When I was a young modeller, just setting out on my epic journey to becoming an old modeller, I remember being impressed by the senior guys in my team, who seemed to have an uncanny knack of knowing what the results of my model were going to be before I did.
I would go off and spend many hours wrestling with detail and complex messy formulae (this was before I was any good at modelling and before I met the FAST standard), only to confirm what these guys already knew.
Or, more annoyingly, I would spend many hours of modelling, only to be told by these guys, in a second, without even looking at my model, that my model was wrong.
So, what was going on here?
What was going on was that these guys had ‘been there and done that’. They were experienced and knew what to expect.
When you are modelling, always know what to expect. Get to know your model’s vital statistics.
‘The model says so’
When working with junior modellers, I have lost track of the number of times that I have been told ‘the model says so’.
The costs have increased, but the price has come down, does that seem right?
“no, but the model says so”
The interest rate has increased, but the total cost of debt has reduced, does that seem right?
“no, but the model says so”
We’re spending £1m less capital, but the price goes up?
“no, but the model says so”
‘The model says so’ is never an acceptable answer. As a modeller, think through the logic of what you are doing and always know what to expect. It may well be, and very often is the case that the model is right, but you have to be able to explain what is going on, both to clients, managers and to yourself. Get to know your model’s vital statistics.
What to do
Think through the key dynamics and vital statistics of your model. What are the main variables which are likely to be moving around and how would you expect movements in these variables to impact the model.
Test out these key variables.
This will vary from sector to sector and for different types of modelling, but by way of example, for a project finance model, I would always test out the following to get an idea of how the model reacts.
Capex impact on Unitary Charge (Price) – test out what a £1m increase / decrease does to the price.
Annual operating cost impact on Unitary Charge (Price) – test out what a £1m increase / decrease does to the price.
Lifecycle / maintenance cost impact on Unitary Charge (Price) – test out what a £1m increase / decrease does to the price.
Debt Interest rate – test out what +/- 10 basis points does to the price.
Equity IRR – test out what +/- 10 basis points does to the price.
Once you are armed with these statistics you will be able to quickly rationalise whether changes you are making to a model make sense. You will spot errors or anything that doesn’t ‘look right’ as you go.
Furthermore, if you are operating within a single sector, it is likely that the vital statistics of similar models will be broadly comparable. You will build up experience and a knowledge of what to expect.
Follow this process, and this will make you a better modeller. You will be able to impress in meetings, by knowing what the model is going to do, before doing it. You too can become the guy who impresses / annoys the newbies.
