Checks and balances

November 29, 2010 — 14 Comments

fatman1

In a conversation over breakfast this morning an architect friend was relaying a conversation she in turn had had the previous evening with someone influential in the world of economics. He was wondering why in architecture and design there are relatively few examples of massive and systemic failures. Buildings rarely fall down and products rarely fail to work basically as intended. He compared this to the financial world where ‘new designs’ often fail systemically with significant unforeseen downsides. Of course there are failures in design and the occasional product recall to prove it but in general his observation is accurate. The design process as applied to products, buildings and other tangible things is quite reliable.

This led me to wonder about two things. Firstly, why is the design of tangible things so reliable and secondly are there lessons from attempts to design in the abstract world of economics that may be useful to all design thinkers?

My response to the first question would be to wonder about the inherent transparency of tangible things. It seems to me quite a bit easier to inspect and test designs that are tangible versus designs that are intangible, and often extremely complex. Even in their nascent form as drawings many people can assess the viability of a new product idea during the design process. It is also relatively easy, and getting easier all the time, to create prototypes that can be tested under the laws of physics, markets or any other environment one might be interested in. The design of physical things also has a long tradition and hence many past examples to draw on when imagining potential failure modes.

So, what happens when we leave the world of tangibility and enter the abstract. Examining software design might be a good first stop. Here we see many of the same characteristics as in the tangible world (fast prototyping, iteration, reasonable transparency) that help mitigate against catastrophic failure but we also see some of the characteristics of the abstract and complex that signal potential danger. The network effects of modern software mean the ultimate impacts of our design can be hard to understand and imagine in advance. The relative ease of iteration and innovation makes change constant and the impact of that change hard to assess. As we move even further from the comfort of tangibility toward financial systems, social networks, health care systems and the like the predictability, evaluation and transparency of new designs become even more of a problem and the risks of dramatic failure increase.

As far as I can see there is little inherent in the design process that protects design thinkers from these same failures if we choose to tackle abstract, intangible questions such as services, systems and networks. Instead we might imagine how to apply the same rigor and discipline to the design process that has emerged from hundreds of years of practice in the tangible world. We might concentrate on how to make the process of the design of the intangible as transparent and open to observation as the design of the tangible. We might develop prototyping environments that allow us to learn through failure without catastrophic implications. We might accept that we need better mechanisms for criticism and feedback so that we begin to establish a body of knowledge about what works, and what does not, in the design of these things that don’t go ‘thud’ when we drop them.

Tim Brown

Posts

14 responses to Checks and balances

  1. Tim,

    A really good book that addresses design (or lack there of) of public policy is “If We Can Put a Man on the Moon: Getting Big Done in Government” by Eggers and O’Leary. They identify six basic problems in public policy design and implementation. One that is of particular interest to design thinkers is what Eggers and O’Leary call the “Design-Free Design Trap”.

    In short, I would say the problem with public policy failure is a systematic failure. But Eggers and O’Leary are extremely on point that the failure of policy and legislative development is that policy and legislation are rarely designed for effective implementation.

    I personally think that design thinking has a lot of application in the world of public policy and government.

    Thank you for this post, your book “Change by Design”, and your work a IDEO. It has helped me a great deal.

    Thanks,
    Brian

  2. Systemic failure with regard to tangible products is found in the effects of products. The (in-)famous story of the IDEO-designed toothbrush is the first one that pops in, but the new plastic-bag “continent” growing in the Pacific Gyre is a rather massive example. These failures also tend to be quite long term, and involve “new” materials and technologies we generally don’t figure the long-term effects of. Designers are more aware of this now, but the same will likely happen when we start designing at nanoscales with, again, new materials and technologies.

    Thanks for your writing, Tim, and your good design thinking!

    Looking forward,
    David

  3. Interesting…if I may add my observations to this line of thought…

    A business system (either at an enterprise level or the larger ecosystem it operates within) is like a money machine… all too often, the problem is that the people who really understand the money (intangible) don’t fully understand the “machine” (the intricacies of how tangible products and services are created), and the people who operate and understand the tangible machinery on a day-to-day basis don’t really understand the (intangible) money – or value creation – part.

    (To date, this great divide has been largely institutionalized between the thinkers and doers…but that is a topic for another discussion.)

    It is relatively easier to understand the money machine at an enterprise level (a single financial reporting unit) – but it is much more challenging to understand it at the ecosystem level. (Is Cisco a $35 billion company – or really a $150+ billion ecosystem?)

