SMART forecast explained
What is a SMART forecast?
The SMART forecast in PMI is a daily breakdown of your monthly productivity/hours forecast . SMART forecast allocates the month forecast to each day based on the expected activity levels of hotel using the cost driver forecast and historic staffing patterns. This can be used to plan staff schedules in line with the monthly forecast.
This is displayed in the labor cockpits to review when planning staff schedules. It is shown as the yellow line on the graphs.
What is the purpose of SMART forecast?
SMART provides a daily allocation of the hours determined in the monthly forecast by either the forecast hours or forecast productivity (depending on ‘Locked target’ setting).
SMART Forecast uses Machine Learning (ML) to recognize trends from previous periods and allocate the daily hours appropriately.
Use this as a guide to how many hours you should schedule each day in order to meet the monthly forecast set, while ensuring the number of hours is in line with the department’s needs on the day.
SMART Forecast is used for productive hours only. This does not account for any non-productive hours (training, fire alarm, sick leave etc.)
How is SMART calculated?
The calculation depends on the ‘Locked target’ specified in the cockpit settings. This can be reviewed and updated by users with admin rights.
Locked target should be set to ‘Productivity’ for operational departments, whose hours will fluctuate based on the activity of the hotel. Use ‘Hours’ locked target for admin departments where the staffing is fixed, and hours do not vary much.
If ‘Locked target’ in Tools is set to ‘Productivity’:
SMART hours forecast for the month = Primary cost driver forecast divided by productivity forecast
For example, a hotel expects to have 100 room nights (primary cost driver) this month, and the Housekeeping target is to clean 2 rooms per labor hour (productivity forecast). So SMART forecast would calculate Housekeeping hours needed for the month as 100/2 = 50 hours.
SMART Forecast then uses Machine Learning to allocate the hours per day. It looks at the expected activity each day, as well as trends from previous periods to determine a daily staffing recommendation.
If ‘Locked target’ is set to ‘Hours’:
SMART productivity forecast for the month = Primary cost driver forecast divided by hours forecast
This is displayed as ‘SMART Month-end’ in the cockpit.
SMART Forecast then uses Machine Learning to allocate the hours per day. It looks at the expected activity each day, as well as trends from previous periods to determine the daily staffing recommendation. The daily forecast is displayed in the graph, and are summed together to get the SMART MTD value.
For example, if there are many check-outs on a Sunday, but housekeeping usually waits until Monday to clean all rooms, the SMART Forecast will recognize this and suggest more hours on the Monday, even if activity levels are higher on the Sunday.
The same logic applies to SMART Budget.
SMART Last year hours = current month cost driver divided by last year’s actual productivity
This can be used to see if your current productivity has improved or not compared to last year.
Please note – Due to COVID-19, historic data from March 2020 up to (and including) February, 2022 are disregarded when analyzing historic data to predict future trends.
SMART MTD: The sum of the SMART forecast for the month to date days.
SMART Month end: The set monthly productivity/hours forecast.
Summary of what PMI considers when calculating hours in the SMART forecast.
- The monthly productivity forecast.
- Expected activity levels.
- Historical staffing patterns.
- The use of primary and other cost drivers to identify trends and allocate daily hours.
SMART settings
Use only primary cost driver in SMART
SMART looks at a wide range of cost drivers to identify historic trends and allocate daily hours based on history. There is an option to change SMART to use only the primary cost driver to allocate daily hours.
This is useful when;
- Staffing patterns have recently changed, thereby making the historic trends an inaccurate reference point for scheduling future hours, or
- The department should staff only based on the primary cost driver for the same day, and nothing else.
- For example, a standard restaurant without the need to prepare anything in advance and unable to save work for coming days.
- Departments are outsourced as cost usually solely depends on the cost driver level according to contract.
- This setting will be enabled as default for departments that have the cockpit type setting on ‘Outsourced’
‘Use only primary cost driver in SMART’ can be enabled by users with Administrator rights. Select the Tools icon on the relevant cockpit, and choose Settings. To check if this setting is being used, hover over the SMART forecast legend icon in the main graph.
