This piece of text was written by a copywriter, and Steve had the great idea of letting the typography sink into the loch. Isn’t it beautiful?

Cruise Loch Ness' emotions run deep30
Cruise Loch Ness’ emotions run deep. (View large version31)

Creating empty carousels and sexy hero images is easy. Adding meaning through design is more difficult. But that’s exactly what people should be paying us for. To put it bluntly, if your job is to re-skin Drupal themes, you’re not a graphic designer.

Does Everything Need A CMS? Link

Content management systems often force your hand when you do a layout; for example, by requiring content areas to be variable height, or using overlays over images because you don’t know when the client will change the background. With a CMS, we have to be generic.

Without a CMS, we can be much more bespoke, making use of creative typography, fixed positioning, animation, unique illustration, and so on. And chances are you might not need to CMS everything. So before you start limiting your ideas, ask yourself two questions: how rare is the content I’m designing for, and how likely is it to change?

Do you really need to CMS that bit of content?32
Do you really need to CMS that bit of content? (View large version33)

Rare content types – that is, content that isn’t replicated throughout the site – usually don’t need a CMS. And the longer the lifespan, the more you can invest in it. Common content, like a news article or a product listing, is likely to need a CMS, because they tend to have relatively short lifespans, and their content type is used across multiple entries on the the site.

Waterstons: building better businesses34
Waterstons: building better businesses. (View large version35)

The feature (above) we did for Waterstons36 is a good example. This whole feature area is hard-coded. We wrote it. We commissioned an illustrator for it. We animated it. And the client can’t change it. This doesn’t mean it can never change. Here’s their first campaign after the launch of the site. New copy. New illustration. New animation.

Five ways to performance improvement37
Five ways to performance improvement. (View large version38)

And here’s what we launched at the end of last year:

The Great Quest: Raising quality and lowering costs39
The Great Quest: Raising quality and lowering costs. (View large version40)

Is this more expensive than a standard hero image with a text editor? Well, yes. Is it more inspiring? We certainly think so. Is it worth it?

“The best money we ever spent”

– Michael Stirrup, Financial Director at Waterstons

Right. Moving on.

The Holy Book Of Don’t-Do-This-But-Do-That Link

As a community of designers, researchers, and developers, we love to share our expertise, and that’s one of the best things about this industry. Ironically, it is also one of the main obstacles to creativity.

We’re simply too good at telling each other what to do and what not to do. Everywhere you look there are helpful guides, tips, and hints. The problem is, the more guides there are, and the more we repeat them, the more they start to sound like commandments.

  • Thou shalt not define thy line-height in pixels.
  • Thou shalt not base thy breakpoints on devices.
  • Thou shalt not take the name of the lord Performance in vain.
  • Thou shalt begin with mobile.
  • Thou shalt not put important shit below the fold.
  • Thou shalt not worship the carousel of false idols.

Et cetera.

Sharing our expertise is ironically also one of the main obstacles to creativity.41
Sharing our expertise is ironically also one of the main obstacles to creativity. (View large version42)

Our rulebook is problematic for two reasons. The first is arrogance. We start to think we’ve solved it. That we’re finished. That, if everyone just follows best practice, we’ll create the perfect web.

Now, there will always be best practice and best practice is usually good practice. We absolutely need a standards movement. We absolutely need to consider performance, and usability, and all the other things that make web design difficult. The danger is that when too many people walk the same path, it becomes very uncomfortable to create an alternative, particularly as Twitter is full of firebrand preachers who take it on themselves to sniff out blasphemy.

The second problem is this: with so much focus how to do stuff, how to implement stuff, we forget our most valuable assets: meaningful ideas. By creating so many rules for ourselves, we miss out on those ideas. We put blinkers on, and we fail to explore concepts that don’t fit in with the current dogma.

Breaking The Rules Link

I have a confession to make. I use parallax. And carousels. Even hamburgers. Shocked yet? I doubt it. But I can sense your judgement. As our industry matures, we see the rise (and sometimes fall) of contentious issues that polarize our community. Yesterday it was hamburger menus, today it is page bloat. Tomorrow? Who knows.

