Hierarchical Forecasting

Aligning forecasts across various levels and functions within your organization can be both complex and time-consuming. Hierarchical forecasting streamlines this process by enabling you to create accurate forecasts at every level—whether it's weekly operational forecasts by product or long-term strategic forecasts of market trends. Once the forecasts are generated, hierarchical forecasting ensures alignment by assigning greater weight to higher-performing forecasts, resulting in cohesive forecasts optimized for accuracy

One source of truth

Hierarchical Forecasting breaks down the organizational silos and unifies the organization towards one source of truth, while improving the forecasting accuracy.

This interview with one of the contributors to the Hierarchical Forecasting research, Nikos Kourentzes, Professor at Skövde University discusses the concept and value of Hierarchical Forecasting.

Value achieved

By reconciling your forecast with Hierarchical Forecasting, you can:

Key features in Indicio

Frequently asked questions

What is hierarchical forecasting in Indicio, and why is it useful?

Hierarchical forecasting in Indicio means you forecast simultaneously at multiple levels, for example total, region, channel, product, or customer. Instead of only doing top down or bottom up, you model all levels and then reconcile them, so the hierarchy stays consistent and overall accuracy improves.

This lets you use different model types where they fit best. For instance, you might use simple univariate models at SKU level while running advanced econometric models at country level that include regional market and economic drivers.

Which hierarchies can I model (products, customers, regions, channels, time)?

You can model flexible hierarchies across the dimensions your business uses. Common examples are SKU, country, segment, channel, and time. You can also combine several dimensions in one hierarchy to match your reporting structure.

How does Indicio reconcile forecasts across levels (top down, bottom up, middle out)?

Indicio supports several reconciliation strategies across the hierarchy, including:

  • Identity/OLS with an identity error covariance matrix
  • Variance scaling (weighted least squares and structural variances)
  • Generalized least squares (with shrunk or sample covariance matrices)
  • Bottom up aggregation
  • Custom setups where you can lock specific hierarchy levels
Can I override forecasts at certain levels while keeping hierarchy consistency?

Yes. You can override forecasts at any chosen level, for example category or region, and then reconcile the hierarchy so all other levels adjust consistently. This lets you apply business judgment without breaking alignment between totals and detailed views.

How does hierarchical forecasting impact accuracy, bias detection, and reporting?

Hierarchical forecasting lets you use the right model at each level and then reconcile everything, which typically improves overall accuracy. It also makes it easier to spot bias by comparing errors across levels, and gives you consistent, drillable reports that match how you manage the business.

Explore more features

Explore all features

Virtual demo

View our click-through demo

Experience the ease and accuracy of Indicio’s automated forecasting platform firsthand. Click to start a virtual demo today and discover how our cutting-edge tools can streamline your decision-making process.