Analysis
Seasonality
Also called calendar pattern.
Seasonality is a repeating pattern tied to the calendar, such as weekday rhythm, month-end, or the holidays. It shows up every cycle and is not growth or decline.
How it is measured
Look for the same shape in the same slot across several cycles. Weekly patterns need at least four weeks. Yearly patterns need at least two years. Compare each point with its own slot, not with the one next to it.
Describe it with a number: the typical swing from the average for a slot. Use that to judge new data.
Worked example
A tax-preparation site gets 2,100 daily sessions in March and 95 in July. Every year the pattern is the same: a ramp starting in late January, a peak on April 14 and 15, then a fall.
A new marketing hire sees a May drop of 60 percent from April and calls it a crisis. A chart of the last three Mays shows drops of 58, 63, and 61 percent. The actual May number is in line with the pattern.
How it differs
Seasonality is a repeating shape. A trend is the long-term direction after the shape is removed. A line can rise in a season and still trend down.
Common errors
Calling the seasonal low a problem. Using one year of history. Comparing a peak to a trough. Mixing weekday and weekend. Forgetting a moving holiday like Easter.
In practice
Overlay the last two years on one chart. Write the known peaks and troughs into your plan. Compare every month to its own past.