Breaking web conventions

Punks: the conventional non-conformists

Conventions are paradoxical. On the one hand they make our lives manageable, but on the other their nature prohibits progression. With usability testing and surveys constantly informing our decisions, are we web designers in danger of becoming shackled by our own conventions?

Modern technology – the Internet – is a beautiful thing. We have an unprecedented array of tools available to research, manufacture, play, share, create, distribute and, a godsend to marketers and designers alike, track data. We track reading patterns, we track colour combinations and the effects of copywriting. We track visiting time, button clicks, conversion rates and monitor sizes. Never before have we known our audience more intimately, and never before have we been more scared to offend them.

Through blogs, shared research and usability gurus we’re taught how our users behave and what makes them tick. We’re taught to test everything, incrementally honing our designs. We’re taught what works and what doesn’t – at least according to the latest A/B results from our favourite internet heroes. What we aren’t taught, though, is when to disregard the rulebook, because online the metrics rule and conventions are King. Yet the road from Kingdom to Tyranny is short and we should take care not to become enslaved by our precious metrics.

“And the guards of a king are citizens, but of a tyrant mercenaries.”

Aristotle, Politics

How conventions are formed

Conventions are patterns, and it’s our mind’s ability to create, recognize and categorise patterns that make them so efficient. If we didn’t have patterns, communication – and life – would be a very arduous indeed. Imagine if you didn’t recognise the basic pattern of a door handle: every single door you encountered would require you to learn how to open it. Patterns, and therefore conventions, are definitely an intrinsic key to our survival.

“A system that can create its own patterns and recognize them is capable of efficient communication with the environment.”

Edward De Bono, Lateral Thinking

Edward De Bono is the inventor of the phrase “Lateral thinking” (if you haven’t read the book already, go do it after reading this article), but his most noteable contribution is perhaps his theory on the mind as a self-organising memory system. According to De Bono, our mind arranges all our experiences into categories – patterns – in order to deal with the potentially indefinite number of impressions encountered daily. As means of explaining how experiences implant themselves on our mind’s “memory surface”, he uses the analogy of a shallow dish of table jelly:

“If hot water falls on this jelly surface it dissolves a little bit of the jelly and when the water is poured off a shallow depression is left in the surface. If another spoonful of water is poured onto the surface near the first spoonful it will run into the first depression tending to make this deeper but also leaving some impression of its own. If successive spoonfuls of hot water are poured onto the surface … the surface will become sculpted into a jelly landscape of hollows and ridges. The homogeneous jelly has simply provided a memory surface for the spoonfuls of hot water to organise themselves into a pattern. The contours of the surface are formed by the water but once formed the contours direct where the water will flow. The eventual pattern depends on where the spoonfuls of water were placed and in what sequence they were placed.”

In other words, new experiences reinforce the pattern of similar, past memories and our interpretation of new events is influenced by previous ones – to the point that a door handle that swings upwards would seem completely inconceivable and probably have most of us stumped for quite a while. The advantages of such a system is, of course, that we’re able to quickly read and respond to our environment. But there are also disadvantages, such as:

1. Patterns grow perpetually stronger and more rigid since they influence all future interpretation in their favour.

2. Thus, it is extremely difficult to change established patterns.

3. The sequence of arrival of information (or creation of patterns) defines how future information is arranged, meaning we’re unlikely to ever arrange our information in the best possible way.

4. Established patterns grow larger and larger as individual patterns are strung together creating long sequences so dominant that they constitute major patterns of their own.

5. The mind creates, feeds and uses clichés.

“In particular the mind is good at establishing concept patterns but not at restructuring them to bring them up to date”

Edward De Bono, Lateral Thinking

Common design conventions

We can arrange design conventions into four types – convention by necessity, convention by economy, convention by usability and convention by hegemony – and, for the most part, they are all helpful in creating products that work.

For example: If magazines generally didn’t have the masthead at the top of the cover, recognising your favourite title amongst a hundred others in a magazine rack would be almost impossible, and if a web site’s main navigation was in the footer it would be excruciatingly difficult to find what you’re looking for. These are conventions by necessity – the product simply wouldn’t work without them.

Economic conventions are equally easy to understand: If every magazine, book, letterhead or poster were of a different size not only would it cost a lot to print, it would also result in huge waste in paper production – creating a set of conventional paper sizes makes financial sense for everyone involved. Economy is also the reason why templated web sites are so abundant, as skinning a pre-built template is much cheaper than creating something from scratch (whether this approach actually makes for effective web sites is a different matter).

Conventions by usability exist to make products easier to use and aims to match our collective intuitions: Pushing a door handle down is physically easier than pulling it upwards, and left-aligned type works well because we, in the West, intuitively read from left to right (that’s another big old convention right there).

Whilst the original reason for doing things a certain way might change or disappear completely over time, the pattern relentlessly grows stronger with every passing day.