Are carousels inherently bad? Of course not. Should parallax be banned? Don’t be daft. Do hamburger menus work? Well, I guess the jury is out on that one.

When I was designing magazines, no one ever talked about click rates, or conversion funnels or usability patterns. We simply had to trust that our readers knew how to open the magazine, how to find the contents page and how to read page numbers. And most people did. I think. (We had no accurate way of checking.)

On the web, things are not that easy. There are so many things that can go wrong when you can’t give everyone the exact same experience. Does this behave the same way on a slightly older version of Firefox? Do people know what the hamburger icon means? Do they even scroll? Luckily, we can measure things!

Worshipping metrics43
Worshipping metrics. (View large version44)

We desperately love metrics, and we lap it up when someone publishes their results. If you’re into your hamburgers, you’ve probably seen James Foster’s A/B tests, which proved that putting the hamburger inside a box, so it looks like a button, increases use by 22.4%. It also proved that switching the lines for the word ‘Menu’ makes 20% more people click.

I’m using the word ‘proved’ quite generously here. What I actually mean is that Foster’s tests showed us particular results from a particular period of time on a particular site, using a particular set of hamburger icons. In other words, as far as everyone else is concerned, the tests prove nothing.

Why? First of all, because A/B testing excludes the rest of the alphabet. We cannot know what would happen if the button was in a different color, a different size, or in a different context altogether. What if the home screen only consisted of a hamburger menu and nothing else?

A/B testing excludes the rest of the alphabet45
A/B testing excludes the rest of the alphabet. (View large version46)

Second, there are simply way too many factors at play, making us unable to pinpoint exactly what determines the results. It could be differences in the technical background of the audience. It could be affected by what mobile phone they were using. They could be influenced by the apps they are already used to. By their emotional state. By the temperature outside, or whether they were in a rush, or in a meeting, or on the toilet.

Or, as this simple A/A test47 shows, it could be coincidence.

I’m not knocking A/B testing (or James Foster, for that matter). By all means, test your stuff and see what works best for you. My point is that someone else’s A/B test rarely has takeaways for your own project. Despite our best efforts or intentions, we simply don’t have the tools, or data, to say how well something works across all contexts.

Despite our best efforts, the answer to most design related questions is still “It depends.” Throwing pseudo-scientific stats around Twitter doesn’t change that.

Think about this: if carousels are inherently bad, how come practically all smartphones use them as the main means of navigating? Most of us use a carousel every single day. Which means that as long as people a) know how to operate the carousel, and b) actually want to see what’s on the next slide, carousels may actually be all right.

It means that at least in some contexts, carousels work. And the same goes for hamburgers, or parallax, or even pages weighing a ton: it depends.

So I say fuck the rules (responsibly). Rules are stifling.

Sure, we’ll make mistakes. Lot’s of them. We’ll make horrible things, bloated things, painful things. But every now and then, we’ll make something extraordinary. Something wonderful.

And I’d much rather have the web be a mix of horrible and wonderful than an ocean of beige.

Pattern Fatigue Link

The term ‘lateral thinking’ was coined by Edward de Bono48 in 1967 and made famous by his 1970 book49 by the same name. In this book, he talks a lot about patterns and how difficult it is to move beyond patterns that have already been established. As an example, he provides the reader with a puzzle that involves combining an increasing number of shapes to create a larger, simple shape (such as a rectangle).

Adding each piece is easy until you get to the last one50
Adding each piece is easy until you get to the last one. (View large version51)

Adding each new piece to the existing shape is easy. Until you hit the last piece. What then? Most people really struggle at this point, but as de Bono points out, the solution is to reevaluate the whole pattern – not just the final piece.

Ignoring the initial assumptions make the final solution easier to reach52
Ignoring the initial assumptions make the final solution easier to reach. (View large version53)

I wonder if we’re stuck right now. We’ve got all these patterns, and they fit so well together that it’s almost impossible to imagine anything else. It feels like our patterns are complete.

