ETS Holt Damped (Damped Trend Exponential Smoothing)

ETS Holt Damped is a time-series forecasting method within the ETS (Error–Trend–Seasonality) framework that updates estimates of the current level and trend over time, giving more weight to recent observations. It includes a damped trend, meaning that while short-term forecasts follow the existing trend, the impact of that trend gradually decreases over longer horizons. This helps avoid unrealistic long-term projections.

The method is typically specified as ETS(A,Ad,N), indicating additive error and trend components with no seasonality. It is particularly suitable for data with a trend but no clear seasonal pattern.

In North Data Company Intelligence, ETS Holt Damped is mainly applied to annual financial KPIs such as revenue. It performs well for datasets with limited historical observations and produces stable, conservative forecasts by preventing the trend from continuing indefinitely.

To improve robustness, data may be transformed before modeling to handle different scales or extreme values, and then converted back to original units after forecasting. Forecasts are typically generated for a short-term horizon (e.g. up to three years), where results are more reliable.

Overall, the method provides a practical balance between capturing recent trends and ensuring realistic long-term behavior, making it well suited for benchmarking, planning, and forward-looking analysis.

*ETS(A,Ad,N)
A specification within the ETS framework indicating a model with additive errors (A), an additive damped trend (Ad), and no seasonality (N), commonly used for non-seasonal time series where trends are expected to flatten over longer forecast horizons.