Conventions by hegemony, however, stands out from the first three in that it’s dominated and fed by other conventions – even if they’re aribtrary or coincidental. For example, clocks turn left to right not because it’s cheaper to manufacture or easier to use (right-handed people actually tend to prefer drawing circles counter-clockwise), but because they happened to be invented in the northern hemisphere where the shadows on sundials already moved in that direction. Fast forward a few hundred years and the pattern has stuck – it would now seem ludicrous to create a clock that turns in the other direction, even in the southern hemisphere! This example highlights the particular disadvantage of hegemonic conventions: Whilst the original reason for doing things a certain way might change or disappear completely over time, the pattern relentlessly grows stronger with every passing day. At a certain point it will be so ingrained in us that we’re unable to see other, potentially better solutions. Thus it is conventions of this type we must question and challenge the most.

Tyrrany of the metric

Let’s consider a common feature of web sites – navigation. Conventionally, navigation is almost always aligned horizontally along the top or vertically on the left hand side. Why? Whilst navigation has to be above the fold and easy to find (conventions by necessity), there is little evidence it absolutely wouldn’t work in the lower half of the screen or on the right hand side. “But our reading direction is top-left to bottom-right”, you might argue. That’s true, however that only really matters if the navigation is the first thing you want your users to look at, and that shouldn’t be a given (on most sites I frequent, the content is a lot more interesting to peruse than the navigation). The main reason why navigation follows such standard pattern, however, may be the overwelming amount of usability and eye-tracking studies that tell us that’s how people use websites. Fair point indeed, but consider this: Do we design web sites a certain way because that’s how people use them, or do people use web sites a certain way because that’s how we design them?

Case in point. Most web designers try to fit as much content as possible above the fold and we do this because, according to Jakob Nielsen’s Scrolling and Attention (2010), the average user’s viewing time is 80% above the fold against 20% below it. Considering the sites tested for Nielsen’s survey – standard layouts with a combination of top navigation, left hand sidebar and a main column full of text (and images, in one example) – the result is hardly surprising. But what if we ran the same survey on sites that deliberately hid content below the fold – would we get the same result? Take The Great Discontent, for example. Upon opening the site, the whole browser window is filled by huge visuals telling me nothing about the content below, apart from the name of the interviewee. If I want to know more I have to scroll, and I gladly do so because of the intriguing imagery up front. Should the content happen to be interesting I even scroll all the way to the bottom – contrary to what usability studies tells us – just like I would read a good book all the way to the end. I wonder, had TGD followed the conventions advocated by Nielsen’s scrolling study would I have given it the attention it deserves?

if what we’re testing already adheres to the conventions at play, we can’t expect the results to teach us anything new

We’re also told that web users read web pages in an F-shaped pattern (again somewhat unsurprising given the Western reading direction). Furthermore, when examining the sample designs provided with the article, it is clear that neither of them have been designed to inspire anything but F-shaped reading. Let’s say instead of traditional F-shaped layouts, we tested circular layouts, or layouts with attention-grabbing content along the right hand margin. Would that change the findings? The point to take away here is if what we’re testing already adheres to the conventions at play, we can’t expect the results to teach us anything new. Yet as soon as something is proven to work, we have a natural inclination to stick with it – trying anything else suddenly involves a risk – and our conventions grow even stronger.

Usability metrics gathered by the likes of Nielsen play a vital role in figuring out how people behave online and the quality of your average web site would no doubt be worse without them. However, we should be careful not to reach grand, sweeping conclusions about all web users based on relatively small sample selections – the fact that something works well in a defined context doesn’t mean it’s perfect across the board. Think about A/B testing. Although measuring whether “B” is better than “A” is unequivocally beneficial, it doesn’t say anything about “C” to “Z” – and by accepting “B” as the convention just because it outperformed “A” we’re denying ourselves the possibility of finding a truly amazing solution further down the alphabet.

Whilst relying on metrics when making design decisions may decrease the risk of failure, it also perpetuates conformity and hinders the transgression of outdated conventions.

Breaking free

A long time ago (or so it seems), most designers had no metrics to rely on. Focus group and customer surveys were expensive and time consuming, meaning good designers relied on intuition and experience to make clever decisions. Nowadays, in the online sphere at least, there are so many free surveys and tests and formulas telling us what to do that we’ve become afraid of trying anything else. Granted, following conventions and letting decisions be ruled solely by metrics can pay off – there definitely are situations where 0.05% increase in conversion can make a real impact on your profits – but what’s right for Amazon might not be right for the site you’re working on.

…to make real progress and find the patterns of the future we have to try new things and dare to fail.

As much as data gathered on other projects can be helpful, we have to retain our intuition and curiosity. Numbers can’t think, nor can they generate ideas. Consider innovations like the original iPod, with its game-changing scrolling wheel, or pinch-to-zoom technology. These didn’t come about because some study said “that’s what people want” – those studies came afterwards – but through creative thinking, intuition and a self-proclaimed wish to break conventions and set new standards.

Remember, only a fraction of what’s out there has actually been tested under scientific conditions, not to mention the bottomless pool of ideas yet to have been conceived. I’m not saying conventions aren’t useful – they absolutely are, and the world would be a very difficult place without them. But to make real progress and find the patterns of the future we have to try new things and dare to fail. Over and over again.

Image credit: todosnuestrosmuertos