PMI Knowledge base

Generic filters
Filter by Custom Post Type
Article contents (TOC)
Print

Non-rooms Live forecast using PMI prediction

Please Share Your Feedback
How Can We Improve This Article?

Non-Rooms PMI Prediction provides an automated daily forecast for non-room revenue and activity drivers. It helps you build a more accurate Live Forecast by identifying patterns, seasonality, and trends unique to your property. The system updates daily as new data is received.

Availability: Non-Rooms PMI Prediction is only available for chains where this feature has been activated by d2o. For activation or more information, please contact support@d2o.com.

The automated forecast is designed to be a reliable starting point. If you see new trends or information that PMI cannot know—such as events or operational changes—you can override the value manually.

Imported to know that this module is only available in departments with an existing OTB import.

This article explains how to work with the Live forecast when PMI Prediction is enabled.

Step 1: Review the Prediction Value

The Prediction value shows PMI’s automated forecast for the selected day.
If the number looks higher or lower than expected, start by reviewing the Prediction details.

How to access Prediction details
Click the PMI Prediction value. A pop-up window will show

  • When the forecast was generated
  • Lead time
  • Season of the selected date
  • Models used
  • Historical dates available in the same season, evaluated by the algorithm
    • OTB status at the same lead time
    • Actuals
    • Pickup/change patterns
These details help you understand why the prediction was generated and whether it reflects the situation accurately.

Note: When Live Forecasting is set to segment level, prediction details are only shown per segment. Hover over the value for a tooltip, and click + to view each segment.

Your responsibility
Even when using PMI Prediction, you remain responsible for the final forecast. If you expect a different result, override the value as needed.

Step 2: Make Manual Adjustments When Needed

If you want to override the automated value for a future date, you can take over the day manually.

How to take over a day

1. Click the robot icon on the left side of the row.

  • The cell opens for editing.
  • The icon switches to a person icon, showing the day is now manual.

2. You will edit the pickup value, which will update the Live forecast. The sum of OTB + Pickup = Live forecast.

  • The Live forecast will be saved and keeps its value even if the OTB changes.
On manual edit, the OTB plus pickups will provide the Live forecast values that will be saved.

For properties forecasting on segment level, click the + icon to expand the row before editing. 

To return to automated forecasting, click the person icon to switch back to the robot. PMI Prediction will apply again after the next data update.

Note: The machine learning (“AI”) model improves over time as it learns from data. Properties without any historical actuals will need more time to adapt across seasons

When Should You Override the Automated Forecast?

PMI Prediction covers regular trading patterns well, but some situations require manual input unless OTB level already reflects a different level than usual:

1. New events – Dates with little or no comparable history, including:

  • Concerts
  • Conventions
  • Conferences

2. Public holidays – Holiday effects can vary from year to year and may need manual adjustment.

3. Short school holidays – Properties with short booking windows often need to handle these manually.

4. Closed periods – PMI Prediction can detect recurring closures, but variations in exact start or end dates cannot be predicted.

Examples:

  • Closed periods that shift slightly year to year
  • New closures not present in historical data
  • OTB appearing inside 28 days before a closure where historically there was none (PMI will assume the property is open)

In these cases, manual overrides are recommended.

For more details about the prediction logic, see the articles PMI prediction explanation and Collaborative Forecasting with PMI Prediction.