# Introduction to Rearranging ## Overview Usually, we think about our Models in a certain "direction", based on what's most important to us. Let's say we are building a revenue forecast for our company selling widgets. The first Calculation we might think about is `Revenue = Widgets Sold * Price`, which looks like the below **(1)** in our Model:

The same Calculation might, however, just as easily be expressed as `Widgets Sold = Revenues / Price`, which would look like the below **(2)**.

The way in which we express Calculations usually depends on what we consider to be inputs and outputs of our Model. In **(1)**, we know how many widgets we are selling and what the price is, and we want to calculate the revenues generated. In **(2)**, we know the revenues and price, and want to calculate the number of widgets sold. Often, particularly when we are building Models that are a mix of **actual** and **forecast** data, what we consider to be an input and output will change over time. For instance, in our example above, we want to forecast future Revenues as an output, but we only know historic Revenues and Price, and not the number of Widgets Sold. For historic periods, therefore, we want `Revenues` and `Price` to be inputs, and in future periods, we want `Widgets Sold` and `Price` to be inputs. We can achieve this using **Rearranging**. The next section walks through an example of how to do this in Models.