First in a series on prediction. I hope.
Today is a simple if abstract discussion. The math of strategy. Now the majority of my peers at ASP specifically avoid the quantitative, and most strategists I’ve encountered are MBA’s, historians, or soldiers that are not that keen on math, and don’t have much use for it beyond the simple arithmetic of “our big number beats their little number.” (which isn’t even always true).
TED, FRY, Complexity, Systems and Strategy
By now, many of us are familiar with TED Talks (Technology, Entertainment, Design). Personally I was chagrined when I found out Ted.com was taken. But at least it’s being put to good use.
The TED talks are about “Ideas worth Sharing.” They started out very cool, and have gotten more commercialized over time, but more often than not, are fascinating; if typically very focused on technology (due to their silicon valley roots). There are also independently organized TEDX talks that really embody the “diamond in the rough” meme that started TED. It’s a great place to see well communicated ideas.
The other day on my lunch hour I happened upon a TEDx Talk from Britain that well, hit my sensibilities as a physicist, mathematician, and strategist squarely.
Dr Hannah Fry, according to her website, is a “Dr. of Fluid Dynamics, researcher of Complexity Theory and all round bad-ass.” If you take a look at her PhD Thesis, it’s 200 pages of graduate level mathematics that pretty well qualify her self description (As do her blogs posts on Python scripting).
OK, enough background. SO what?
Point is, Dr Fry did a really well developed TEDx Talk earlier this year that demonstrates in layman’s terms the power of modern mathematics. Specifically they created a mathematical model given available information and science that accurately recreates what happened in the 2011 London Riots, and appears to predict the behaviour of riots.
OK, I’m biased. I’m an astrophysicist that does quantitative analysis of business intelligence and project planning and execution for a living. My pile of books to read at home includes math texts on catastrophe theory, game theory, and complexity theory. But I confess that it was John Mauldin that first showed me the value of applying complexity theory to strategic thinking.
Here’s how Math works. Your classical physics and math is great for directly predicting simple things. This is the majority of everyday engineering and construction. We can accurately mathematically predict how well engineered machines and processes will work before we build them (e.g. machines, electronics, factories). We can use math to measure and control the consumption of measurable quantities, be is Gas in our cars, grams of protein for athletes, sales revenue, market share or money in our bank accounts. But you are limited by what you can measure, and it has to be “simple” mathematically speaking.
1 – There are many things you can’t measure.
2 – Some problems require you to measure too many things.
Now in the 20th century, science got around some of those problems by using statistics (i.e. simple pattern recognition). We can mathematically model and predict how huge systems work if their behavior is simple when treating them as a whole. This works for things like galaxies, stars, gases, explosions, fluids, water, etc. I don’t need to track every molecule of water in a pipe, just the sum of its parts and everything works out.
The problem is the middle ground. A complex system is something that has too many variables to measure and solve simply; and not consistent enough to treat as a single simple whole (i.e. a business, a war, the economy).
Now in the 21st century (with the help of computers), complexity theory is starting to get good. Researchers like Dr Fry are starting to create mathematical models that accurately reproduce and potentially predict the behavior of real life complex systems, using rationality and complex math. This is possible by basing the model by adapting existing models that work in analogous scientific systems.
For example assume:
How viruses spread = How information spreads
Shopping travel habits = Rioting Travel habits
Predator Prey movement = Rioter / Police movement
Adapt those three models into a single predictive model, and they actually do model and predict how fast, how long, and where riots will spread; based on input conditions and changing variables. The model is roughly as accurate as the input data, and does a good job of recreating riots that have happened. But even here, the precision and accuracy of the model is based on how many measurements you can put in. Historical data plus known science becomes a constraint on future prediction. That’s mathematical modeling 101, at least it was when I was in college.
Notice business kids, they use the known science of shopper behavior to accurately predict how rioters behave. Personally I had forgotten there was a science to shopper behavior. Point – there is tons of data and science out there, if you take the time to find it (there nothing new under the sun, odds are somebody else already knows the solution to your problem).
