I have found the following to be essential in creating a coherent system of modelling.
Externalisation of knowledge
Models quickly get complicated. There can be thousands of calculations all interlinked. You will not be able to keep track. Additionally, the calculations themselves can be complex. You, therefore, need to rely on the structure of the model to tell you or remind you what's happening and how the elements interlink. Additionally the more that this is embedded into the structure of the model in a way that others can understand, the easier it is for them to quickly grasp what's going on in the model. This is essential for review, quality control and the ability to make modelling a collaborative endeavour.
Atomicity & simplicity
This means that the elements of the system are small, separable and as simple as they can possibly be. This applies to both the workflow and the structure.
For example, if your model is made up of small, simple units of analysis or calculation, it allows complexity to build up at the content level. These small blocks should tackle on calculation step. However, they can be combined in countless ways to deal with any modelling problem. And each step within the structure remains individually understandable. As you continue to build your model, it's going to become increasingly complex. Building it out of simple "atomic" components is the only way to manage that complexity.
The pieces fit together seamlessly.
In order for a workflow to become routine, the pieces must be simple and repeatable. If they are not the behaviour change will be too difficult.
The elements must fit together into a coherent and interlocking process where all the bottlenecks and frustrations are removed.
You can trust the system
It allows you to let go of having to keep everything in your head. When you no longer have to try and keep track of lots of information in your head, you can focus on what matters - the work, the ideas, the insight.
Everything is taken care of
If the structure we are using doesn't take care of everything that needs to be taken care of, the neglected bits will irritate us, elements will be neglected, and the missed items will come back to burn us later. This will then lead to "out of structure" hacks that degrade the overall system. Software developers call this "technical debt". We'll talk more about this concept later.
One of the biggest objections to adopting a structured modelling system is that it impedes creativity and prevents us from thinking for ourselves. In the next chapter, I'll talk about why this is a myth.
Sign in or become a Financial Modelling Handbook member to join the conversation.
Just enter your email below to get a log in link.