Who Wants to Pay for a Building Twice?
A Provisional List of Challenges for Practicing Human-Centered Architecture
Over drinks with a group from MIT doctoral students late one evening, an accomplished computer scientist asked me why architects do not prototype buildings. My glib answer was that buildings are among the most expensive things we buy as humans, and many people don’t want to pay for them twice. Most couldn’t, even if they wanted to.
Since the initial question, some ten years ago, about prototyping buildings I’ve been thinking about the limits of human centered design (HCD) and the need to create design methods that embrace human fragility without giving up the need to take on questions larger than any one person. This exploration started at Helsinki Design Lab, where we found that HCD alone could not tackle problems that are both long term and systemic in nature, and it continues today at Dash Marshall, in both our Civic Futures work and the buildings and spaces we design.
We start architectural projects with an interview protocol that’s derived from design ethnography I learned while at IDEO, we use paper/tape prototypes when possible, and we often suggest to our clients that they save 30% of the budget for post-occupancy rejiggering of their space, but we’ve had no takers on that one (thus far). These are paltry examples compared to the robust HCD process that is common in industrial, digital, and pretty much every other sector of the design world.
If the goal is to make things that people want (buildings, vegetable peelers, energy grids, stationery, etc), then the designer can either guess wildly at what people want and hope to be correct, or they can narrow the field of options by using some limiting factor. If you’re Frank Gehry you might limit your field of exploration to buildings-that-look-like-scrunched-up-paper-and-have-expressive-tectonics because you’re famous and you have a career’s worth of evidence that people want that kind of building. If you don’t happen to have a global personal brand, you need another homing beacon that will pull you (and the output of your design process) towards the zone of human relevance.
For people who design and engineer digital products it would be heresy to suggest that you can build something once and get it even 80% correct, let alone nail 100% of the issues. Human centered design was adopted so voraciously in the digital design community because it provides a helpful scaffold to get from “this idea could be interesting” to “this idea works and people are enjoying it.” Architects could benefit from at least two outcomes of the HCD process.
First, the architect needs evidence that can be used to inform and defend their design decisions. Second, they some way to tests those design decisions and either invalidate them cheaply and quickly, or generate confidence that they were correct and should be pursued further. Collected below are open questions relating to these two topics in the context of architectural practice.
Evidence
Evidence comes in many forms, including facts and opinions. Facts, like sun angles, may be calculated. Opinions, such as the widely held architectural belief that an atrium creates social connectivity, may come from an individual (very weak), may be validated by architectural precedent (somewhat weak), or may be the result of a structured insight-gathering process (generally stronger). Though many architects do not think of the stimuli they are responding to during the design process as “evidence,” we use that term because evidence is exactly what non-architects expect to be used to justify a decision.
When your client or the community review board asks, “why did you design it like this?,” it’s up to you as the architect to choose the basis of evidence that you use to explain (or perhaps defend) the decision. In this setting, your own professional opinion may get you a little leeway, but many well-intentioned designs have fallen on the sword of unsubstantiated claims. It seems too common that architects substitute charisma for evidence because there is often very little or no extra-disciplinary evidence to be had. In this case, a reliance on charisma is not a personal failing but a disciplinary one. A stronger basis of evidence within architecture could also help deflate the ‘star’ culture that still seems to dominate.
With regards to HCD in architecture, the question is how we can use HCD to generate a convincing basis of evidence for architectural decisions.
How to gather feedback from people who do not exist yet? HCD relies on insights gathered from living people, but buildings are designed for a 30+ lifespan, meaning that some of the future users have not been born yet. What needs will they have and how can they be anticipated?
How to anticipate cultural change even if it cannot be predicted? HCD results in products that are finely tuned to humans, but humans change with their culture, especially over the long lifespan of buildings. There’s some answer in the concept of shearing layers (a notion put forth by an architect that has, somewhat curiously, become more popular in the digital design community), but this is a methodological problem for architects to reckon with. How do we balance present known against future unknown?
Who pays for user research? In HCD, the client pays for research because it has been proven within industry that human-centered products do well in the market. This is now widely understood, but it took a lot of investment by private design studios and public advocacy groups like the UK’s Design Council to get here. Without the same basis of evidence for positive outcomes from human-centered buildings, there’s a chicken-and-egg problem. Who will invest first? (I think it’s an opportunity for universities).
What’s the right balance between direct and indirect users? HCD offers tools for understanding the “user” as a diversity of different archetypes, but buildings add a further layer to this, which is the person who never enters a building but is impacted by that building’s presence in their city nonetheless. How would the user research protocol look for a building such as 432 Park Ave? It’s large enough to be seen from multiple states! There’s spade work to be done figuring out how to balance the interests—and often competing interests—of direct users who will literally use a building and those who will not, but are affected by it nonetheless. As a side note, this phenomena is now being experienced by Facebook and other large digital platforms in the wake of the Russian influence operations during the 2016 election cycle. Whether you use Facebook or not, their design decisions have an impact on your life. Designing for even the people who are incidental users of your work is a very hard thing to do, and so far they Facebook to be failing.