So at this point, ya’ll are saying, cool Ted. So some math geeks can do some math that models a riot. So what?
Ok kids. Any organization; a family, an army, an NGO, government or business is a system of interrelated parts that exchange information and resources to accomplish a variety of activities to accomplish organizational goals while also doing activities it needs to do to survive. E.G. A Family may go on vacation – focusing on logistics, travel and entertainment, but still have to deal with tactical needs of food, shelter, sleep, and family politics plus the long term requirements like bills, insurance, and a home.
A business may be focused on selling it’s product, with the normal cycle of R&D, Marketing, supply chain, manufacturing it’s product – but also has to have HR, Legal, Accounting, Facilities, IT, Taxes, and possibly PR and Lobbying to support the simple goal of building and selling a product.
Point is – all the functions of the business form a system of relationships that exchange information and resources in a constant and definable way. Time sheets feed accounting so they can do payroll, but need charge numbers on the time sheets to feed financial cost controls and bucket costs for filing taxes. These are mathematically speaking, complex systems like what Dr. Fry works with.
Hopefully, the system thing is making sense. System, ecosystem – simply a collection of processes that flow resources and information through the organization to accomplish what needs to happen. Working in an Engineering firm, I see plenty of technical process flow diagrams for things like refineries and power plants – and the methods are the same as when we are doing business processes. There is a reason why there is a whole “systems thinking” movement in management consulting. It’s accurate to the real world, and works when done well. I was amazed when I started studying business strategy, all these cyclical strategy frameworks were almost identical to a standard engineering control theory.
Funny to me at least, I bring up that point at an ASP meeting fully of strategy consultants that developed these business strategy frame works, and they are honestly ignorant of their parallel development of the existing engineering science that is many decades old.
Well, kids, there is a math to your organization’s system. And an easier paradigm than complexity theory. Google the term “topology”. It’s a branch of mathematics developed to study shapes; but what topology really does is map the relationships of different groups with different functions. Which means two relevant things. 1 – there is a math for quantitatively modeling the relationships; i.e. the flows of information, resources, and processes within your organization. And that math is also the math of databases; so if you can find someone willing to learn the theory of a homeomorphism (one to one vs. one to many relationships), you can have a model or framework for understanding exactly how to map the system of processes of your organization; and even use it to streamline the series of databases that IT already maintains to try and manage that existing system via IT based business intelligence (ERP). This allows you to data mine your big data intelligently if you understand the topology, the relationships between how different sets of data interact.
Then to miss Fry’s point. If you can find some graduate students in complexity theory, you can model the behavior of your organization’s processes and accurately and precisely predict how they will react under different scenarios. You can take what good managers do instinctively, and improve on it.
Now, let’s be honest. The vast majority of us have not studied graduate level mathematics like complexity or topology. But you don’t really need to do formal math to be effective. Math is just the science of patterns and relationships. It’s built on rules and numbers. But the rules work without the numbers (accurately if not precisely).
If you can understand the difference between a one to one and a one to many relationship, and notice that a pattern of a one week delay in supply chain results in a two week delay and cost in manufacturing – You can sketch out some simply process flow diagrams to model the systems of your organization, and see the patterns of what goes right, what goes wrong, and will have a legitimate engineered approach to troubleshooting and understanding what goes right and what goes wrong. It works for plants filled with a billion dollars of machinery, it can work for your organization filled with people doing various interdependent functions working towards team production.
Learn the concepts, and it can paradigm shift your ability to understand and solve problems. If you can define and recognize the consistent patterns of your organization, you can predict how changes will affect those patterns.
We covered, well, a lot in a small space. I obviously have too much applied math to get of my chest.
