AI Explained | How much historical data does ChannelMix need for my model to run?
<! -- ChannelMix Monitor Blue Paper --> <! -- GA4 User Guide --> <! -- Keys User Guide -->
Machine learning models typically require large amounts of historical data for the model to train itself correctly. For instance, to learn a year-over-year trend we will need at least 90 days worth of valid data. It's important to note that while we can model as little as 90 days worth of data, the more historical data you have available, the more factors we can model more accurately, and the more confidence the model will have.
What does "valid data" mean? Data being valid means that there are at least 2 distinct marketing channels with activity and activity for your target variable (typically sales, leads or customers).
Every month ChannelMix adds your most recent media data into the model. As new data is added, the model will update and refine its predictions based on these new results.