    A few people understand the money side of the ecosystem really well. Of those, some use that knowledge for their own gains, and perhaps to the detriment of the ecosystem (Goldman Sachs?). Some thought they understood it but didn’t (AIG credit default swaps group?) Some who did not understand it used models that turned out to be incomplete (almost all the banks). Some who understood it really well bought into the hype and theory when they should have known better (Alan Greenspan)…

    When I think of prototyping in the area of intangibles…I see a lot of potential in the area of simulation because its impractical to prototype these things in a tangible way.

    I like what you say about being open to observation, criticism, and feedback. “We know what we know. We know what we don’t know. We don’t know what we don’t know.” In a world of increasing complexity, I feel like we don’t have enough respect and appreciation for the last part of that phrase. Until then, expect more “thuds”…

  4. Tim,

    I agree that this is a significant and increasing problem for designers, however I don’t think the solution is as simple as the one you propose in your last paragraph. In fact I suspect that approach is the one currently being unsuccessfully used by economists.

    To borrow terminology from The Black Swan by Nassim Nicholas Taleb: tangible products are normally from Mediocristan and complex interconnected systems are from Extremistan.

    In Mediocristan there is randomness and uncertainty, but it is relatively small and manageable, variation tends to fall in the normal distribution that designers are comfortable dealing with. Extremistan on the other hand is ruled by low probability but high impact events that tend to fall in power-law or fractal distributions which designers – and people in general – are much less comfortable dealing with.

    I don’t think that it is possible to increase the transparency of the design process in Extremistan to the same level as Mediocristan. To do that you would need to be able to fully and accurately model incredibly complex systems with unpredictable inputs. Furthermore these models could only be validated against historical data – the method economists use and look how that turned out – just because something has not happened in the past does not mean it won’t happen in the future. Even assuming that you could accurately model the system and generate unpredictable (unforeseen based on historic data) results, how do you set an acceptable threshold for the probability of total collapse? Just because the probability of total collapse is 1 in 1,000,000,000 doesn’t mean it won’t happen after 1,000 events.

    Just now as a design thinker, all you can do is: determine if you are in Extremistan or Mediocristan, try to design appropriately and hope for the best.

    Thanks,

    Steven

  5. Hi Tim, great questions and an interesting post. The one element that I think hasn’t been fully touched on is the role of people and human behavior.

    While I’m certainly no expert in design, I would imagine that when an architect is designing a building, for instance, he or she understands that there really is only one desired outcome – that the building stands up straight and stays up (granted this is oversimplifying but it’s basically true). There is really a “right” and a “wrong” here – the building either stands up or falls down, an idea that guides the architect throughout the design process.

    When you’re designing systems, networks or other intangibles (especially when it comes to social innovation), I think there is often more grey area surrounding the desired outcome. Personal motivations and drivers (ie: can I make money in this deal, what’s in it for me, what do I want to see happen versus what do others want to see happen?) can come into play here, which colors the final outcome and ultimately determines whether a project is successful.

    In my opinion when we design for the intangible, there’s more room for human psychology to play a role – which can be both a good and a bad thing. The trick is to figure out how to 1) leverage peoples’ passion, commitment, experience and interest to design solutions for these problems while 2) creating the checks and balances needed to ensure that the building ultimately “stays up”.

    Thanks,
    Ashley

  6. “Firstly, why is the design of tangible things so reliable and secondly are there lessons from attempts to design in the abstract world of economics that may be useful to all design thinkers?”

    Because the design of physical objects relies on models of materials behaviour and the abstract world of economics relies on models of human behaviour.

    Materials parameters, performance characteristics and tolerances are relatively easy. People are hard if not impossible.

  7. +1 for Lauren Elliott. The regulation of tangible things that kill people is not entirely fictional, unlike finance. In the UK an architect has personal liability for 10 years. And, yes, poor products don’t usually make it out of the door unless the company has tons of money from other sources (one in Redmond comes to mind).
    Plus product designers don’t have to listen to the nonsense of economists!
    Design of intangibles is potentially subject to systemic failure; let’s see how Facebook works out.

  8. Thanks for this great post.

    I have a few scattered observations.

    Is it possible to make the intangible more tangible in the design process?
    software design would a great example, turning interface into object
    (screens / post-it notes)

    From service design’s perspective, it seems this is how we approach
    the intangibility. By prototyping all the touchpoints and have people “walk through” them to simulate the experiences.

