FAQ: Understanding and Using APE

This article provides essential insights into utilizing Absolute Percentage Error for precise, property-level forecast accuracy assessments.

Ashley Dehertogh avatar
Written by Ashley Dehertogh
Updated over a week ago

What is Absolute Percentage Error (APE)?

APE is a statistical measure that quantifies the accuracy of forecasts by comparing the predicted values against the actual outcomes. It is calculated using the formula:

This formula provides a clear, percentage-based indication of the accuracy of property-level forecasts for an entire stay month.

How does APE benefit my forecasting process?

APE enables precise measurement of forecast accuracy, allowing you to identify discrepancies in your forecasting strategy. By understanding how close or far your predictions are from actual results, you can make targeted adjustments to improve forecast accuracy for each stay month.

Can I compare forecasts between different groups of properties using APE?

APE focuses on assessing forecast accuracy for individual stay months within a single property. You can use this to compare the accuracy of forecasts between properties, but should not be used to assess the forecast accuracy of a group of properties or stay months. Using APE will sum the actuals and forecasts across the properties to calculate the error percentage rather than building the mean across the error percentages for a more accurate representation of the data. Instead, Mean Absolute Percentage Error (MAPE) should be used in these instances of aggregation. This metric will be made available in the near future to support effective portfolio-level forecast accuracy assessment.

How can I access APE within FLYR for Hospitality?

APE can be explored through two specific dashboards within the FLYR for Hospitality platform: the Property-Level System & User Forecast Accuracy and the Property-Level User Forecasting Assessment dashboards. These tools are designed to help you analyze forecast accuracy in depth.

In addition to the dashboard templates, users seeking deeper customization can explore APE, along with all underlying data, through the Forecast Evolution Explore feature within the Data Explorer. This option offers enhanced flexibility, allowing for tailored analyses and reports to meet specific analytical needs, ensuring your evaluations precisely align with your unique requirements.

What's the difference between the Property-Level System & User Forecast Accuracy and Property-Level User Forecasting Assessment dashboards?

  • Property-Level System & User Forecast Accuracy Dashboard: This dashboard enables you to compare user-generated forecasts with system-generated forecasts, illustrating the evolution of forecast accuracy for a specific stay month.

  • Property-Level User Forecasting Assessment Dashboard: Focuses on analyzing the user's forecast accuracy, providing detailed insights into the forecast evolution and segment performance within a stay month.

How do I use APE to improve my forecasts?

Reviewing APE scores helps identify where your forecasts diverge from actual outcomes. Use this analysis to refine your forecasting methods, especially in areas with higher discrepancies. Continuous monitoring and adjustment based on APE scores are key to enhancing the accuracy of your forecasts over time.

Is there a tutorial on how to navigate and use the APE dashboards?

Yes, an embedded video tutorial is available below to guide you through navigating and maximizing the features of both the Property-Level System & Property-Level User Forecast Accuracy and Property-Level User Forecasting Assessment dashboards. This resource aims to equip you with the knowledge to effectively use APE in your forecasting process.

Where can I find more information or get help if I have questions?

For additional information and support, please reach out to our Advisory team via the chat. We are dedicated to assisting you in leveraging FLYR for Hospitality to its fullest potential, ensuring your forecasting efforts are as accurate and efficient as possible.

Did this answer your question?