What is a forecasting model?

A forecasting model is a method that predicts a future value, such as next week's harvest, next month's orders or next quarter's price, from past data and known drivers. A good one also states how wrong it is likely to be. Wikipedia sets the test: "the accuracy of forecasts can only be determined by considering how well a model performs on new data that were not used when fitting the model".

Kinds of model

Some models extend the past: trends, seasons and cycles in the series itself. Some add drivers, such as weather, prices, orders or building activity. Machine learning models can combine many drivers.

The simplest model is the benchmark. Naive forecasts, the same entry notes, "provide a benchmark against which more sophisticated models can be compared": same as last week, or same as this week last year. A model that can't beat that isn't worth its cost.

How to judge one

By its error on data it hasn't seen. Cross-validation holds back periods, forecasts them from what came before, and measures the miss, many times over. That error only means something beside two other numbers. One is the size of what is being forecast. The other is the naive forecast's error on the same periods.

In a timber business

Forecasts support harvest and haul planning, log buying, production scheduling, inventory and pricing. Some drivers are themselves hard to forecast. A December 2024 review of 75 studies of wood products demand, by the Centre for International Economics, warns that using such a driver "could lead to a less reliable projection than simply allowing a time trend in wood demand". That review is about long-run demand projections. The same logic applies to shorter forecasts: a driver you must predict first brings its own error.

What it isn't

A forecast isn't a target or a plan. A scenario is different again: it asks what would happen if something changed, not what is likely.

Quarri for forest management is built around how a forest operation runs, from the cruise to the settled account.

Quarri is an AI-native data platform for the timber supply chain. It connects buying, production, sales and inventory for forest management, sawmill, wood products and pulp, paper and packaging operators.