Simple volume forecaster

Paste historical volume below (one observation per line). The forecaster fits a Holt-Winters model with the seasonality you choose, and projects forward. Auto-fit will search for the best smoothing parameters.

Your data

Historical volume (one row per period)

Accepts “date, value” or just “value”. Daily data is recommended. Need at least 2 full seasonal cycles.

Seasonality (period length)

Forecast horizon (periods)

Auto-fit smoothing parameters

Forecast

Observations—

In-sample MAPE—

Chosen alpha / beta / gamma—

Naive seasonal MAPE (baseline)—

Forecast table

Period	Forecast

How to read this. Holt-Winters tracks three things at once — level (the baseline), trend (gradual movement up or down), and seasonality (repeating weekly/monthly pattern). The auto-fit searches for the smoothing weights that best fit the recent past. Compare in-sample MAPE against the naive seasonal baseline; if your fit isn’t materially better than “repeat last week”, the model isn’t adding value — and that’s a useful signal too.

How it works

This simple forecaster takes your historical daily volume and projects it forward using exponential smoothing with weekly seasonality — capturing both the underlying trend and the day-of-week pattern that dominates most contact-centre demand. It is a fast, defensible baseline: the honest starting point before you layer on the marketing events, product launches and seasonality that history alone can't see. Paste your daily numbers and it returns a forecast and a chart.

Common questions

What method does the forecaster use?

Exponential smoothing with a weekly seasonal component — it learns the level, the trend and the seven-day pattern from your history. It is a strong baseline that often beats far more complex models, especially with limited data.

How much history do I need?

More is better, but a few months of clean daily data is enough to capture the weekly pattern. Cleanse out one-off spikes (outages, campaigns) first, or the model will treat them as normal.