Calculating: Seasonality

Imagine you sell umbrellas. You had a great sales month in April, but was it great because of your new marketing campaign, or because April is the rainy season? By dividing your April sales by the April Seasonal Index, you can compare April's performance to March’s on a level playing field.

| Method | Best for | Ease | Accuracy | |--------|----------|------|----------| | | Clear, stable seasonality | Easy | Moderate | | Seasonal ARIMA (SARIMA) | Complex, autocorrelated data | Hard | High | | Holt-Winters (Exponential Smoothing) | Trend + seasonality | Medium | High | | X-13ARIMA-SEATS (Census Bureau) | Official stats, trading day effects | Hard | Very High | | STL (Seasonal-Trend Decomposition using Loess) | Robust to outliers | Medium | High | calculating seasonality

Ever feel like your data is trying to tell you a secret? You notice sales always dip in November but skyrocket in December, or website traffic peaks every Tuesday morning. That "secret" is —predictable, repeating patterns in your data. Imagine you sell umbrellas