Best practice for modelling / analyzing seasonalities?

How do I?
1

Hi guys,

Do you have any tips, nifty functions or general best practices to analyze & model seasonalities ?

e.g.

  • strong vs weak months / weeks / weekdays of revenue
  • seasonality of online marketing cost
  • seasonality of customer inquiries
    Etc?

Thank you & Best Regards
Fabian

2

Hi Fabian,

Great question! There are many approaches, but these two are the most used:
1. Seasonality as a % of annual total
— In this approach you can create a seasonality variable that applies a % of a the annual value over 12 months (7 days, 30, days, etc) so that adding Jan-Dec would add up to 1 or 100%.
— then use another input variable for your annual target
— the resulting formula would read as : annual target x Seasonality|month
by placing month in the time portion of the formula, Causal will always pull the seasonality associated with that month Jan, Feb, etc.

2. Seasonality as a variance to the avg
— in this approach your formula should be written as normal to get a monthly value and then multiplied by Seasonality|month at the end
— Seasonality should equal 100% for “avg” months and 105% or 90% for above or below avg months.
— This approach works well if you have monthly trends that need to be shifted for seasonality

In summary, the first approach divides an annual target throughout the year and the second adjusts monthly numbers for seasonality.

Hope this helps!
Ariel

2 Likes