Creating a model, at its core, is a way to understand the world. It is a process that allows us to translate complexity into clarity and help us make better decisions, especially before heavy investment is made or before important decisions are taken. It also helps uncover hidden behaviours and relationships thus predicting what might happen next and also provide the best outcome possible or to use the best variables possible for a given solution or scenario. 

Modelling starts with the purpose that you are trying to achieve. A model should never exist in a vacuum or in isolation, as it is built to answer questions and to solve problems. These could be how a product will behave under certain circumstances, simulating a process to reduce waste or testing what happens when you tweak one or more variables in a production line. Or even understanding how proteins behave in food systems, building economic models or understanding how manufacturing processes respond to different settings and mechanistic variables, or maybe even how a new formulation might perform even before stepping in a lab, or creating a first prototype. In essence modelling becomes a powerful ally, with the first step being understanding what you are trying to find out or what is the problem that you are trying to resolve. 

Once the purpose is clearly defined, the next task is to get to know the system, and this mostly comes with breaking it down. Understanding what goes in, what comes out and what happens in between is essential. In a food context for instance, this might involve ingredients, temperatures, moisture levels, microbial interactions and how all of this connects to each other and how these evolve together. 

After understanding the system, we then choose how to represent that system. Some models may rely on known scientific principles, often called mechanistic models. Other models are driven by data, picking up the patterns that might not be obvious to the human eyes. The best results sometimes come from blending the two creating a hybrid approach to modelling, by combining the two approaches (first-principles and data-driven models). The choice depends on the nature of the problem and the data available, as well as whether you have more knowledge of the system as well as quality data. Data is of course essential. It may come from experiments or previous studies. This information becomes the fuel for the model through shaping the model, calibrating and giving it meaning.

After this, you build. You build models using tools such as Python, R or specialised simulation software and we begin translating the system into something the computer can work with which could be an equation, a set of rules or a learning algorithm. We test the model and compare its behaviour to reality, refining it as we go. It is an iterative process often requiring a fair bit of tweaking and creative thinking. Once a well-built model is ready it can offer a new lens for decision making. It can help companies predict outcomes without trial-and-error, reducing costs by optimising processes, accelerating innovation timelines and making informed strategic choices, especially in environments where real-world experimentation is time-consuming, risky, or expensive. 

Here at Funis Consulting, we see modelling not as a technical afterthought but as a cornerstone of innovation. By combining scientific understanding with modern computational tools, we help clients simulate, predict and optimise across a range of domains, from food systems, to industrial processes and beyond. While every model we build is unique to the problem at hand, the goal is always the same: to make the invisible visible and the complex comprehensible. 

Modelling is about numbers and equations but it is also about insight and foresight. Increasingly, modelling is becoming a quiet but essential driver of smarter business in a world where agility and accuracy matter more than ever.