Using Machine Learning to Forecast Sales
How we Forecast Sales:
We use a machine learning procedure for forecasting time series data based on an additive model where non-linear trends are fit with yearly, weekly, and daily seasonality, plus holiday effects.
This means that sales can be accurately predicted even if they have a strong seasonal effect after just a few seasons of historical data.
This model is robust to missing data and shifts in the trend, and typically handles outliers well.
Where to see your Sales Forecast
There are two places you can find your sales forecast. At the top of the insights page as shown below and in the Goals section once you have setup a goal:
The forecasted amount is displayed in the top left for the current period. The value shown is the Total Sales forecast for that period, that includes both realized and unrealized sales.
The solid blue line is a graph of realized sales for the period, while the dotted blue line shows the forecasted sales remaining in the period.
How to use Sales Forecasting
The best way to use Sales Forecasting is with the Goals feature. There you can set a metric to track over a certain period, like monthly sales. To learn more about how to setup and track your goals see our Goals documentation.