Min/Max
This setting allows you to manually set the minimum and maximum number of hours that SMART should calculate on any given day. SMART will not forecast outside of these settings.
This is useful if you have little or no historic data in PMI, or if your staffing patterns have changed significantly compared to past periods.
You can also set Min/Max hours for a holiday season. For example, if you will be closed over the Christmas period, you can set the Max hours to 0 to ensure SMART will not allocate hours on those days.
Troubleshooting tips
This section provides additional troubleshooting advice for readers experiencing specific issues. While it’s packed with useful insights, you may skip it if you’re not facing any related problems.
Why is the hours total different in the table and the graph?
The difference in hours comes down to the type of forecast being displayed:
- Management Forecast (Table): This is the forecast set manually by management, based on their expectations and strategic goals. It reflects planned hours to meet operational needs.
- SMART Forecast (Yellow Bar in Graph): Generated by PMI’s SMART system, this forecast uses machine learning and historical data to provide a dynamic suggestion. It factors in monthly productivity goals and current hotel activity.
The variation between these two forecasts highlights the difference between static planning (Management Forecast) and data-driven predictions (SMART Forecast). Use this insight to adjust plans as needed for better alignment with real-time performance and goals.
Manual changes overwritten at Live forecast import
Legacy PMI System
- Issue: Pickup column doesn’t “gray out” when importing Live forecasts.
- Cause: This is the design of the legacy PMI system.
- Effect: Manual changes in the pickup column will be overwritten with each import.
New PMI System
- Preserving Manual Changes:
- Enable Editing: Click the robot icon (if auto-forecasting is enabled) or the import icon next to the day you want to edit.
- Edit Value: Enter your desired value in the pickup column (only editable column for future days).
- Save Changes: Press “Save”.
- Result: Your changes will be preserved even after a new import, indicated by a person with a pencil icon.
Glossary of Terms
Stuck on a term or abbreviation? Check out our PMI Glossary for quick definitions and insights into common terms and abbreviations. Simplify troubleshooting by understanding the language behind the processes!
Excluding segments from driver-based calculations
It is not possible to stop Machine Learning (ML) from including an existing segment in the driver-based calculation. If the ML Live Forecast doesn't align with your expectations, you have the option to manually override it or construct a custom driver to make necessary adjustments.
Enhancing forecast accuracy: Leveraging PMI’s auto forecasts
Discrepancies in the imported Room Night live forecast can impact the generation of live forecasts for arrivals, departures, and stayovers. To address this, it's recommended to utilize PMI's auto forecasts. These forecasts are typically more accurate as they rely on machine learning and historical data. Ensuring the accuracy of the imported Room Night live forecast is crucial, as it directly influences the precision of your live forecasts. Regularly review and adjust your forecasts to maintain their reliability and effectiveness.
Different hours for similar revenue on different dates can vary!
Recommended hours for similar revenue on different dates can vary due to a combination of factors including the monthly productivity forecast, historical staffing patterns, and other factors recognized by the machine learning capabilities of the SMART forecast.
Difference Between Live Forecast and PMI Prediction
If there is a recalculation of the PMI Prediction due to season generation, new file import, or manual calculation, it takes time for the ML algorithms to fully update. The Live forecast values will only update once all dates in the viewed period are completely updated. Thus, there might be a delay, and the Live forecast might not immediately match the PMI Prediction.
Can you activate Live forecast accuracy report in PowerBI?
The PMI system integrates Live Forecast accuracy into its PMI Index calculations. Although Live Forecast accuracy contributes to the PMI Index, it is not explicitly shown as a separate report in PowerBI. Note that the "Center of Excellence" is an additional module available in PMI at an extra cost, known as the PowerBI module. It offers a range of specialized reports that can be tailored to meet your specific requirements. For more information about the PowerBI module please contact support@d2o.com.
Can you activate a Live forecast accuracy report in Power BI?
The PMI system integrates Live forecast accuracy into its PMI Index calculations. Although Live forecast accuracy contributes to the PMI Index, it is not explicitly shown as a separate report in PowerBI.