Creating Demand Forecasts


In this video I’ll show how you can create demand forecasts for the products that you sell.

I’ll assume you’ve already imported some sales history - if not then you may want to check out the previous videos on importing sales data.

To get started, go to the Forecasts dashboard and click on the “New Forecast” button. This displays the Forecast Wizard where you can configure the forecast.

You can enter a description for your forecast, which can be anything you like: “Bob’s Forecast”, “Quarterly by Category”… whatever. This is just to help you identify it later on.

Under this you select a forecast interval. This can be either Quarterly or Monthly and determines how “granular” your forecasts will be. If demand for your products differs significantly from one month to the next, then you may want to try Monthly forecasts. However, if you don’t have much sales data for your products then consider going with a Quarterly forecast instead - aggregating the sales data may make it easier to detect trends and patterns. If in doubt (and if this is your first forecast) probably just stick with the default, which is Quarterly.

Next, you can choose the horizon for your forecast - which is simply how far out you want to forecast demand. As a rule of thumb, the horizon should be at least twice your lead time but no more than one year. We’ll discuss that furtherl in later videos on inventory planning but, for the time being, you can just leave this 12 months.

Finally, you can add one or more attributes to your forecast hierarchy. All forecasts have at least one level in their hierarchy, which is the “Root” level. The root level is a forecast of the aggregate demand for all of the products that you sell.

Adding additional levels to the hierarchy lets you forecast demand for specific categories, subcategories etc. I’ll talk a lot more about this in future videos, since choosing a good hierarchy is key to generating accurate forecasts. In broad strokes though, try to structure your hierarchy so that products that are likely to have similar sales patterns are grouped together.

Using the sample data, I’ll just go with Category, Subcategory and SKU.

Now we simply click Generate Forecast and then sit back and wait for the software to build our forecast in the cloud. This may take a little while. We’ve just asked the software to generate a root level forecast as well as separate forecasts for each category, subcategory and SKU that we manage.

Depending on the number of SKUs you manage, the report might take anywhere from a couple of minutes to a couple of hours to complete, but you’ll get an email letting you know when it’s ready.

Once it’s complete, you’ll see your forecast back on the Forecasting dashboard, where you can click on it to open it in the forecast explorer. You can use the forecast explorer to navigate, view and edit the various nodes in your forecast hierarchy. There’s a lot to see in the Forecast Explorer though, so I’ll leave for another video.