    It also seems we need to have a new set of vocab for system ergonomics to
    evaluate the structure and stability of any system. It reminds me Tim’s previous post about the visual of networks, an example of transforming the intangible to more tangible.

    Human behavior cannot be fully predicted. People will use a chair as a table, a table as a chair. Does this means the design fails? It does when someone sit on the table breaks it… designer should welcome these type of “misuse” and design for them, not avoid them. But how?

  9. Thought-provoking post, Tim.

    A few additional thoughts…

    I think the tangible vs. intangible dichotomy is an interesting one, but it clearly breaks down. The great challenge, as you imply, is when one is dealing with systems that have emergent properties. When you’re dealing with a single product or service or even abstract entity, analysis and a design process definitely make sense, and can lead to a result that’s not far from what is desired. When the system gets more complex, meaning lots of interactions between lots of entities, with unpredictable outcomes, then all the design thinking and process in the world will only get you so far.

    In my opinion, when it comes to complex systems with emergent properties, it’s all about testing and refinement. Iterative development and change management are key, and systems that are designed without an explicit way to accommodate these things are risky at best, disastrous at worst.

  10. This is an article about the most important type of design thinker, the one with the abilit to say no. I hope you enjoy it. Thanks. http://thinkbig-lab.com/blog/do-nothing-with-urgency

  11. I do agree that there are some things to be learned about checks and balances in the field of structures. On the flip side, I think there could be a lot to be learned about the innovations that civil engineering can learn from software.

    On another note, here are my thoughts on what I know about the different disciplines:

    As a former civil engineer, I know that structures undergo a LOT of checking, double checking, and triple checking. Each line of calculations and drawings get checked regularly. Also, a lot of it includes a fairly strict algorithms and regulations; there aren’t many different ways to calculate the same structure. (These algorithms also include a large factor of safety.) As others have noted, civil engineers are also very aware that a failure can cause a lost of license, lost of life, and even imprisonment. (The sad truth is, though, that structures still fail more often than they should.)

    As a current software engineer, I know that although there is a quality assurance process, there’s still a lot of room for mistakes. Because of resource constraints, code is rarely checked line by line. There’s also many different ways to do one thing. A failure in software is typical, and can easily be fixed with patches or version upgrade. Usually, there’s little to no chance for lost of life or license when you screw up in code. Deadlines trump correctness in software projects.

    Also, I’m nowhere close to being an expert in finances but from what I do know there are so many more unforeseen factors in economics. Civil engineering doesn’t have to factor in the psychology of consumers and international relations.

    Hopefully, the different perspectives I have from my experience in both civil and software engineering will provide more food for thought about this topic.

  12. It occurs to me that economic system designs are possibly being implemented by people who are most interested in the ability to exploit the system over its lifecycle. If this is the case, they may create intentional complexity to obscure their behavior and prolong the life of the system. These systems collapse under the weight of either the exceptional greed of a few or when the means to exploit the system becomes known to many. The collapse of banks caused by as few as one rouge derivatives trader is an example of the former, and the global economic crisis caused by “fraudulent” mortgage lending and the broad securitization of those sure-to-fail assets is an example of the later.

  13. If the complex and fault-prone structure of credit default swaps and subprime mortgages had been mapped and made visible, including “touchpoints” as LEO TC refers to, the faulty design would have been visible and available to humans making a decision to invest. If anyone saw a building being constructed upon sand, there would certainly be loud questioning of that likely point of failure. The point of failure — one of them — of that economic collapse was in the so-called rating of the financial instruments and ridiculous incentives to lenders to give money away and sell the liability. But then again, maybe people would still make irrational decisions to take a chance with their money.

  14. Fascinating post and comments!
    I would argue that this is more of a psychological issue – that people tend to think less critically about an intangible process or product. I know that sounds over-simplified, but it is a common human flaw, is it not? The tendency to spend more loosely with a credit card than with cash, for example. If humans were to look at economic products as if they were tangible products, they would probably be designed more critically. Of course, designers are generally not economists/policy-makers, and vice versa. It seems to me that the kind of thinking required to function successfully in these two spheres has not yet been integrated in America.
    Wouldn’t it be interesting if we asked an economist to build a physical representation of a mortgage? What would it look like? How could we identify different aspects of the mortgage? It’s like explaining String Theory to people – you have to start with a line, and become a cube, etc. Visualization of mortgages, and also of the higher level economic structures that they rely on, might bring some valuable insights.
    ~Amy

Leave a Reply

*

* Copy This Password *

* Type Or Paste Password Here *

Text formatting is available via select HTML. <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>