Designing for one context, shipping to another: The perils of supposedly Universal Design
Picture this. You live in a colder-than-freezing place, and design and manufacture a jacket for your usage. It’s a pretty good jacket, and so you ask your local friends if they like it. They do like it, and you end up manufacturing and selling jackets to your whole neighborhood, then town, and state.
You’re doing great, and so you decide to go international. You sell to markets in places with similar climates as yours, and the business does well. Then you expand all over the world, and suddenly find that the jackets don’t sell as well. You’re confused, but push on. You keep offering your winter-friendly jackets in tropical countries, and never realize why your jackets don’t sell as well.
In the industry, this is loosely what is known as ‘universal design’, the idea that something can be designed such that it works for everyone, across a whole spectrum of potential users. Obviously from the above example, you can tell how I feel about it.
In this story, and in some other cases, a universal design fail can just be passed off as silly or misguided, but ultimately benign. Users and non-users alike can simply laugh at some failed product or misfitting marketing, and move on. But in more severe cases, universal design can lead to inaccessible products that often exclude the most historically marginalized populations, most commonly users with disabilities. And in some extreme cases, products designed and built in one environment but deployed in another, can spell death for users in contexts in which they were not designed. Just ask Drs. Advaith and Ajmal Asif, of Kerala, India.
At 12.30am on Sunday the 1st of October in the Gothuruth area of Kerala’s Ernakulam district, the two, and three other friends, were driving back from late-night shopping. Completely in their senses and not under the influence of alcohol, Dr. Advaith was relying on navigation software to get through an unfamiliar area. They were following the navigation which was indicating that they should go through what appeared to be a waterlogged street, a common feature in the Indian monsoons. Without thinking twice about it, Dr. Advaith drove on the suggested route, and found themselves going into a river. The car began sinking and although three of the five were able to come ashore safely, Dr. Advaith and Asif lost their lives underwater. All because of their reliance on a service that is globally renowned, but ultimately not designed for their context: Google Maps.
This is not even the first instance of Google Maps leading users horrifyingly wrong into life-threatening or ultimately deadly situations. The first author themselves has lived experience of Google Maps directing them on to an unfinished flyover in India. Faulty directions took a family in New South Wales, Australia off the highway onto a dirt road that left them stranded in the Australian outback where they were stranded for two days, drinking water from puddles until they were rescued by a wide community search. A Spanish tourist in Rio de Janerio was shot and nearly lost their life after driving into notorious gang territory after Google Maps led them astray en route to the Christ the Redeemer statue. Tourists and hikers in the Italian town of Baunei were continuously misled on to secluded mountain roads instead of their intended tourist destinations over a 140 times in the span of a few months, leading to locals putting up signs saying “No Google Maps” in various parts of town. Even in the American contexts, Google Maps can mess up in terrible ways. In September 2022, a man in Hickory, North Carolina was directed by Google Maps to drive across what turned out to be a broken bridge that had collapsed almost a decade ago. The man fell to his death and his widow sued Google for the incorrect information, despite the fact that Hickory residents had suggested edits to Google Maps to indicate that bridge as broken two years prior to the incident.
Google Maps was designed in and by American researchers, based on the CIA-supported acquisition of geospatial data company Keyhole, the company whose satellite technology pioneered Google Earth. Designed and tested mostly in American contexts, Google Maps was made available to the rest of the world based on this satellite coverage, which disproportionately captured American/Western places and was geared towards gridlike city designs like a lot of American/Western cities. To expect that it would work the same in remote corners of India, Australia, Kenya, or other places that are not built the same as the places where it was trained is not only inaccurate, but also deeply dangerous.
Unfortunately, such design is not uncommon, and Google Maps is far from the only proponent of the one-size-fits-all model. The modern craze of LLMs and tools like ChatGPT across the world is demonstrating how well it performs in languages like English and contexts like the US, but the performance curve quickly tapers off in different languages and contexts, especially ones that are traditionally underprivileged. This pattern of designing in one context and shipping across others, without consideration for the difference in cultures, traditions, and practices, is an unfortunately common but symptomatic practice of designers with high degrees of privilege. Either intentionally or otherwise, designers can wield their privilege to design and deploy their products across different contexts without proper groundwork, living under the myth of universal design.
If you’re reading this and are a designer (we’re all designers, either by intuition or by practice), I implore you to consider the contexts in which you are designing, and whether your own mental models align with them. In cases where they do not, try working with the local communities for whom you are designing, talking about their specific needs and use cases. Work with them as ‘humans’, and not ‘users’: treat them as real people with real lived experiences, rather than users to extract information from. Most importantly, test in local contexts and think about how you can maintain and fix errors.
None of this is necessarily novel information: the existence of my field of Human-Centered Design predates a lot of these thoughts that I am putting out here. But I felt this to be an important story to tell, both in the honor of the people who lost their lives as they relied on products marketed to them but not designed for them, and for those who might yet be working to design what they intend to be universal design.
