sometimes we have a very narrow view of the world

Lost in translation: a primer for cultural differences in design and analytics

This article was written to accompany the summer 2016 Qlik Innovation & Design Research Digest, which is driven by a study into the BI maturity of organizations in the APAC region. I find this particularly interesting as the benchmarks and ideas that define the use of BI in an organization are on the whole based on Western business practices. In many cases still deeply rooted in the European book keeping practices of the 1900’s. This means that although we can assess BI maturity based of adoption, methodology and practice we should be careful that we do not ignore the subtle cultural differences that are also at play when we do so.

When we consider data and analytics we often think about the universality of numbers and the global language of mathematics as trumping differences in culture. This may be true at the atomic level but when we wrap these calculations in models, KPIs, ‘best practices’ and all the trappings of business intelligence usage, that universality is no longer so absolute. This is because data requires analysis, which is an activity performed on an object, and that object whether dataset, report or process is influenced by the culture in which it sits.

Culture influences the way we work with data and analytics in a huge variety of ways. From the way data is displayed, to the processes that it informs and more acutely, the way it is shared and utilized by people in an organization. Much of the BI data we deal with in our organizations ends up being ‘designed’ into reports, dashboards or visualizations, so let’s start with that; a few thoughts on how cultural differences can effect design.

How it looks is how it feels

The way something looks has an effect on how people understand and process the information it holds. Graphic and information design has been exploring this for decades, both striving for neutrality and simultaneously realizing that there’s always a voice, an authority and an identity. Those voices enable us to differentiate brands, to inform people via signage, and to assess the authenticity of a document. They set the tone of the communication. The design choices we make shape what is communicated. As the typographer and graphic designer Neville Brody once ironically said, if we were honest we’d only ever need one typeface.

The ‘design voice’ of a piece of information varies depending on the stylistic fashion of the times and the viewer/reader’s level of visual literacy. Both of which are framed by current cultural trends and the viewer’s historical cultural understanding. The understanding I personally have of aesthetics is framed by my western upbringing. The modernism, white space and ‘clean lines’ I love and think of as a clear authoritative design voice is a western cultural construct. When it comes to internationalism and localization, just porting a design from one culture to another is unlikely to be effective. Often it will underperform and in some cases even miscommunicate.

For example, there’s an old marketing anecdote about the Cola executive that tries to communicate the benefits of their drink in Saudi Arabia. He lacks Arabic language skills so he decides to convey the benefits as a pictorial story laid out through three posters, like so:

First poster: A man lying in the hot desert sand totally exhausted.
Second poster: The man is drinking the Cola.
Third poster: The man is now totally refreshed.

But no one told him that Arabic is read right to left.

It’s a fun example that’s often used to help teams designing for non Latin reading markets be more aware of their own constructs and assumptions. But it’s not only the obvious things like language, design concepts and metaphors that trip you up. In fact even the very building blocks of visual design (graphic & information design) are subject to culture nuances and misunderstanding.

Let’s take a quick look at color

I say quick, because color is a subject that is simple on the surface but will swallow you like a black hole once you start looking into it. So here’s a glimpse at some of the problems, starting with red and green:

  • Designers rely heavily on a color’s implied meaning, but that meaning is cultural. The most common example that trips up information designers is in finance. For western stock exchanges and money markets, the index grids and price tables often use red for negative, meaning loss and green for positive, meaning growth. However in China it’s the opposite. Red for growth and green for loss. Red in Chinese culture signifies good fortune and financial prosperity (gifts of money are given in red envelopes) and green by contrast can signify exorcism and infidelity. The western use of green for growth is rooted in it’s association with nature, this also the same in China but the not with monetary growth. In this case cultural understanding trumps established western best practices.
  • Red/Green color blindness (Deuteranopia & Deuteranomaly), affects the ability to distinguish red from green and is one of the most prevalent forms of color blindness. However this varies not just by gender (men are far more susceptible than women, due to it’s link to the X chromosome) but also geographically. And it varies a lot from country to country, for men in the US it’s around 7%, but in Norway it’s as high as 9%, whilst in Japan it’s 4% and for the men of the Fiji Islands it’s only 0.8%. (1)
  • Historically the availability of a color plays a role (at least in the physical world). A color’s current meaning is deeply linked to its historic use. But that use may not have been as orchestrated as contemporary readings of the color may imply. The red barns of North America are painted as such, not as a reference to the red or falu barns and houses of Scandinavia, and also not ‘so that cattle can find their way home’ (cattle are also red/green color blind). The reason is a little more practical. Red was the cheapest available color, barns where painted red or left bare.