There are many examples of so-called completed patterns online. Hamburgers. Horizontal navigation. Pagination. We even have patterns that dictate our page layouts, like the F pattern.

“F for fast. That’s how users read your precious content. In a few seconds, their eyes move at amazing speeds across your website’s words in a pattern that’s very different from what you learned in school.” – Norman Nielsen Group

People read differently on the web, we’re told. In fact, they don’t actually read our pages, they scan them. They start at the top, going across. Then they move down the page a bit and go across again, before they scan the content’s left side, in a vertical movement.

Every website, ever54
Every website, ever. (View large version55)

It’s no surprise we design to this pattern. We put our navigation at the top, our content directly below it, and our sidebar on the left-hand side. It makes sense because that’s how users read websites.

But this is a chicken and egg situation. What comes first, the F reading pattern, or the F design pattern? I think people read our sites like they do because we design them that way. I think that if we changed our design patterns, the reading patterns would change too.

“We need creativity in order to break free from the temporary structures that have been set up by a particular sequence of experience.”

– Edward de Bono

In order to become more creative, we must think more like children. We must dare to doubt the obvious, to be curious about the mundane, and to ask the silly questions. Questions like, “Why do we always put the logo in the top-left corner?” or “Why do we always put the navigation at the top of the screen?”

We don’t have to. There are tons of places we could put the logo, and there is no reason other than expectation not to put the navigation at the bottom of the screen. Why not ditch the navigation altogether?

Why are we so obsessed with having every page accessible from anywhere, instantly? Without the navigation as a primary means of moving around the site, perhaps the content can do the job instead. Perhaps every website could have a contents page, leaving all the other pages free of clutter. Perhaps that means we’d discover a whole lot of content we wouldn’t necessarily seek out. Perhaps we’d consume content in a different way. Maybe getting lost is a good idea sometimes?

What if we designed for landscape instead of portrait?

Screens aren’t just getting smaller, they’re also getting bigger. And most of our non-mobile viewports are landscape. Why do we think a long, vertically scrolling wall of content is the best experience for a landscape device? What if we broke our layout into sections, spanning the full width of the viewport?

What if websites read like magazines?

You may think I’m an idiot. That I’m going too far. That challenging patterns that have taken years to develop is unnecessary and dangerous. Well, you wouldn’t be alone in thinking that.

“One of the biggest problems in modern web design is that everyone keeps reinventing things that don’t need reinventing.”

– Matt Hill56

For the record, I have since spoken to Matt, and he pointed out57 that he’s not advocating a lack of creativity, rather “lamenting the constant re-engineering of solved problems at the *expense* of being creative. IE, designers tweaking pointless things and ignoring the big picture.”

A fair point. Looking at the bigger picture, I believe Matt – and many more – would agree that we’re not done inventing yet. We’re not actually finished. In fact, as far as I’m concerned, we’ve barely scratched the surface.

So, Is The Internet Killing Creativity? Link

I have discussed content, rules and pattern fatigue as major obstacles to creativity online. So where do we stand? Is web design dead yet?

In a word, no. The internet hasn’t killed creativity, and web design isn’t dead. But generic solutions58 are. And the robots really are coming. There are so many tools, there are so many frameworks. There’s SquareSpace. There’s the Grid. And others will follow.

Soon, it will be virtually free to create a nice looking website. There will be super sexy templates. There will be themes. There will be payment plans and production lines. If it can be done with an algorithm, it will. Maybe it even should.

But remember this: machines have no imagination. Robots work to a specific set of rules. They follow the recipe. They can’t create something that’s original and meaningful. They have no imagination. If we want to survive the Matrix (my robotic dystopian metaphor of choice), we need to maintain ours.

So we’re left with a choice. We can take the blue pill: we carry on as we are. We perfect the recipe. We choose the familiar. And eventually we hand our jobs over to robots.

Or we take the red pill: we go beyond the obvious and enter Wonderland. We fail. We fail again. And in the process, we offer what the machines can’t: memorable, original and creative experiences.


