Shove, actually: nudge theory in the built environment

You may not know it, but you’ve been nudged. Possibly several times a day. It might have been that last-minute health drink you bought at the office canteen. Or an arrow that got you to walk the long way between subway platforms and avoid the crowds. Or perhaps when a frowny face on a digital road sign reminded you to drive slower.

Cues in our environment have always influenced our actions. But since US academics Richard Thaler and Cass Sunstein brought the idea of “nudging” to prominence in their bestselling 2008 book, Nudge: Improving Decisions about Health, Wealth, and Happiness, the application of behavioural science has blossomed. By using psychology to understand how people make choices, the research shows, environments can be subtly altered to favour certain outcomes. The potential impacts are far-reaching, from boosting retailers’ profits to solving societal problems such as obesity and climate change — or the spread of disease, as when the UK government’s “nudge unit” helped shape its response to the Covid-19 pandemic.

Now the built environment is catching on. Architects and engineers are hiring behavioural scientists, or offering specialist services such as “workplace nudging”. The fast-growing WELL and Fitwel certification systems both recognize buildings optimized for healthy behaviours, such as eating fruit, drinking less coffee or taking the stairs.

But the simple message that designers can help drive “good” behaviour conceals a host of tricky questions: who gets to decide what “good” is? Is there room for a designer’s intuition if impacts on human behaviour are rigorously codified? And, if our actions are so readily changeable, are we in the end just robots?

In one sense, nudging has long been part of the designer’s arsenal. Back in the 1920s, London Underground’s Frank Pick was pioneering an approach to navigating the Tube that was responsive to human intuition. He ditched the old, cluttered map that organized stations by geography, instead introducing the schematic design still in use today.

Behavioural economics remains central to Transport for London’s (TfL) approach, says Ian Pring, its customer marketing and behaviour change planning lead. His team spends its time considering everything from platform markings that nudge people to mind the gap as they step off the train, to how station design affects perceptions of journey times. “It’s pervasive really,” Pring says. “Our job is to try to get people out of cars and onto public transport, and into cycling and walking. We also have a remit to keep people safe on roads and on public transport systems. So it’s arguable that behaviour change is at the heart of pretty much everything we do.”

Elsewhere in the built environment, the approach feels more novel. Joanna Frank, who was director of active design under former New York City mayor Michael Bloomberg, says interest has grown significantly over the last two years. She now runs the Center for Active Design, which promotes architectural and planning solutions to public health problems, specifically by translating thousands of peer-reviewed studies into practical design strategies, and is the licensed operator of Fitwel.

A decade ago, Frank was “jumping up and down” trying to persuade people of the virtues of the behavioural approach. New York had solved its infectious disease problem by improving sanitation and living conditions a century before, and her circle were convinced that they could do the same for chronic disease with architecture that encouraged activity and social interaction.

Frank sees it less as some shiny, newfangled power than an effort to tune in to what’s already going on. “Our environment is affecting our behaviour all the time, so it’s about understanding that. It isn’t that we’re suddenly saying, hey, let’s influence everyone’s behaviour using the environment. It’s really, let’s understand that we are influencing behaviour, and how can we optimize it.”

She gives the example of a meeting room: if the room has a table of standard height with chairs around it, everyone will sit. But if the table is standing height with stools, then people might sit or stand. Once you recognize how the furniture is influencing behaviour, you can do something about it — especially if armed with research on the choices people are most likely to make.

Susan Carruth is also keen on research. As head of behavioural design at Danish architect 3XN/GXN, her team includes an anthropologist and two PhDs, researching subjects such as the impact of plants on behaviour. She describes her job as “borrowing, stealing and raiding” from different disciplines in order to better understand “that relationship between mind, body and environment”. To some degree architects have always focused on behaviour, she says, but with a data-led approach there is room for so much more rigour.

Connecting the dots

“There has been a blossoming of research in other fields so we can measure and qualify and quantify things far more accurately than we could before,” Carruth says. “We can start to be much more precise, and that’s having a great influence: we connect the dots in a way that’s based on science rather than assumption.”

