I’ve been thinking about decision making for a while and wanted to point out how important the steps before the decision part is and how you need to use different decision making processes depending on what the problem and solution are.
This post isn’t a super clear set of thoughts, it’s more a collection of different related ones with an eye towards something more important brewing. But hopefully you get something out of it. I’ve been fairly vague in some areas but attempted to give real life examples where I can.
The Decision making Process
First you have to identify the problem. This requires good quality sense making. Daniel Schmactenberger and the people at the Consilience Project are doing some great work in this area especially with a focus on collective sense making and fighting against propaganda. Although there’s obviously more to it than just that. Often you’ll need specific skilled people who are experts in the field.
For example, when I’ve got an issue with my bike and I think it could be one of 3 different things I can take it to a bike mechanic and in a quick moment they often identify the problem. I can often do this with websites, depending on the problem.
Then there’s the solution making part. You need to work out what the potential solutions are before you can make a decision.
Whilst sense making requires parsing information and trying to uncover the truth of the problem, Solution making requires understanding what resources are available and using imagination and logic to develop solutions.
The solutions obviously differ in various ways. If it’s a small one off problem then usually a small quick junky solution can usually suffice but if the problem is something whereby you or many people might die, you’ll want to put a lot more effort, thought and skill into it.
A general example of this might be how to secure your furniture onto the trailer. If you take the time to actually strap the furniture with rope and use proper knots then you can drive further, faster and over more bumps without losing the furniture and having it possibly hit someone behind you as you are driving. It always seems to take so long to apply the ropes then undo them, but it’s worth it.Wearing a seat belt is a similar solution to a similar problem, but applied to people sitting inside of cars and it happens so regularly that we’ve made wearing a seat belt easy (you don’t have to install one each time) and we’ve got rules and regulations and standards around it. Even if not every country requires you to wear them all the time. Actually coming from Australia and even Europe where seat belt wearing is required to the Philippines where it’s not, I was shocked. However when you spend 3 hours travelling 12km in rush hour traffic (which lasts half a day), well, I can literally run faster than the traffic some times. So people just don’t travel particularly fast in Metro Manila. Maybe for 2 minutes you’ll get above 30km/hr. Where as in Australia you are regularly going 50-60km/hr and can fairly easily go 110km/hr, so wearing a seat belt is required because traffic accidents at those speeds are way more devastating.
When it comes to making decisions usually you’ll be weighing the different solutions and using different variables to do so. Standard ones include Time, Financial Cost, Environmental cost, mental and physical effort, upfront capital versus long term costs, or many other factors.Here the processes used will differ according to the type of problem and possible solutions.
When I’m proposing a solution to a programming issue and people want an estimate they want to know how many hours work it’ll be (and what my working rate is).
To help make such estimates I use the ECU matrix of Effort, Complexity and Uncertainty.
Some things might be a lot of effort, but easy to do. Maybe it’s a lot of copy/pasting and tweaking, but not something that’s hard. Some things might have a lot of complexity, which usually means it’s got some hard optimisation problem or is some core piece with lots of connected / moving parts associated with it and needs extra care with lots of testing both it and the associated parts.
Uncertainty usually means we don’t know the full extent of the work involved. We often won’t know how long it’ll be until we start it, or maybe it’ll be an iterative process and we won’t know how many iterations it’ll take to get right.
Something that’s easy might be adding a new field to a form. The systems I use (Yii2 Framework) make this easy.
One of the worst projects I worked on turned out to be importing data from CSV files.
It was a lot of effort because there was lots of data 100k+ data points. There was a lot of complexity because there was lot of parts that related to each other. Being Sports fixtures (games) it required the teams, field locations and dates to match.
Initially I assumed very little uncertainty because I assumed it was quality data. However it turned out there was a lot of noise in the data as it had been created by hand and we couldn’t match different sections easily as the primary key often had extra cruft to it.
So myself and another developer would iteratively work on getting the data input and would see that we had 60% of the data import without issues in the initial attempt. We asked how good quality the data needed to be and was told 100%. We baulked and tried to explain the issue but they just asked how long it would take. We told the project manager it would be 4 days to do an iteration, fixing what data we could and ended up importing 78% of the data properly. Then after another few days of intense work we got it to 83% this continued for over two weeks. The thing is, the data was for Sports matches and events that were happening during that 2 week period. By the time we achieved the goal of nearly 100% data integrity it was useless. The project manager at the time didn’t want to understand the nature of the problem nor solution. It turned out we could’ve just focused on 1/7th of the data and been done in the first 4 days and fix up any issues by hand, besides, few people were going to be using it. To me the issue was a project management issue and the wrong problem / solution.
