Noise is slowing us down! Especially when it comes to professional opinions. The variability in what each of us believes is the correct answer, and the time spent in discussing this, is often time wasted. Not to mention giving less than optimal results.
There’s an interesting book on this topic by Daniel Kahneman and Oliver Sibony. Noise, they say, is an error of judgement. It leads to inconsistency in our decision-making. It’s the reason the same question can have many different answers. They point to studies showing huge variation in court sentencing depending on which judge is presiding. This is an example of noise. There are plenty of equivalent examples in medicine, engineering and other professional services. Kahneman and Sibony say noise is part of being human, but it leads to huge costs.
One of the solutions they propose is the use of algorithms, described as the ultimate in noise-free decision-making. While there are limitations to the use of algorithms, there is certainly a place for them.
Reducing noise for the questions we have already agreed on
We expect our legal system to provide a just outcome regardless of the judge that oversees our case. We expect the right medical opinion for our health regardless of which doctor we see. There are many situations in professional services, such as these, where experts should be agreeing on the right answer. In seismic engineering, we work to a building code. We should come to the same conclusions for earthquake loading for example, because the standards set out a gold standard approach. However, the code is highly complex and requires a lot of engineering judgement. As a result, different people will come to different conclusions.
What if we could make the interpretation code easier? Open source software built on algorithms could be very helpful in situations where we use the same foundational calculations. Take for example, earthquake loads. If we use a common library – such as that provided by open source software, all engineers would start with the same base. There would be a commonality of source. Not only would the outcome be more standardised, and therefore closer to the ‘right answer’, we’d save so much time!
Having a shared basis for standards doesn’t limit your creativity or what you design – it just means you start with the same loads.
Peer review and the regulatory environment
In Noise, Kahneman and Sibony recommend that judgments should, where possible, make use of the “wisdom of the crowd” in order to reduce noise. If you ask enough people a question, you will get a better answer than if you ask one. As long as each of these people answers independently. Professional services make use of peer review to provide ‘wisdom of a crowd’. However, it is not practical to include large numbers in peer review.
Even with just one person peer reviewing another engineer’s work, it can be cumbersome. You need to review the engineer’s report, identify which assumptions they’ve made and decide whether they are correct. The assumptions will vary on a project and project basis. Each new project requires a fresh review, as the engineer considers each factor again.
A shared library of workable standards would make peer review and regulatory review much easier. With a shared library, you know the assumptions that are made up front. When it comes to the building code, this is information that we should all be agreeing on anyway. By standardising the way we use and interpret the code, it would save engineers a huge amount of time and expense in repeating calculations.
Furthermore, by sharing the code that sits behind this library, as an open source code, it can be reviewed by an unlimited number of people. This allows for ‘wisdom of a crowd’ in a far more practical sense. Because the code is shared, any changes can also be shared, meaning new advances or learnings are to the benefit of all users. If the building code changes, for example, you only need to change one piece of code and everyone is up to date.
The same stands from a regulatory point of view. If a shared library has been reviewed and approved, it doesn’t need to be checked again for every project. If council for example has a set of approved libraries, and these have been used in the engineering equations, it would reduce the workload considerably.
Where to next?
We’ve spent almost a decade looking at the automation of engineering in construction. Last month BVT shared our open source code for the engineering standards for 1170 and 1170.5. We’d love for other engineers to take advantage of this work. We believe that by sharing our progress, we can help our industry become better faster. You can find our libraries of code here.
I’m happy to run a workshop/webinar on how to access and use these codes if anyone is interested. If you would like to try them out, and you need a hand to get started, please get in touch.