SARIMA

What is SARIMA?

SARIMA is an extension of the forecasting algorithm, ARIMA, that incorporates seasonality into the forecasting process. It is suitable for time series data that exhibit recurring patterns over specific periods, such as daily, weekly or monthly.

To turn those seasonal patterns into real decisions, plug them straight into Hakio’s Fashion Planning for collections, channels, and drops that match the forecast.

What are the components of SARIMA?

The steps involved in the SARIMA algorithm are similar to ARIMA, with an additional seasonal differencing step:

• Seasonal Differencing: In addition to regular differencing, seasonal differencing is applied to remove the seasonal patterns from the data and reveal the underlying trend. In practice, those seasonal insights are only valuable when they guide planning - which is exactly what Hakio’s Budget Planning does by translating seasonal forecasts into SKU-level financial targets.
• After seasonal differencing, the AR, MA, and seasonal components capture autocorrelation, moving averages, and recurring seasonal effects. These patterns become business-critical when translated into decisions - for example, through Hakio’s Buying Module, which turns forecasted demand into purchase suggestions that respect vendor rules, lead times, and MOQs.