Note: this is a write-up of my talk delivered at SmashingConf Barcelona 201559.

Footnotes Link

  1. 1 ''
  2. 2
  3. 3 ''
  4. 4 ''
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9 ''
  10. 10 ''
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16 ''
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21 ''
  22. 22
  23. 23
  24. 24 ''
  25. 25
  26. 26
  27. 27 ''
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36 ''
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
  46. 46
  47. 47 ''
  48. 48 ''
  49. 49 ''
  50. 50
  51. 51
  52. 52
  53. 53
  54. 54
  55. 55
  56. 56 ''
  57. 57 ''
  58. 58 ''
  59. 59 ''

↑ Back to top Tweet itShare on Facebook


Espen Brunborg is Head of Design at Primate, a web agency driven by an overwhelming passion for the web industry and a slightly unsettling love for monkeys, and is an advocate of content-led design, simplicity and typographic principles. He writes about his design convictions at 8 Gram Gorilla and his tweets are occasionally worth reading.

  1. 1

    Love this article. I’ve been bored with modern web design lately and I think you hit the nail on the head with this article. There is definitely a time and place for CMS and templates, but I miss the days of highly creative websites that set a unique tone for the client and amazed users. I’m striving in all my designs to bring the sense of amazement and delight back to websites.

    You’re right that the web isn’t dead, we just need to know when to break the rules and do something unique, and not be afraid to do so.

    • 2

      A traditional graphic design has ALWAYS had “UX” under their belt. A brand system works with TV spots, radio, stationary, website, billboards, flyers, postcards, magazines, etc, etc. All kinds of different “UX” experiences problems that need to be addressed. It’s always about your audience and thinking about their needs. Nothing new here.

      We need to get back to giving the graphic designer all the props and rid ourselves of specialized titles that feel “made up”.

      I’m a graphic designer that loves to read Jakob Nielsen articles. OMG, what do I do now? Change my title? I’m a UX designer!!??? Who do I blame for my conflicting identity? Maybe it’s the powerful business owners, like Amazon, using their influence and money to reduce the graphic designers worth?

  2. 3

    i couldn’t agree more with the idea that technology has a way of taking our awareness and energy away from the natural states that cultivate creativity and imagination. Unfortunately the influences of our capitalist culture find their way into every environment, and this includes those where our most creative and talented reside.

  3. 4

    I mostly agree with your opinion and I feel often the same BUT in the other hand most of the usual clients we are all working for doesn’t expect any “breaking the rules” and I think this is an fact, which doesn’t change that fast.

    And what kind of minimal changes should we do to create unique and smart webdesign? You’re examples were not really different to most of the flat websites, or am I wrong?

    The grade of graphic design is down to a minimal since flat design and I agree with you that robots doesn’t have imagination BUT they are 1) cheap 2) fast 3) easy and a lot of people upload a picture of their company and they are satisfied. Hard topic – and a lot of different views!

    • 5

      Helen_OneLine ethical design

      January 31, 2016 12:42 am

      I hear you. Perhaps though we should be educating our clients that if we don’t break the rules they won’t get what they really want, which is to be noticed.

  4. 6

    I think clients that have no concept of design, GUI of art are the problem.

    How many times have we heard… “its great but can you change just one thing” and after placing “blink” and a waving Homer Simpson on their corporate site, the customers happy!

  5. 7

    Markup rules.
    Accessibility rules.
    UX rules.
    Code standards.
    SEO rules.
    Expert dos and donts.

    We do not provide creativity. We provide rules. Some trying to make a living. Others already living out of this world. A monster like the internet today will crush without rules.
    Out of fear, afraid to fail – that is -, nobody tries to reinvent the wheel. One failure could end your internet future.
    It is just fashion, folks. Who liked the same type of shoes endlessly?
    And… well, beautiful or not, more or less fashionable, big brands are setting trends. So there is something to blame, because this resource says it clearly: design is dying.