Of course many other colors have their own issues. For instance the color name ‘blue’ doesn’t exist in some languages, such as Welsh. Color names can also change what they signify, ‘pink’ was once a term for yellow and it’s current use comes from dianthus flowers, whose notched petals resembled the edges left by ‘pinking sheers’. (2)

How colors are culturally used is even more fun. The blue for boys and pink for girls gender typing is relatively new and has reversed over time with guides as recent as the 1918 stating pink for boys, blue for girls.(3) Cultural events also play their part, such as the use of red to signify love for valentine’s day in the west. If you’re in the US, the ubiquitous Halloween holidays means that orange belongs to them, but in Japan it’s associated with courage and love, and in Egypt orange is the color of mourning. And just because it’s the brand color it doesn’t always mean it’s sacred. UPS’s brown is not well received in many countries so they use blue for much of their multinational marketing and communications — blue being one of the most universally positively thought of colors.

Color is one of the most basic yet most powerful tools to add meaning to information, but its use needs to be well thought through. What’s meaningful in one culture may not be in another.

What about typography?

Typefaces and their fonts are used to give words a voice. All fonts, even those that have been designed to be neutral such as Helvetica, have a voice and a tone. If you look at the explosion of typefaces since type design and printing went digital you’ll get a sense of the diversity of those voices. Almost every organization has a brand font, and each computer and smart phone operating system carefully selects and tunes the fonts available to reflect their desired voice. Our exposure to the use of different typefaces through printed and now screen mediums has helped assign character and weight to their voices. This means they can have an effect of how we process information.

If we expressed the same information in two different typefaces, which would you find more believable?

It’s very likely to be A. In a fascinating experiment run by Errol Morris in the New York Times online (4), he tested to see if one typeface was more believable than another.

Morris posed a quiz to assess if people were optimists. But what he was really doing was randomly showing the same information in different fonts to over 45,000 people in an attempt to assess if the font itself had an impact on the results. It did. The fonts he used were Baskerville, Georgia, Comic Sans, Trebuchet, Computer Modern and Helvetica. The results showed that information presented in Baskerville was more likely to be believed and agreed with than the other fonts. With (unsurprisingly) Comic Sans as the least believable. It would be very interesting to see this experiment carried out on an international scale. The fonts we have come to trust in the US and Europe seem closely linked the historic information sources we have learnt to trust, such as our ‘quality’ newspapers and official documents. As Michael Bierut of Pentagram put it, “In my mind, Baskerville speaks with a calm, confidence-inspiring English accent, sort of like Colin Firth. No wonder it’s so trustworthy.”

The power and history of typography is a deep and wonderful subject, for now we will just look at a few things that influence the international outlook.

First up there’s the three major writing systems: Alphabetic (such as Latin and Cyrillic character sets — ordered, standard sets of letters representing phonemes), syllabic (such as Japanese Kana — a set of characters representing discreet syllables) and logographic (such as Chinese — characters that represent objects or ideas). These aren’t rigidly separate systems as different languages adopt aspects from multiple systems. The differences in the systems mean that each contains their own typographic methods, approaches and quirks. Things that are common graphic design approaches in one often do not transfer to another. None of the following typographic techniques for alphabetic systems transfer directly to syllabic or logographic systems:

That’s right no caps, no bold, italic, underline (in syllabic and logographic emphasis is often shown by using additional marks), and many typographic ideas like ‘sinking text’ will not work. This impacts how we can use typography to layer on additional meaning and importance.

And I’m yet to find the equivalent of this passage, using logograms:

“Olny srmat poelpe can raed tihs.