Some of those dots are mainstays of nudge economics. Nobel prize winner Daniel Kahneman’s thesis that we operate via System 1 and System 2 thinking — the first quick and instinctive, the second slow and deliberative — has spurred the use of intuitive visual cues in place of written signs. TfL, for example, has started painting blue footprints on the right-hand side of escalators to show people where to stand, and which direction to face.

Handwashing is another example: while research shows written hygiene signs to be largely ineffective, a 2014 study in The American Journal of Public Health found that when hand towels were dispensed automatically after someone used the toilet, soap use rose 13%.

Research around cognitive bias has also highlighted environmental shortcuts for influencing behaviour. Salience, our tendency to prioritize information that stands out, is one — in play when designers make stairs easier to find than lifts (which has led to a three-fold increase in movement at the International Olympic Committee’s Lausanne headquarters, says Carruth).

Social proof, our natural urge to behave as others do, is another. Mirror neurons in our brains fire both when we perform an action and when we see someone else perform it, “biologically compelling us to feel and behave as social creatures”, notes Carruth. Designers can harness this to promote things like communal eating in an office, she explains. So they might place the canteen where everyone will see it as they pass by, and alongside other things that draw people in, like exhibition spaces. These they call “watering holes”, because they are like wildlife habitats that pull different species together.

Desire paths — human rebellion over the predetermined route
Photo Flickr/wetwebwork

Such insights may open new doors for architects. But with so many requirements on them already, from space constraints to energy performance, is the need to nudge a burden too far? And does such a scientific approach risk suffocating architectural ingenuity?

Done right, behavioural insights should support the design process, not alter it, says Edward Gardiner, behavioural design lead at the UK’s Warwick Business School. Rather than adding extra layers, it’s about shifting mindsets to look at challenges from a behavioural perspective.

And intuition can be wrong, he points out. Gardiner quit a career in advertising after becoming disillusioned with the “lack of evidence and insight” going into many campaigns, and says we can’t assume that those in positions of influence — advertisers or architects — just know these truths and will apply them systematically. “People think that their own intuition may often be right when it’s not.”

Indeed, when it’s done badly, nudging can backfire. Lessons from history should warn us to tread carefully. For example, the rather paternalistic attempts by mid-20th-century architects to shape the lifestyles of apartment dwellers often foundered, with profound effects on communities.

The Pruitt-Igoe housing project, a 1950s complex in St Louis, Missouri, is perhaps the most notorious failure. Debate continues over the precise causes of its decline, but the vision of thriving vertical communities, where residents left their small kitchens to socialize in shared areas, did not materialize. Communal spaces theoretically owned by many were cared for by none. Elevators that stopped only on certain floors were meant to lead neighbours to interact in stairwells and corridors — instead they made them prone to muggers. The towers became a crime hotspot; the last was dynamited in 1976.

Tread carefully

More recent examples abound of design that is at odds with users’ instinctive preferences. Desire lines, one term for the informal paths created when we walk where they want rather than where we’re expected to, are an emblem of human rebellion against predetermination, with a whole community of devotees on Reddit. And Carruth recounts how one tech giant, in a bid to boost productivity, carefully placed its new office restrooms in out-of-the-way spots — thereby unleashing the fury of pregnant employees.

Unintended consequences are probably inevitable given the complexity of human behaviour, points out Jared Thomas, technical principal in behavioural sciences at WSP in New Zealand. “Engineers designing a bridge will be able to look at three or four qualities, measure those and say, ‘I can predict with 98% certainty that, under these parameters, this bridge will stay up.’ Our ability to predict human behaviour is probably more like 30%.”

This inherent uncertainty means that it is crucial to identify and measure behavioural indicators of success, and to move fast to adapt if things aren’t working as planned. Thomas gives the example of a road-safety project in Wellington to stop cars from exiting driveways onto a cycleway without giving way. It was a case of “looked but failed to see”, he says, a phenomenon where the brain focuses so narrowly — in this instance on other cars — that it blocks out everything else. So the team tested a range of interventions for alerting drivers to the possibility of cyclists. Paint that mimicked a pedestrian crossing combined with a speed bump was the most effective, roughly halving the number of misbehaving cars from 61% to 31% and reducing near-misses by 75%. By contrast, studs that lit up when they detected a bike only performed accurately 89% of the time, and — since drivers might learn to rely on them — potentially worsened the situation.