  6. 8

    About those A/B tests; I concur with your thoughts, and stressed that my results were from one particular site only. I believe this kind of testing is only useful when you have a very clear idea of your goals.

    Services such as The Grid and Squarespace, etc, are changing the industry. I only recently volunteered for a site design and build, only to discover the client found Squarespace and I was redundant.

    Oh well, I guess I’ll have more time to do some more controversial A/B testing…

    • 9

      Espen Brunborg

      January 27, 2016 10:30 pm


      I was in doubt as to whether to include a reference to your tests, as I don’t want you to think I’m having a go at you. As you point out, your results were from one particular site only, and in that context they are, by all accounts, completely valid and useful.

      Thanks for not taking it personally.

  7. 10

    Jakub Linowski

    January 27, 2016 8:49 pm

    To say that “the tests prove nothing” is an unscientific, desperate step backwards into the dark ages. When you go against experimentation, you are left with opinion, and the design-as-magical-superhero model. Dark ages. Witchcraft. I say, screw that. Show me the data.

    I agree with you that open mindedness requires some skepticism, and more importantly probabilistic thinking. Agreed, it’s childish to perceive rules or best practices unquestionably and expect them to guarantee results 100% of the time. Your hinting at “good practices” slightly clears us from such naive thinking – thank you for that. It might be what Carl Sagan meant that all scientist are never to be 100% certain and always leave some room to challenge their own beliefs and build new knowledge. I like that about your article.

    But to say that we cannot build knowledge from tests and strive to capture patterns as a means to attain higher probabilities of desired effects, is an attack on design science. Sorry, your argument that A/B are useless because C,D,E,F,G,H,I,J have not yet been carried out is flawed to the bone. Scientists generate knowledge even though an infinite amount of experiments have not been carried out. Infinite amounts of experiments cannot be carried out by definition. True, that strength of a pattern/finding rises when multiple experiments are run. Rigor and multiple comparison need to be made. In that light, Foster’s A/B test about hamburger menus is a starting point and should be thanked for, built on, compared to other tests, but not ignored. Don’t throw stones at someone for sharing design data! It takes guts to share data openly. Foster is the hero here for being transparent. Closing your eyes to experiments is what scared people do – medieval.

    Data does not kill creativity. Creativity comes from discipline and experience as design theorists such as Donald Schön, Kees Dorst, and Bryan Lawson have written about extensively. If you’re not creative, it’s your design process, and not scientific thinking. Don’t crap on A/B testing for killing creativity.

    Do stay skeptical and open at the same time. :)

    • 11

      Espen Brunborg

      January 27, 2016 10:23 pm


      Thanks for the constructive feedback.

      First of all, I’m not saying A/B testing is bad (in fact, I specifically point that out in the article when I say “I’m not knocking A/B testing”), and I’m not “throwing stones” at James Foster (ditto, see comment above).

      What I am saying is this:

      1. If you A/B test your own product, with a large enough sample size over a sufficient time period, you will know with some scientific confidence that A or B performs better than the other. However, that is also all you know. You cannot know how C, D, F etc. would compare unless you test them. Therein lie the false comfort of A/B testing – it is too expensive to test twenty radically different ideas, so we limit ourselves to incremental tweaks and end up perpetuating ‘safe’ patterns already tested by others.

      2. What works well in one context does not necessarily work well in another, and vice versa.

      • 12

        Jakub Linowski

        January 27, 2016 11:05 pm


        To your points:

        1. The degree of change is a variable that can be designed into the experiment, openly. You can have a B that is incremental or vastly different from A. We have ran radical tests as one strategy: We also ran incremental tests. On a side note, nature is incremental and gives rise to a really diverse range of species. :)

        2. The fact that there are various variables acting on participants in a test, do not automatically guarantee that they have an effect. So if someone is sitting “on the toilet” (in your words) during an experiment, that may or may not have an effect. More so, with enough of a sample, those difference are actually irrelevant because both A and B should be exposed to them more or less equally with random sampling.