How do we make it cheap and easy to quantitatively analyze the effect that buildings have on humans? I am not a fan of quantifying everything, but some amount of quant evidence goes a long way, particularly when the client is used to using quantifiable evidence to make business decisions (there’s a separate discussion to be had about the validity of that evidence). To go back to the atrium as an example, what’s the cheapest way to measure the efficacy of an atrium at producing more social interactions? How much would it cost to study a reference atrium and, say, a small lobby? Groups like Humanyze are building some tools for this, but it would be great to see many more options in this space. Situ Studio, for instance, are hacking their own.
Confidence
Unfortunately, with all the best research in the world it’s still possible—even easy—to make the wrong decision. In the tech world “lean” approaches to development rapidly become the norm about ten years ago. Architects who look at the lean development process would find it very familiar: it’s an iterative design process. The difference is that when building a digital product you can repeatedly build and test versions of the actual item. With buildings it’s exceedingly rare to be able to build and test (though I did hear a story about Foster building a 1:1 scale house in a hanger before iterating and then building on site). Instead we rely on models, renderings, and simulations to test the design decisions that we make along the way.
Prototyping through different means is great because it gives the designer a tool to check their assumptions, test design hypotheses, and ultimately generate even more evidence that the design proposal is valid and beneficial. With a good prototype, both architect and client can move forward with confidence, which is no small feat. So when it comes to HCD in architecture, the question is how to expand the tools we use to evaluate and iterate design proposals, ideally in a way that puts architect and non-architect on equal footing.
How do we test large scale experiences? HCD is being used right now to test spatial experiences such a redesigning hotel lobbies and travel journeys, but how do we test something like the redesign of a park? Urbanists with their recent focus on tactical urbanism have provided one way to think about this. You prototype a park by doing something small and scrappy, often without much interest in design quality (at least in the ways it would be understood by capital-A architects). NYC’s experiments with street plazas under Mayor Bloomberg are another example. Dig in to these stories and what you find is that by and large they’re championed by people outside of the architectural disciplines. How could architects build a culture of rapid experimentation at 1:1 scale within the discipline? I’m optimistic about VR but the technology still needs to mature substantially.
How do you test conditions and experiences that have not existed before and are too large to prototype? Buildings are more like an operating system than an app, which is a problem because it’s easier to test an OS than an app. An OS provides the platform for countless activities and actions, intention and unintentional. Some of those activities (say, sleeping) are common and the architect can use analogous examples to understand important factors that influence them. Great buildings, on the other hand, enable activities and moods that have not been experienced yet. The Shed in NYC presents such an example. How does the architect anticipate — and therefore test — the conditions created by the Shed’s retractable roof? No one knows if it will work, how it will work, or what it will inspire.
How do you make decisions about such a thing with any degree of confidence? The answer may be that you don’t, and that’s fine too. Sometimes the rhetoric of HCD seems to presents the possibility of a design process that is free of risk, but a world without risk is a boring one indeed. At the same time, a world without informed risk is foolish, so the question remains, how do we acknowledge the risks inherent in our architectural decisions and still act confidently?
Who pays for full scale prototypes? How might clients be convinced of the value of building 1:1 scale prototypes? The tab includes hard costs of building prototype, soft costs of producing the prototype and iterating the core design based on findings, not to mention the additional time that this introduces to the process overall. There seems to be opportunity here for institutional players who own buildings longterm and therefore have an incentive to design buildings that function well from environmental, social, and cultural points of view. Again I would look to Universities as one possibility for a radical rethinking of the design and build procurement process. Large companies (Google, Facebook, Ford, GE, etc.) could be another option, but they have little incentive to be open with the findings of such a process. Same goes for developers who build and hold, such as the Durst Organization.
Who is responsible when evidence and confidence are not enough, and the design decisions approved and then built are still unsatisfactory? Humans will inevitably make mistakes and that means the outcomes that our buildings produce or influence will not always be beneficial in the ways we would like them to be. Under the status-quo a wide variety of stakeholders pay the price for this. Children suffer the result of a building that performs poorly and uses too much energy or sources it from dirty plants. Workers suffer the result of buildings rented or built by corporate leadership without genuine concern for the human condition. Neighbors suffer when market forces conspire to create buildings that offer little or no encouragement for positive social dynamics. We all suffer when buildings are used for speculation instead of contributing to the habitation, work, and recreation of the city. The biggest question that I have towards a truly human centered architecture is that of accountability.
Who is responsible when a building performs poorly? How long do we wait to find out? How do we trace performance back to design, construction, and operations or some combination thereof? How is it remedied? Who pays for the remedy? What happens when my definition of “performs poorly” is different from yours? American architects stepped away from questions like these during the course of the 20th century, perhaps due to scary questions like legal liability, so it’s no surprise that much of the current output by architects is still perceived as not responding directly enough to human needs. It’s not just architects, of course. Developers and large-scale builders have also structured the delivery of new buildings in such a way that accountability is a hot potato. Failures are often better unacknowledged than accepted and dealt with. Truly human-centered architecture may not be possible without also rethinking the way that we finance, insure, permit, build, and operate buildings. For all parties involved, it’s hard to imagine being human-centered without a little skin in the game.
Comments, suggestions, and examples of human-centered design in architecture are most welcome.