The speed of enquiry had real impact, says Thomas. By drawing up a list of success indicators, and using video to capture them, they established that the lights needed to be removed within two weeks: “Waiting for a crash history to build up would have taken five years.”

Researchers elsewhere are exploring the use of smart technologies and bigger data sets to measure how environments impact human behaviour. The Link Lab at the University of Virginia is a 17,000ft² office space with a network of sensors gathering real-time data on indicators like temperature and air quality while tracking the behaviour of approximately 150 occupants. “The more data we collect, and the longer we collect it, the more confidence we will have in our conclusions,” explains assistant professor Arsalan Heydarian.

Tech advances have also made it more straightforward to run real-world experiments, says Liam O’Brien, principal investigator of the Human Building Interaction Lab at Canada’s Carleton University, which focuses on energy saving behaviours. Digital building controls, for example, mean it’s easy to change settings on just one floor of an office, or set algorithms to progressively drop the temperature.

O’Brien’s research has highlighted default settings as a powerful energy saving tool. This idea, which relies on our inherent bias towards the status quo, is already deployed in fields like organ donation with striking results: according to Stanford University, donation rates exceed 90% in countries where people have to actively opt-out, compared with under 15% elsewhere. Similarly, when O’Brien’s team changed the light settings in a Carleton office so that they had to be switched on manually rather than coming on automatically, lighting energy use fell by 90% over the course of a year.

Another approach shown to drive energy-saving behaviour is feedback. When software company Opower started producing monthly reports for utility customers — complete with smiley faces for those using less energy than their neighbours — household consumption dropped; according to one study by as much as if there had been an electricity price hike of up to 20%.

Master manipulators

The effect of such relatively minor interventions may not just be good news for the planet, but for budgets too. As O’Brien puts it: “Technology is expensive; behaviour is cheap.”

But just because we can influence people’s behaviour, should we? How wary should designers be of seeking to exert such power over others? These are questions those working in the field spend a lot of time considering. Carruth says that ultimately nudge theory is rooted in the idea that there must always be a choice: “It’s not about enforcement, it’s not about taking away options, but about enabling certain choices.”

Nonetheless, there remains the question of which choices are enabled. After all, nudging is not the sole preserve of those wanting us to exercise more and save energy. “Some of the best science in nudge theory comes from research around smoking and gambling,” says WSP’s Thomas. Casinos often use a maze layout that makes punters weave between slot machines, so that they stay longer and part with more cash. Some researchers, including the father of nudge theory, Cass Sunstein, have become so concerned about the ethical dimensions that they have called for a “Bill of Rights for nudging”.

The argument for boundaries of some form is perhaps even greater in light of the fact that nudges in one environment can lead people to form broader habits, as Wendy Wood, psychology professor at the University of Southern California and author of the book Good Habits, Bad Habits, points out: “It’s sort of like smoking. You might smoke with friends having a drink at someone’s house and then over time you start to smoke in more and more places as the habit becomes broader and generalizes to new contexts.” This can of course produce healthy behaviours too — Wood herself ditched her car and moved next to a train station after returning to California from a car-free year in Paris — but it is a marker of just how far a nudger’s influence can stretch.

The ethical question certainly places huge responsibility on the designer “in terms of identifying or deciding what good looks like”, says Gardiner. However, the flipside is that there is no such thing as impartial design: “It’s not like you can design something in a neutral manner that will not influence behaviour. Every decision you make will in some way influence people and that’s a reason to think carefully rather than dismiss it.”

And for anyone worried that nudging is the start of a slippery slope towards mass mind control, we can take some comfort from research that highlights the enduring complexity of humans. One University of Virginia study, again focused on default lighting settings, found that responses varied according to personality type, with participants who were higher on the neuroticism scale more likely than others to start fiddling with the switches. “We’ve got a good understanding that some of these comfort settings have to be usercentred,” says Heydarian, who led the study. “But based on what we saw, nudges have to be user-centred too.”

In other words: we are not robots. Or if we are, we are complex and varied models.