        My main point, I think, is not to blame data or experimental thinking for some of the creativity problems you identified. Experimentation has immense value. And having 1 test result makes us smarter than having none. Having 10 or 100 test results gives us an even greater advantage. There is nothing pseudo-scientific about sharing 1 test result openly. On the contrary, your logic about carousels might be laced with pseudo-science (“if carousels are inherently bad, how come practically all smartphones use them as the main means of navigating”). Pseudo science gives way to science when common / general opinions are actually tested.

        • 13

          Espen Brunborg

          January 28, 2016 11:13 am


          1. I agree. (On your side note, nature works in increments because of evolution, which is completely void of design. There *is* a force of evolution happening in web design, of course, but that’s a topic for another day.)

          2. Yes, “sitting on the toilet” may not have an effect on your test. But then again it may. My point is that we don’t know. Contrary to controlled experiments, most A/B tests have no real way to account for unknown factors. As you say, with a large enough sample, such factors may not matter. But what is a large enough sample?

          Finally, my logic about carousels is not pseudo-scientific. I’m not using the example of smart phones to argue that carousels are always good, only as proof that they’re not always bad. If the majority of the world’s smart phone users – a large enough sample size in my book – are comfortable with a carousel as navigation, it stands to reason that it works in that particular context.

          Thanks for the discussion :)

          • 14

            Jakub Linowski

            January 28, 2016 8:50 pm

            Data trumps popularity, beliefs and fads.

            Popularity is not proof that something if good or bad.

            Smart designers don’t follow fads, but become knowledgeable and use data to their advantage. Unless they are not designers, but artists who create stuff to fuel their own fancy.

            Pseudo-science hides under rocks that cannot be measured or falsified.

            Here is some summed up data about where carousels have been shown to be less effective along with some possible explanations:


        • 15

          That Individual

          February 10, 2016 6:49 pm

          While saying that “tests prove nothing” may seem a bold claim (when divorced from the article’s context, where it points how badly done tests prove nothing), saying that shunning such tests is a step in the dark ages is a very harsh thing to say.

          I’ll start with a harsher counter-claim of my own: if you do not have a solid grasp of (at the very minimum) univariate frequentist statistics you have no business in conducting statistical inference. I make this claim with a background of conducting basic statistics in both a universitie’s psychology department and an economics deparment. While I am not (not even by a long-shot) an expert, I can recognize when something is above my head. In our economics department we see A/B testing as a basic statistical tool destroyed by marketing woo.

          Touching upon “incremental enhancement”, frequentist statistics does not easily lend itself to “increments”. If you are doing repeated A/B testing (on let’s say, navigation) and in one quarter of the year, version B “wins” then in the next quarter you have B against C and B still wins, then you have B vs. D and D wins, with this between-groups experimental design you have stacked the type 2 error (in this scenario) three times, where an ANOVA on A/B/C/D would’ve proved more reliable. Using repeated A/B testing, in this scenario, we cannot guarantee that D is better than A, unless you can guarantee that you had a uniform experimental methodology (which you can’t, cause you don’t). Also, I do hope you run Shapiro-Wilk (or any) normality tests on the distributions of your data to ensure that you can apply the parametric version of your test.

          As a last remark, repeated measuring of data and tweaking of models lends itself better to bayesian statistics, where until recently more complex models were analytically unsolvable. Thanks to developments in probabilistic programming, they can sample the distribution with markov-chain monte-carlo stochastic simulation. I say “they” because only our math/CS departments have knowledgeable enough people to run such simulations. We defer that when we need to use such models it must be made in conjunction with those departments. Sadly, a website rarely (if ever) has a dedicated statistician on hand.

          Statistics is not a plug and play add-on or some dangling charm you add on your toolbelt. P-values are so misunderstood (even in academic circles) that they have been banned altogether from some scientific journals:

          “The perfect is the enemy of the good”. Of course nobody (except maybe(?) the big sites) has a statistician on-hand. That does not save you from the shortcomings of your own flaky experimental design decisions. (OpenStaxCollege has a wonderful statistics manual available for free, that I heartily recommend).