But, I was also partly to blame. I had been asked a month before how I’d like the data and if an Excel spreadsheet was fine. I said yes. but hadn’t known at the time that people would be manually copy/pasting data from websites into various spreadsheets and in a haphazard way, sometimes manually typing it it, with typos, or copying in extra things that aren’t needed. If such a scenario happens again I’d instead setup a webpage where they have to enter the data in and be able to properly link the data correctly.
If the problem is man power and the solution is a set of people to hire, then you’ll want to evaluate those people based on psychometric profiling, how well they work in a team, their skill set and if they’ll find the work engaging and get them into the flow. If they are too unskilled they won’t be able to thrive in the position. But if they are too skilled they’ll likely be bored.
If the problem is predicting what future energy usage levels will be then Philip Tetlock’s book Superforecasting explains well how having a weighted sum of people who are good at forecasting is a very great model for making such predictions.
Power Plant Example
If the problem then turns out to be a lack of energy generation the decision could then be which type of power plant to build then the solution making process should look at what resources are available and make useful suggestions. An example might be that there’s a small coal deposit 400km away in the remote countryside but the city needing power gets enough sun, wind and tidal power that those are all great options. So the decision making process would weigh both the short and long term costs both financially, but also to the community and the environment.Solar Thermal (CST) which includes energy storage that can be used overnight, plus Solar Panels on roofs with wind turbines on the nearby hills would work well together. However there could be a rare species of fish that nest where the wave / tidal power plants would go. Coal shouldn’t be a last consideration given the CO2 output, coal dust and black smoke effects on the township, plus the long distances needed to get the coal there. But also, the cost of renewables goes down over time, whilst the cost of coal and other fossil fuels increases over time.Unfortunately we know that most such decisions are made by politicians not engineers and the politicians are often swayed by fossil fuel lobbying or other indirect financial incentives, such as a shorter time to pay back the initial build costs even if the life time costs of the fossil fuel related project are going to be an order of magnitude higher.Going back 30 years and the Solar and Wind options weren’t viable solutions. It took a lot of effort to develop them to be possibilities. Wave and Tidal power are only just getting off the ground in terms of potential, but will hopefully have strong pickup soon.Returning to the issue, I pulled a fast one by assuming we’d need more energy generation. It could be that there’s other solutions to the problem, like lowering demand by increasing efficiency. This could be done by installing better insulation to reduce air conditioning costs or switching to using Earthship style housing which is one of the most energy efficient forms of housing. Changing lighting or heating / cooling are usually the main drivers of energy efficiency but cooking, transportation and other things also affect the power grid.
So the type of problem and the possible solutions require different metrics and usually different decision making processes.
When most people think about decision making processes they likely think about democracy versus authoritarian styles. There’s things like consensus decision making, meritocracy, or weighted averages and a whole variety of voting mechanisms.
These are especially useful tools when making decisions which affect a lot of people. Most people are familiar with voting for a particular government party, you usually vote for people or a group instead of for general policies, or specific implementations.
There is of course an issue with voting for people to represent you when we currently live in an economic system which expects people to work in their own self interest. Couple that with sociopaths who are good at manipulation and are power hungry and it become obvious that there’s some major work to be done and Democracy has major issues.
Most people will be familiar with voting for a government party that might be pro-renewables or maybe one that’s pro-tax reduction (especially for the rich). Rarely is there a way to vote for specific policies where you can vote for both, possibly because doing so would mean you’d find that road repairs and hospitals or other aspects of society would end up being neglected.
The main difference between a full open democracy versus a dictatorship is in the number of people who need to be appeased in order for the person making the decisions to keep in power. The Dictators handbook explains well how the quality of the decisions being made and which groups are given preferences are often easily detectable based on the elections. These elections can scale in size from having a company board to voting for the president of a local club to the a local mayor and up to country elected officials. If only 1% of the constituents usually vote to get someone to be elected then that 1% are who will be focused on, often at the expense of the 99%. You can see this type of politics play out with the crazy types of subsidies that occur and complex tax credits. Maybe certain types of farmers, bankers and film makers are given extra concessions whilst plumbers likely aren’t given such concessions, based on the way people in those professions generally vote. The smaller the cohort of people voting, the more corrupt the government is also likely to be. Again, sociopaths or full violent psychopaths are a big issue here. But there’s other major systemic issues within most governments whereby the politicians in power are only given a certain amount of wiggle room anyway.
For those who don’t have time to read the Dictators handbook I recommend the CGP Grey video the Rules for Rulers.
Scaling up Consensus?
My main thinking regarding decision making processes is when thinking about a Post-Scarcity Society and how to do resource management and deal with cultural issues.
Many sustainable communities use a consensus decision making process whereby everyone is a part of the process and a majority vote is required, with some people able to put up blockers indicating they’ll leave or not participate if a certain solution is implemented.
Such decision making is very slow and painful. I’ve been a part of it as part of Occupy Adelaide and bringing everyone up to speed with the problem and the suggested solutions can be very time consuming.