I cdnuolt blveiee taht I cluod aulaclty uesdnatnrd waht I was rdanieg. The phaonmneal pweor of the hmuan mnid, aoccdrnig to a rscheearch at Cmabrigde Uinervtisy, it deosn’t mttaer in waht oredr the ltteers in a wrod are, the olny iprmoatnt tihng is taht the frist and lsat ltteer be in the rghit pclae. The rset can be a taotl mses and you can sitll raed it wouthit a porbelm.” (5)

Another area that causes problems is information density. Variance in word length in different languages has long been the bane of designers. This is especially acute in UI and chart design, where short words and category labels can be the biggest problem. Words with under 10 characters in English can expand up to 300% in their width, as Flickr found when translating ‘views’ into Italian; ‘visualizzazioni’! And it’s not just width, the line height and character height needs to be thought about too, which can have a major impact on the original visual hierarchy of a page or screen. If you think abbreviations are useful ways to save space, think again. Arabic words can be very hard to abbreviate as they are often constructed of “very compact, pattern-based roots with prefixes, suffixes and small internal changes to express the precise meaning”, according to W3 guidelines. Chinese and Japanese can have a very high information density compared to alphabetic languages. The logographic and syllabic character construction allows for very dense information display, with a lot of meaning packed tightly into a small area. Europeans with our modernist ‘less is more’ and love of white space can often find this overwhelming. Information density and most importantly the degree at which we are comfortable with it, depends of what we are used to. It’s about familiarity, as the old adage goes, reader’s read best what they read most.

How it flows reflects what we are used to

When we lay out information, we use typographic rules and concepts that suit our understanding of readability and perception. Tabular information is a great example this. We have rules and best practices about left and right aligning of data in columns and how different types of data should be aligned. On the most basic level for English, categorical labels are left aligned and numeric values right aligned with a fixed number of decimal places. These universal numbers (Arabic-Hindu numerals) are odd in this regard as they read right to left, not so surprising when we consider their origins. However for many Japanese readers the right alignment of number values feels very wrong. For them centered values better support their characters and established reading direction from top-bottom. This is something that westerners would consider a design no-no, however numerals are doubly odd in English as they are in fact logograms not alphabetic glyphs.

As we saw with the Cola illustration reading direction has consequences beyond the words on the page. Interaction design and data visualization rely heavily on directionality and flow. When this isn’t linked to the physical and is instead associated with abstract concepts such and time and progress, things can get a bit tricky. Our western understanding that things progress from left to right is closely tied to language and the written word. We read from left to right, progressing forward, turning page after page from the start to the finish of a book. This strongly frames how we ‘see’ time. Visualization best practices tell us that time on a time series chart should progress (move forward) from left to right, with the axis on the left. But Arabic readers, read from right to left. For them time series charts feel more natural if they flow from right to left with the axis on the right. And studies have found that top to bottom reading languages, such as Chinese also give the reader a different view of the flow of time, in those cases this is often from top to bottom.

Language and reading direction influence our idea of time and how we perceive and understand sequence. An implied order exists for how to read a piece of information or a navigational flow, or a series of instructions. That order is cultural and can confuse and impact usability if we enforce a model upon it.

What we perceive is not universal — WEIRDness abounds

Much of our understanding in the west of perception, communication and cognition is founded on psychological studies carried out on westerners. Over the past ten years there’s been a growing body of research that shows that the universal facts we had gathered regarding basic human perception are in fact not so universal. Mostly this is because WEIRD (Western, Educated, Industrialized, Rich and Democratic) people, particularly those from the US, were the main participants in many of the foundational experiments that set our understanding of basic cognition and visual perception. This is beginning to be challenged and research is revealing international and cultural differences in how we perceive the world and make decisions.

For instance experiments with adaptations of the Müller-Lyer line illusion (shown above) have revealed culture differences. Undergraduates from the US were found to be the most susceptible to the illusion and people of the San foragers from the Kalahari unaffected by it. And this isn’t by a small margin. “On average, the [US] undergraduates required that segment ‘a’ be about a fifth longer than ‘b’ before the two segments were perceived as equal”.(6) This has been used to highlight perception differences in industrial vs small scale societies, but there’s also some interesting findings around eastern vs western societies. Unsurprisingly these are less to do with visual perception and more to do with economic decision making. One of the main areas is in the use of holistic and analytic reasoning.