Dr Fry can’t tell you when a riot will happen. But she can tell you how a riot will happen. That allows for some very precise scenario planning. Be ready for an event or change before it happens. And honestly, scenario planning has been around for decades. Even going through the act of inaccurate scenario planning prepares the participants for change – orients them to better understanding the patterns of an organization and how they react to change.
If you understand the patterns, you have a chance at anticipating, managing and changing them.
Here’s the math free example of applying the math of understanding patterns of behaviour and relationships. Charles Duhigg wrote a very useful book called the Power of Habit. He thinks in terms of habits and the behavioural triggers that cause them. Which is simple math: Trigger = Habitual Behaviour; only gets more complex as you see how a system of habits as a process in a population.
So what? Well, how’s this for problem solving. Some years back, Mr. Duhigg spent some time in Iraq during the war. In the city he was in, the Army had a problem that every evening people would gather in the town square, protest, and there were some riots.
Over a few several days, Army leadership observed the habits – people didn’t show up to riot. The town square was also a market. People arrived to shop, ate dinner, and then were already gathered in the evening. Trouble simply ensued after dinner.
Solution? The Army Major tasked to solve the problem simply banned kebab stands from the town square.
So at the end of the day, people in the town square got hungry, went to find food, and left the square. Ended violent protests and riots with zero force, no violence, no death.
Charles Duhigg rightly calls that the power of Habit. Someone like Dr. Fry would likely call that an elegant solution by simply understanding the equation, and changing a variable to zero. I simply call it good strategy.
Dry Fry may be able to predict the pattern of a riot. In one instance, Charles Duhigg witness an Army officer able to identify the pattern and disrupt it, and prevented the riots from happening. In other words, he successfully predicted the conditions needed for a riot. They can do that because in there own way, they do the math to understand why things happen the way they do.
The lessons here are simple:
- You can’t predict the future (unless you are very well informed and good at game theory, then you can a little – but we’ll leave that for next week.)
- But if you can identify predictable patterns of behaviour for a system of relationships ( e.g. an organization), you can predict how that behaviour will happen; e.g. Dry Fry and her complexity models that show how riots happen.
- The better the math (pattern recognition), the more precise the model. But even simple models can yield accurate and useable predictive results for patterns of behavior.
- You can use a system map or process diagram to identify the relationships and patterns of your organization. That will give you a framework for predicting the patterns of your organization.
- If you can predict how your organization will respond to change, you can prepare for and manage that change. Whether a change in the market, or any deliberate strategic change to the organization driven by leadership. If you know how the organization works as a system of processes, you can deliberately engineer the desired results.
And the big points
- You know change is coming. You know the world is not standing still. This is a tool that can help you be ready for change, and respond and change your organization and it’s system of processes in a faster and more appropriate way. Simply put, an appreciation of math combined with some leg work, and you can adapt faster and better than your competition.
- Yes kids, this is another approach to process improvement.
- Even better – If you can do that internally with your organization, you can probably also map the system of processes for your value chain, supply chain, market, competition, etc. And get the same sort of feel for predicting and appreciating changes external to your company. Allowing you to adapt your external view and relationships just like the internal above.
- Even better, if you take the time to understand how things work in the world, you may be able to manipulate (influence) the process to get results more in your favor. I.E. it may be possible to alter a riot in progress, or a market shift to change the outcome predicted by the model, and achieve a more desirable outcome.
You can’t predict when events will happen, but you can be ready with a good response for when they do. Yeah, it’s basically the boy scout motto. But there is legitimate math that can help you manage the problems that are too big to measure, and too unpredictable to treat simply.
By rigorous study of the math of the patterns, we can learn them, and that allows us to predict, adapt to, and potentially disrupt, prevent or improve those patterns.
If you understand the patterns of conflict, you have the potential to predict, adapt and manipulate it to a more desirable result.
And in my mind, that is what strategy is all about.
If you made it this far, thanks for reading my pedantic rant on math.
-Ted Galpin SPP
Pingback: Can Bruce Bueno de Mesquita predict the future? | StrategicScience.Org