Most of the time we delegate decisions to a culture code or set of basic principles. For example CORENA the non-profit in Australia which provides interest free loans is focused on reducing CO2 emissions. This is primarily through getting solar / renewables installed. But it could also be by helping have the lights replaced with LED or thermal insulation installed to reduce air conditioning bills. With such a target aim it’s usually easy to understand what is and isn’t within the purview of the org.
Within society we have basic morals. Don’t kill, don’t steal and the like which are considered important foundations towards how to live your life. We’ve then got laws and rules which add more nuance or behavioural modification attempts towards dealing with the complexity of society. You aren’t likely to go to jail because you took a pen home from work, but you might for doing drugs or even worse, dealing drugs.
Sometimes there’s other options, it’s just we haven’t adopted them, or consider the costs not worth it. A great example is speed limits and traffic lights. These are more soft solutions and considered mere suggestion by some people and for others, they aren’t in a physical and mental state of mind to be able to abide by them (e.g drunk drivers). This often leads to car accidents.
But Jacque Fresco (RIP) had a great idea about having poles come out of the ground or the road come up at school crossings so that cars would hit the poles not the pedestrians. Those are more hard solutions than simple traffic lights, although have a whole different set of issues.
I’m also reminded of a story I heard regarding the Earth Song community in New Zealand and their issue around cats. The sustainable community was designed and built over the course of a few years with relatively little conflict. They had a good architect who created a great design and people brought into it and got it built. But once people started living there an issue came up. Cats.
There’s generally dog people and cat people and this community had both. However cats have a behaviour of killing birds, even if those cats have been domesticated and are given food regularly. There’s a great theory I heard that cats are actually communal animals and consider you as part of their community. They bring you birds and mice they’ve hunted because that’s their way of both pulling their own weight and also thanking you.
The unfortunate thing with a community of 80+ people is there’s just so many cats that there’s a lot of dead birds. For people who care about the natural environment and put a lot of effort into growing trees and plants to attract local wildlife, it was an issue.
Over the course of 5 years they thought about and experimented with a variety of different options. Firstly, getting rid of all the cats. This doesn’t make the cat lovers very happy, having to give away their fur babies, but it also has an unintended side effect. Wild cats instead take up the territory and those hunt for food not just occasionally for sport. So they kill more birds. It was determined that the area could be serviced by 2-3 domesticated cats. So instead they let some cat lovers become cat owners, however that makes the ones without cats not very happy. Eventually an optimal solution was found by having the cats be community pets. They are looked after by the community, this means the costs of desexing and other measures are born by the community and so there are no specific cat owners.
Note that my memory on some of the exact aspects of the cat story are likely to be wrong in detail but hopefully correct in essence.
Whilst consensus decision making is a great system and means there’s a large number of people doing sense making, solution making and decision making and everyone feels they’ve had their input, it unfortunately doesn’t scale beyond 80 people very well, let alone the 150 people of dunbar’s number. So how do we scale quality decision making systems to tens of thousands of people?
What we need is a variety of decision making systems, coupled with good sense making and solution making and to know which techniques we use when.
Actually that’s mostly my way of saying that I’m not sure. It’s going to take a lot of experimentation.
The idea for this post came when reading How to Mislead with Facts https://consilienceproject.org/how-to-mislead-with-facts/
The Consilience Project was created by Daniel Schmactenberger and does some great work around sense making because they see bad collective sense making as an epistemic based existential threat.
It’s an article about how people can use verified facts to put forward a narrative, a form of propaganda which strays far from the truth because there’s a lot of nuance or context missing.
An example that was brought to my attention is the site
swprs.org/covid19-facts It’s pitched as “Fully referenced facts about covid-19, provided by experts in the field, to help our readers make a realistic risk assessment” however when reading it myself and many others got angry at the way the “facts” were being used to bash Covid, lockdowns and the vaccines, because they used very cherry picked points and didn’t provide the full context. It makes the problems with Covid tests and lockdowns seem massive and unjust whilst vaccines are deemed unreliable or more dangerous than the disease. But that’s because of the biased perspective not showing any of the positive effects. Covid19 has a mortality rate more around 1.5% globally whilst the vaccines are on the order of 0.0000369% by what I could find.
When I challenged people who were sharing this bias I was met with a strong emotional resistance. They were attempting very hard to mislead with facts.
Then I read this article.
Want to Be More Innovative and Creative? Science Says Avoid the Dreaded Einstellung Effect
This is an interesting article that’s somewhat related. It’s about how you can easily get stuck thinking that the first answer you have is the correct one, but it might be less than optimal.
It shows that if you are stuck and keep thinking about the one option then pretend it’s not an option and move on to see what is possible.
These thinking tools can hopefully help us come up with better solutions to problems and hopefully then have better options to make decisions about.
Hopefully in the future I’ll get the chance to help create some useful decision making systems.