The 10 Commandments of Modular Financial Modeling

I. Never build from scratch

There are good reasons why most things - from websites to your mobile phone to the car you drive - are not built from scratch anymore; it's time consuming, expensive, risky, and results in a lower quality end product.

These reasons all apply to financial models which - like websites, mobile phones, and cars - can almost always be built more efficiently and more effectively using pieces of pre-existing financial models.

Any financial modeler who tells you that building financial models from scratch is a good thing either doesn't understand modular financial modeling or is motivated by conflicts of interest.

#NeverBuildFromScratch

II. Sharing is caring

Sharing static templates does not work because every financial model is different but, in the same way that an almost infinite number of different houses can be built using the same pile of bricks, an almost infinite number of different financial models can be built using different combinations of modules. Especially if we all share...

Modular financial modeling makes it easy to share modules (i.e. pieces of financial models) with colleagues throughout your organization, thereby ensuring that nobody ever wastes time reinventing the wheel.

Instead of dying in frustrating trying to refactor a colleague's spider web template you can now walk around your office asking if anyone has built a debt module, valuation module, or dashboard module similar to what you need, then plug it into your financial model in seconds before heading the gym for a workout.

III. Divide and conquer

Possibly the coolest thing about modular financial modeling is that it makes it easy for multiple people to build a model at the same time, each building different modules before inserting them all into the same workbook and watching them link together like magic!

When combined with leveraging existing modules, dividing and conquering in this way can reduce a multi-month financial model time frame down to one or two weeks.

Common divisions of build work include allocating the building of operational modules to one team member and financing modules to another, or different team members each building one entity before combining them into a single multi-entity consolidation model.

IV. Use the 40/40/20 rule

Prior to modular financial modeling, building financial models was a largely unstructured process of hacking and iterations which we often jokingly referred to as do, undo, redo. This process wasted huge amounts of time, which is one of the reasons why few financial models have a user guide and die within months of handover to users.

Modular financial modeling reduces model build time so significantly that the actual model build time should be around 40% of the project time, with the first 40% being model scoping and design and the final 20% being model documentation and handover.

The 40% scoping and design step of a financial modeling project is by far the most important contributor to the success of the resulting model, yet amazingly in a pre-modular financial modeling world this step was often completely skipped.

V. Make it scalable

The only constant in life is change, and financial models are no exception to this rule. So why build financial models that can't roll and scale with the underlying business?

An amazing side effect of making modular financial modeling work is that everything must be scalable, otherwise it would not be possible to, for example, insert a DCF valuation from a 5-year financial model into a 10-year financial model, or a revenue projection with 10 categories into a model needing 20 revenue categories.

This dynamic nature of modular financial models - from the time series periods to categories to modules themselves, means financial models can now be made scalable. And once you start building scalable models you can never go back.

VI. Avoid redundancy

For many years non-modular financial modelers have tried to get around the limitations of static Excel templates by building significant redundancy into their models - most commonly via excess categories, excess time series periods, and excess complexity just in case model users might want them.

Building redundancy into models increases the model size, complexity, and risk, which leads to more errors and higher development and audit costs. So it should be avoided at all times.

For example, instead of building a fixed assets module which allows for multiple depreciation methods and variable capex timing, consider building a range of fixed assets modules with varying levels of complexity, and only use the more complex modules when required my model users.

VII. Avoid megamodules

One of the primary benefits of modular financial modeling is module reusability. The more complex and specific any one module becomes, the more it resembles a template - which we all know don't work - and therefore the less reusable it becomes.

For example, if revenue and debtors are included in the same module - a mistake often made by accountants living in the land of debits and credits - it becomes impossible to replace just the revenue drivers within a financial model with different revenue drivers, which would be easy to do by simply replacing the revenue module if it were a separate module to debtors.

Megamodules can be appealing at times because they remove the need to think about links between modules, but in the long run you will never regret modularising your financial models with more granularity.

VIII. Include checks in every module

The beauty of modular financial modeling is that it quarantines complexity. Nothing is actually that complicated if you break it down into small enough pieces, and this could never be clearer than when building complex financial models.

Historically, financial modelers have included a few general checks in their financial model - the most common one checking that the balance sheet does in fact balance - and then left it to model users or other model developers to determine why the error has been triggered. In a modular financial modeling world this approach is lazy and borders on negligence.

Checks should be built into every module which contains anything checkable, ideally when building the module, thereby ensuring that model users are always able to quickly isolate the source of errors in a financial model.

IX. Cherish your content libraries

Modular financial modeling completely changes the financial modeling landscape. While historically the financial modeler most adept at inserting rows, using keyboard shortcuts, and copying and pasting in record time was our hero, s/he now looks obsolete when compared with a financial modeler with a wide range of existing module libraries.

There's a good reason why modular financial modeling has been banned from every financial modeling competition globally. It would be like allowing a machine gun at a knife fight.

Developing your own reusable and scalable content libraries allows financial modelers to commoditize their valuable skill set, moving away from selling their time to selling highly valuable solutions with a decreasing marginal cost.

If you build it they will come.

X. Version control on a module level

Version control is something everyone tells you they're doing but few people are actually doing well at all. This is primarily because static Excel templates don't work and hacking into them is an exhausting mess without the overhead of having to monitor changes.

Version control on a module level is much easier, and much more valuable. Some modules - such as working capital, fixed assets and corporate taxation modules - are so commonly able to be reused without customization that your irregular changes to them can easily be documented and version controlled.

Version controlling modules is already close to achieving one of our life goals at Modano: Pre-audited financial models. This is something that seemed impossible even 10 years ago, but with scalability and our ever-increasing depth and breadth of module libraries, many of our users are building completely customized financial models without actually changing the logic in any of the modules they've used to build these models. So they're effectively getting pre-audited financial models without having to pay for an audit.

Amen to that.

Michael Hutchens

Modano