Ancient Greece’s Aristotle was able to provide strong empirical evidence of a spherical Earth. Yes, that’s more than 300 years before we started using our current calendar! But what’s fascinating is that the legacy of modeling Earth as a flat disk can be felt today, and “flat Earthers” are still more than a comical reference.
So how is this related to Finance today? Back in the days when people were using quills to write on a thick brownish roll of paper, Richard Millar introduced the term “Business Intelligence” [in his Devens' Cyclopædia of Commercial and Business Anecdotes (1865)]. As you can imagine, in 1865, calculation and data storage tools were so primitive that today’s average internet user probably generates more ad telemetry data per day than the whole 19th-century accounting industry (see the “doubling curve”).
To model business processes and to be able to do any predictions and analysis people in the 1800s had to be creative. So, they came up with business modeling suitable for abacus and quills, and this included the simplifications of business we use today: the payroll-based salary, the standard workday, and industrialized cookie-cutters.
Fast-forward to today, like the abundant flat-earthers of the early ages; we’re still modeling most of our processes as formulas suitable for paper and abacus calculations. Given the abundance of data and availability of computer power, we look like laggers trying to understand the world through simplified approximations and linear trendlines. What’s most dangerous is that we’re still teaching the younger generations to use these primitive models to run a business.
Businesses should switch from linear formulae to data-driven finance. This will allow companies to not only get an immediate boost in accuracy, but also permit them to become truly effective in predictive modeling, thus increasing overall operational efficiency.
Why is modern finance so outdated that it can be compared to Flat Earth? What do we do with this, and how can technology help to improve the situation? Let's talk about it in greater detail.
Intro: Data-driven finance vs. non-data-driven finance
There is one term that I hear unpleasantly seldom - dematerialization. In the US, cardboard consumption decreases even though Amazon sends everything in its cardboard boxes. Millions of packages are delivered daily, all packed in cardboard, but we use less of it than before.
How can that be even possible? That’s what obsession with optimization brings. This is a clear example of the efficiency data-driven finance can deliver compared to traditional cookie-cutter simplified modeling.
First, let's discuss the real difference between these two approaches:
Non-data-driven finance operates on a linear lever. For example, the company has a sales department of 10 employees. You multiply that number by 12 and by salary, and you get an annual sales department budget. How much insight can you get from this formula? How many decisions can you make?
For data-driven finance, you will have a table for each employee with all associated parameters such as salary, KPIs, past performance, contract terms, and location, etc. The financial model works with this table and generates predictions based on all listed parameters, not the employee count alone. The model can tell when the sales department will need to hire additional reps, what their real targets are, and what the optimal budget allocation is, given the stakeholder strategy.
What does this mean for the real world?
The data-driven approach provides a model that predicts when exactly a new employee is hired. It then predicts all associated and tracked parameters for this hypothetical employee. If these actions are made with confidence and high accuracy, the management understands what will change if the new rep is hired.
Now imagine what happens when all those numbers no one wants to calculate get calculated: the company grows 10%, so we increase IT spend by 10%. That doesn’t even sound funny in 2021! If you just start by representing your products and services as tables with multiple parameters and create rules by which they interconnect (like the sales department uses 10% of your IT servers) - that’s already a very good start. Gradually, you will gain a fundamental understanding of what your success depends on.
We need to teach people data-driven finance
If data-driven finance is that good, why isn't everyone using it right now? That's a good question, and the answer is simple -
It’s obviously more complicated than the abacus calculations the majority of people are used to.
Well, finger-counting is much easier to do, but we teach times tables for a reason. Something only becomes easier with practice, as it comes more naturally as you grow used to the idea. Data-driven finance is exactly the same.
How to solve this problem
Solving the issues surrounding the move from linear to data-driven finance mostly concerns how we educate those in finance. Dealing with data may seem more complex than traditional approaches, but however complex it seems today, we should stop teaching young professionals the simplified math models.
Humanity became able to achieve everything we have now by moving to more complex things step by step. Today, we do not study the Flat Earth concept before moving to the Spherical Earth, or we would risk seeing many more flat-Earthers.
We do not teach kids finger-counting before showing them the multiplication table. We just know that the latter is much more efficient and will get the student the value finger-counting can't provide.
Flat Earth and finger-counting are outdated, just like linear finance. And so, we should not stick to this obsolete method, especially as now we have technologies that can make it work like a charm. Modern AI technology is making data less alien to finance people. I founded HyperC with this exact goal - to take a step away from math and into the direction of data.
Final thoughts
Let's move from linear models in Excel to a data-driven approach. It is more complex at the beginning but provides higher efficiency in the long run. Start looking at the numbers in your Excel, and think about how much actual data is hidden under them.