          Wonky procedures done on data are not science. It is an artifice of science, a parade, to please those that would wish the reassurance of “scientifically proving” their decision without the effort of scientific rigor.

  8. 16

    Jakub Linowski

    January 27, 2016 8:58 pm

    Oh an Mr. Kadavy is also guilty of discouraging scientific thinking by running an under-powered insignificant A/A test and claiming A/B tests don’t matter.

    I spent hours pouring over articles, learning just how to run a reliable, significant A/B test, and I came to this conclusion: It Doesn’t F*cking Matter.

    Childish. Medieval. Anti-science. :)

  9. 17

    Below are my initial thoughts to the article, forgive me if I have misinterpreted anything. And of course these are just my opinions, I am 100% open to discussion.

    1. A website doesn’t have to be “creative” to be good. Look at Smashing Magazine, does it have a “creative” design? No not really. But is it a good website? IMO yes.

    2. Creativity isn’t just about the visual design. Creativity can be incorporated into all parts of the UX, e.g. creating a unique UI interaction that does its job really well.

    3. I disagree that you have to throw away the rules (assuming of course you have the right rules) to be creative. IMO creativity is creating a unique experience while sticking to the rules.

    4. If a generic solution works… well there is probably a reason why it is a generic solution? Of course the generic solution is probably not the best solution but it is better for a small business to have WordPress template website than to have no website at all.

    • 18

      Espen Brunborg

      January 27, 2016 10:39 pm


      I don’t think you’ve misinterpreted anything :)

      1. Yes. Creativity for creativity’s sake is seldom a good idea.

      2. Yes. Typography, illustration, photography, UI elements, content – creativity works across all aspects of UX.

      3. Yes and no. Not all rules are good. Some are arbitrary. Some make no sense. For example, “Never use hamburger menus” is a stupid rule because it makes too many assumptions, whereas “Always prioritise legibility” is a good rule because it is universally applicable.

      4. Yes. Generic solutions are generic solutions for a reason (they’re cheap), and a generic solution is generally better than no solution.

      • 19

        I agree with your reply, thanks for the follow up :).

        I guess just after reading your article it felt a little like you were disregarding the science behind web design while pushing the creativity whereas IMO the two should go hand in hand. Though now I think maybe that wasn’t your intention, it was just how it came across to me.

      • 20

        Patrick Ashamalla

        January 28, 2016 12:30 am

        It’s interesting that the legibility rule was the example. One of the most important points I have ever read on design was in David Carson’s book, The End of Print. The line, which has stuck with me since I first read it almost 20 years ago, was, “Don’t mistake legibility for communication.”

        Not to discount your point. Your example (or its intent) is valid. That being said, if it’s worth making a rule about something, then it’s also worth breaking it if improves the outcome and if you are in a position to do so.

      • 22


        I think there is a missunderstanding between rules and guidelines.

        Red is a warmer color than blue

        Dont use the hamburger menu, unless you have a special reason for it.

        We have rules and it is often not really clever to throw them overboard because customers are familiar with those rules even if they don’t exactly know them. So if you decide to not stick to them, you absolutly must have a reason for it.

        You design a website about winter sports. But it should be a dynamic, hot site full of action. So, yes, you can say “allright, i go for a more red appeariance instead of using colder colors.” You have a reason to do it.

        To the toppic of internet killing creativity: No, it doesn’t. We are currently facing a time where standards getting truely standards. And this is not a bad thing because it helps people using the stuff we design. Look at books. They mostly look the same with some difference here and there. Its because this form of reading media is good and it works. So why trying to create a new form of book then? What you can do is what i would call “micro-design”. Not the complete product but the small pieces that come together like typografie, Layout of images and text etc…

  10. 23

    Infinite Digital

    January 27, 2016 11:14 pm

    I agree the internet is not killing creativity.

    But generic solutions are not killing creativity either imo.

    The very basics, fundamentals and structure that “generic solutions” provide us are the basic building blocks on which current design can rest on and take-off from. This realization, or the lack thereof, is what is really killing creativity.

