The Three Levels of Forecasting Maturity

At IQinAbox we speak to a large number of stakeholders in and outside the industry, and when it comes to forecasting - and the using data to make decisions - there are broadly speaking three categories.


Level 1: Experience based decision making

Experiencebased decisions will come natural to most people. This basically means that we use our experience to plan our actions and make decisions based on what we know works, what worked the last time, etc.


This will be sufficient for many companies and many decisions, especially if you are dealing with a very stable environment and predictable conditions. In most cases though, you may ask yourself if you can use some of the data you already have, to automate this or get a better insight...


Level 2: Basic statistical modelling

...and, of course, you're right! This is where the Excel-spreadsheet (or similar) comes in handy.

You start developing some basic models based on previous data on when you ordered feed, and you may even start generating some consumption averages per pig.


This will improve your forecasts and basic decisions. A simply Google search will show that almost any experiment done in the field of "standard Excel-spreadsheet solution" vs. human being will always end up with the human losing...

Level 3: Advanced statistical modelling

...but they will also show that there are even more advanced ways of looking at data that outperform Excel every time - and yes, as you have probably guessed, this is IQinAbox's solution.


The models that IQinAbox use are dynamic, so when the world or even a small batch of pigs changes, the model changes with it. The models do not rely on averages and they don't need two years of new data to come up with a new view of the world. This is done "on the fly". It also means that you can use the model to catch outliers, when not everything is within the norm. For most industries, 2020 and 2021 will fall into that category!

Furthermore, because it is done in the cloud, you can have your own data set, while at the same time benefitting from the millions of calculations that have been done on other similar data sets. Is there no end to all the great things that can be achieved?


What can I do if I am not at level 3 yet?

Well, to be honest, not many producers or companies are at level 3. But the world is moving in that direction and fast. The more complicated the issue is (like reducing CO2 or managing your supply chain during a global pandemic) the more we need advanced statistical models that can manage multiple variables dynamically.


They are not going to be perfect, but they will be far, far better than anything you have experienced before, and will even mitigate some of the issues of poor data quality to some extent...but never entirely :-)