This feature is part of Assembled's Pro and Enterprise plan. Please see our Plans page for additional details about our Assembled plans and associated features, and please contact the Assembled team at support@assembled.com if you’re interested in using this!

Overview

This document outlines the various configuration options available for the Seasonal forecast model, enabling you to tailor it to your organization's specific needs.

For more information, visit Forecast models and Finding your forecast model.

By default, the seasonal model generates a different weekly forecast for 60 days in the future. 

There are 2 parts of the Seasonal model:

• Arrival patterns
• Growth

Arrival patterns

You can configure and add the following arrival patterns in the Seasonal forecast model:

Configuration

The following arrival patterns can be changed from the default Seasonal forecast model configuration:

• Week of year
  • Default: Week of the year over the last 12 months
  • This can be adjusted to account for arrival patterns such as “higher volume around summer”
• Day of week
  • Default: Day of the week over the last 10 weeks (70 days)
  • This can be adjusted to account for arrival patterns such as “no volume received on the weekends”
• Time of day
  • Default: Hour (or 15 minute interval) of the day over the last 2 weeks (14 days)
  • This can be adjusted to account for arrival patterns like “higher volume around lunch time”

Additions

The following arrival patterns can be added to the Seasonal forecast model:

• Week of month

  • Default: Week of the month over the last 12 months
  • This can be adjusted to account for arrival patterns such as “lower volume the first week of every month”

• Holidays

  • If your company experiences volume fluctuations around United States or Canadian federal holidays, these arrival patterns can be added to your Seasonal model to account for volume fluctuations around these holiday dates:

    • United States federal holidays
    • Canadian federal holidays

Growth

By default, the Seasonal model uses 52 weeks for the trend line averaging weeks. This works well in most cases. That said, the number of weeks used for growth can be adjusted to increase forecast accuracy if necessary.