    The basic rules, structures and patterns are there not to bind us but to provide a starting point. And I’m surprised the author doesn’t pick up on this considering their ardent support of Flat Design. (The majority of arguments I see against Flat Design is that it’s becoming a cumbersome library of rules and patterns that are restricting the way we design. And I whole-heartedly agree because while the basis of Flat Design is great, the every growing number of rules that bind Flat Design has gotten out of hand).

  11. 24

    Your absolutely right. Brilliant article, concisely expressed!
    That’s my rule book binned, i’m starting over (‘ ‘ ,)

  12. 25

    Read a very brilliant and inspiring article and, simultaneously, see a UI kit ADS is amazing

  13. 26

    Vladimir Jovanović

    January 28, 2016 10:08 am

    I have watched this lecture in Barcelona SmashingConf 2015. and it made left best impression on me. Especially because I am one of those who started as “web designer” and who sees that everything is just a template today – flat design, material design, apple design guidelines…

    On the other hand, I tend to code a lot as well and I can see where is the main problem – creative process has become too complex. Today I need 2 weeks just to make foundations in code, before I start to create something. I need to chose approach, to set up gulp process and plugins, to set up git repo, to select how will I make grid, icon fonts… This has become tiresome. I just want to dive into creative process like we had 10 years ago when Flash was highlight of web design, I don’t want to be slave to “revolutionary” frameworks with a lifespan of a mosquito.

    I feel that the change is near. We must work harder to make it come faster ;)

  14. 27

    Good article, I would say that people’s fears of failure is probably a big decider in such creative ideas. The process of getting something that is wildly creative and innovative to a point where a client may be happy for you to use it with out any tried and tested data behind it, is likely to be difficult. The bigger the client’s worth to you or your business, the less likely you are to take risks.

  15. 28

    This article rhymes one of the recent talks from Andy Clarke about UX killing Art Direction and is one of the best articles I have read in ages. I guess it’s because it really speaks to me, coming from graphic design myself and having been doing web design for 21 years now. I’ve seen rise, decline and fall of creativity throughout the years, as you described… more than fall or murder, I’d point to laziness but it’s kind the same to many.
    IMHO, the discussion around flat/non flat is completely pointless because by defending the use of this trend or another you simply fall into using someone else’s clichés and definitions. Define and use your own style (which may comprise the outcomes of a trend, I give you that) but not someone else’s, otherwise you’re not a creative: you’re merely an executor.

  16. 29

    reading this in some parts, i almost heard the villagers outside the castle with their pitchforks.. but we’re with you buddy… they may take our tags but they’ll never take our freedom!!!

  17. 30

    For too long, the medium has been the message. Likewise social. After a decade of pouring a lot of money and content in with little return, clients want guarantees and correct answers. Hence the rapid rise of UX patterns, user journey work, prototyping tools, split testing and design by data.

    But sadly that’s the right answer to the wrong question.

    Why aren’t audience X or Y converting? Maybe it’s because they don’t care enough about what you are selling or saying. You can have the best user experience, lots of traffic and a water tight SEO strategy. But if the product isn’t great and it looks and sounds a bit ‘me too’ you’re in trouble.

    Creativity is seen as a risk but successful companies don’t grow unless they take them. And tools like macaw, webflow etc might make that risk much more affordable.

  18. 31

    Thank you for this great article. It gives me inspiration…

    • 32

      Saerts, I concur. As I’m starting to enter the web design industry, I’m glad to hear that there is still room and opportunities for creative design.

  19. 33

    It’s a good thing that websites are starting to look the same! It’s a clear sign that the medium has matured and we’re ready to move onto other design problems.

    Does anyone have an issue with most books looking and functioning the same?

  20. 35

    Leanne Tremblay

    January 28, 2016 6:59 pm

    This is by far the best bit on design and content I’ve read in a long time. I’m from the content side: content strategy, writing, and content marketing strategy. I would love to see more innovation when it comes to content/design working together, but don’t know where to start. I love/hate best practices! Thank you again.


↑ Back to top