Holistic thought involves an orientation to the context or field as a whole, including attention to relationships between a focal object and the field, and a preference for explaining and predicting events on the basis of such relationships. Analytic thought involves a detachment of objects from contexts, a tendency to focus on objects’ attributes, and a preference for using categorical rules to explain and predict behavior.” (7)

Both cognitive systems of reasoning are available in most adults, however research has shown different societies have a tendency to rely of one form over another. Westerners (American, Canadian, Western European) are far more likely to be reliant on the analytic system, at the expense of the holistic system. With Asian, Arab, Malaysian and Russians being distinctly less reliant on analytic thinking.

“Westerners are also more likely to rely on rules over similarity relations in reasoning and categorization. Chinese subjects were found to be more likely to group together objects which shared a functional (e.g., pencil- notebook) or contextual (e.g., sky-sunshine) relationship, whereas Americans were more likely to group objects together if they belonged to a category defined by a simple rule (e.g., notebook-magazine; Ji et al. 2004).” (8)

These cognitive systems effect how we assess information when making decisions. When we present information the role of associated context may well be more significant in the decision making process in eastern cultures compared to western. If we are presenting a narrow isolated piece of data, will context be assumed? Designing information so that any relevant context frames it is crucial to helping both holistic and analytical thinkers and thus is essential to designing with an international outlook.

What’s more important, analytic maturity or effective decision making?

Can we judge BI or analytic maturity? As James Richardson discusses in his piece on the survey results, yes and no. We can’t really judge country against country based on the artefacts and methods that are used. It can only be on the adoption of practices and how deeply a culture of analytics is established in a given organization. An organizational culture of analytics, is an interesting idea but may not easy to create especially when we consider what we understand about the cultural differences in reasoning between western and the eastern cultures.

Often we simply port our business tools and processes to new markets and geographies, without even thinking — at best just translating the text. But how we read, interact with and communicate the insights from data is always framed by the culture in which we do it — it’s never simply the numbers. Yes, there are initiatives to establish international standards for business reports and charts, such as the IBCS (9) and the 20th century tyranny of English as the international business language. But these organizations and ideas are secondary to the culture and activities that people participate in daily. They can be adopted, but at best they are merged into broader cultural understanding and hopefully a more nuanced view of the world.

As with all design and communication, in the end it’s all about people; knowing them, what motivates them, what frames their activities and what they understand. What BI needs to supply to those people is useful, meaningful information that works with them and fits their culture: a range of assets and processes that support evidence based decision making and not the gut instincts and assumptions of a few managers. After all even those most ubiquitous of universally understood logograms, our everyday numbers (the Arabic-Hindu numerals: 0123456789) aren’t always the first choice everywhere.

image source unknown

References:

1 — Harrison et al. (1977): Human Biology, Oxford University Press, Oxford (via Wikipedia)

2 — Victoria Finlay, The Brilliant History of Colour in Art (2014)

3 — June 1918 issue of Earnshaw’s Infants’ Department, a trade publication (via — http://www.smithsonianmag.com/arts-culture/when-did-girls-start-wearing-pink-1370097/#hppmiV1mGeTsO1Al.99)

4 — Errol Morris, Hear, All Ye People; Hearken, O Earth (Parts 1 & 2) http://opinionator.blogs.nytimes.com/2012/08/08/hear-all-ye-people-hearken-o-earth/?_r=0

5 — There is no Cambridge research, but it’s a fun typography game, for the history and use see: http://www.mrc-cbu.cam.ac.uk/people/matt.davis/Cmabrigde/

6,7,8 — The weirdest people in the world? Joseph Henrich ,Steven J. Heine, Ara Norenzayan — BEHAVIORAL AND BRAIN SCIENCES (2010)

9 ­- The International Business Communication Standards: http://www.ibcs-a.org/