Share via

Forecast in excel Seasonality

Sinead Nolan 0 Reputation points
2026-03-24T13:01:56.65+00:00

Hi there,

How is seasonality accounted for in Forecasting in excel, is there a numerical value that is attached to it? I'm trying to manually verify the values gotten through forecast and there is a 0.5% difference between the excel value and the value I'm manually calculating using y=mx+c (linear regression) so I'm hoping that there is a numerical value given to seasonality.

Microsoft 365 and Office | Excel | Other | Windows
0 comments

1 answer

Sort by: Most helpful
Most helpful Newest Oldest

  1. AI answer

    2026-03-24T13:02:13.8933333+00:00

    Excel’s built‑in ETS forecasting functions model seasonality explicitly and do not use simple linear regression y = mx + c. That 0.5% difference is expected if manual calculations are based on a straight line while Excel is using an exponential smoothing model with a seasonal component.

    Key points:

    1. How seasonality is handled
      • Functions such as FORECAST.ETS use the Exponential Triple Smoothing (ETS, AAA) algorithm.
      • Seasonality is modeled as a repeating pattern of a certain length (number of periods).
      • Excel automatically detects this seasonal length unless a specific seasonality is supplied.
    2. Numerical value for seasonality length
      • The function FORECAST.ETS.SEASONALITY(values, timeline, [data_completion], [aggregation]) returns the length of the repetitive pattern Excel detects for the time series.
      • This is a numeric value (e.g., 12 for monthly data with yearly seasonality) and is the seasonality length used by FORECAST.ETS when the same parameters are passed.
    3. Other numerical seasonality parameters
      • FORECAST.ETS.STAT can return ETS algorithm parameters, including:
        • Gamma parameter of ETS algorithm – the seasonality weight parameter (higher value gives more weight to the recent seasonal period).
      • FORECAST.ETS.STAT also returns error metrics and the detected step size, which are useful for validating the model but are not part of a simple linear regression formula.
    4. Why manual y = mx + c differs from Excel’s forecast
      • FORECAST.LINEAR (or classic FORECAST) corresponds to linear regression y = mx + c.
      • FORECAST.ETS uses ETS with trend and seasonality, so its forecast will differ from a linear regression line, especially when a seasonal pattern exists.
      • To match Excel’s ETS forecast manually, the full ETS equations with the detected seasonality length and parameters (alpha, beta, gamma) would need to be replicated, not just a linear regression.

    To inspect the seasonality Excel is using, call FORECAST.ETS.SEASONALITY on the same values, timeline, and options as the FORECAST.ETS call, and use FORECAST.ETS.STAT to retrieve the gamma (seasonality) parameter and other statistics.

    References:

    AI-generated content may be incorrect. Read our transparency notes for more information.

    Was this answer helpful?

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.