Financial modeling, as a profession, has developed at an incredible pace, especially in the past two decades. Previously classified under financial analysis, financial modeling now stands out as a highly technical and quantitative skill which extends into many arrays of approaches, complexities, practices and other industry standards that keep changing.
The skills set required by a capable financial modeler is so diverse that retaining such talents has become expensive. Institutions have tried to build and sustain operating models in-house, but the market dynamics keep forcing them to outsource to financial advisors again and again. It was worth the try though when a “spreadsheet” is costing you over $1 million to outsource!
So, efforts have been made to standardize and unify modeling practices to reduce the dependence on outsourcing, by minimizing the role of advisers to hopefully model audit and verification. This has succeeded in many sectors and projects.
And the question facing any CFO or FP&A Manager, when appointed with a new modeling task, would be: Do we build it from scratch or do we adapt to an existing model?
Before we dig deep into the pros and cons, we have to remember that we are dealing with lots of human judgement and bias towards background, experience, workstyle, time pressure and many other factors facing the modelers put to the task. And yes, that of fear of being accused of “re-inventing the wheel” always clouds such debates.
As a manager, you want less risk, even if you sacrifice some innovation in building a new model, but rather you want to deliver on time. Therefore, CFOs with no modeling background might always encourage adapting to an existing template. They might not know that despite project/sector similarities, it sometimes might actually take less time to build from scratch rather than fully understand and comprehend the mechanics of an existing template.
I am going to try to list some pros and cons, in a SWOT for a new build vs an adapted model, but I am sure many points could be added here and there and some of my points can be debated.
| Adapt to an existing template | Build a model from scratch | |
| Strengths |
|
|
| Weaknesses |
|
|
| Opportunities |
|
|
| Threats |
|
|
Well, to conclude, there is no conclusion! … as it will always be a case by case situation depending on:
- The background of the manager in charge
- The capabilities of the modeler(s) in charge
- The nature of the new venture or task
- The needs of the recipients of the model.
And probably other factors too.
Another question comes to mind, what about AI, could it relieve us from all of that? … asking an engine to produce a model. Probably yes, but that would work for simple business plans, at least for now, let’s see what the future holds.
Imagine hypothetically that AI took over this skill. If this were the case, then it would come down to a game of parameter inputs. But as long as modeling continues to be a philosophy in dealing with figures, using judgment and channeling feedback to operations to optimize parameters, then the industry can remain immune to being taken over by AI.
