"Base" and "Season" are two of the sales drivers, that contribute to your total sales (not just media generated sales), that are quantified for any given period and is presented in the drill-down of the Total Sales KPI, in the Overview page.
Here's an explanation of what they mean:
Base
Base is a purely mathematically derived measure. By definition, base is the part of the modeling objective that cannot be attributed to any of the explanatory variables in the model. Seasonal fluctuations, yearly, weekly, even daily, are accounted for by HamiltonAI but is not a result of a tangible input. These fluctuations therefore goes in to the base as a way to describe sales that just happen but varies depending on the day of the week and time of the year.
Another big part of the base are reference values for various explanatory variables where a natural reference is different from zero. Weather, for example, is a variable where it does not make sense to say "how much would I sell if there was no weather?". In this case we instead use a reference defined as "average weather conditions" i.e. average temperature, wind speed, precipitation etc. Fluctuations in sales due to weather relative to the reference weather is attributed to the weather variable. To make sure all effects sum up to the total objective the reference weather has to be added to the base. A similar reference is also used to branding data, macroeconomics data etc. Media is an explanatory variable for which the natural reference is "zero media spend" and the base is therefore independent of media data.
Season
Season is utilised to account for temporally periodic effects that are not accounted for in other similar variables.
A common examples is the identification of certain characterics patterns, that are repeated every Christmas, January, Easter, summer etc. But the season also accounts for patterns within weekdays. E.g. sales on Monday may be larger than on Sundays. Finally it may also account for monthly patterns, i.e. differences between sales in the beginning of the month vs end of the month.
Illustrated examples: Total sales with distinct seasonal effects (weekly- and holiday season)
A caveat of the season is that it sometimes, may encapsulate the base. This often happens in cases where the base cannot be separated from the season in a meaningful way.
Building on the simple example for the Base (above): sales depends only on media and season. If the company stopped spending any money on media, they would still have sales. This remaining sales, are attributed the base, but now with a season, it depends simultaneously on the weekday, the day of the month and time of the year. I.e. the base is inseparable from the season.
Illustrated example: Season encapsulating Base
In this detailed (and ”smoothed”) example of seasonal fluctuations, we see that the seasonality almost ”covers” the whole sales spectrum, with highest total sales on Mondays and almost no sales on Sundays. In this case, there is no way of defining a Base, as the whole sales spectrum fluctuates over time