Labor Forecasting

The labor forecasting feature uses advanced machine learning to forecast how many bookings you will receive on a given date by hour of the day, as well as how many staff members you will require to serve those clients. It considers the productivity of staff members scheduled to work that day to determine whether you need to add or remove staff members scheduled to work at that time, ensuring cost-effective staffing levels. In this article, we will walk through how the labor forecasting feature works. 

Things to know: 

• The ability to access the labor forecast is governed by the "Custom Work Hours" access control setting.
• An additional subscription is required for this module. Please reach out to support@mytime.com or your account manager to activate this module.
• This is controlled by a hidden setting that only the MyTime team can enable. To enable the hidden setting, please contact our Support Team at support@mytime.com.
• Labor forecasts are calculated only for staff members designated as service providers. Staff roles that do not perform services will not be included in labor forecasting calculations.
  
• If a staff member clocks in but does not clock out, that clocked time will not be displayed on the forecast user interface. This is because we are unable to determine when the staff member concluded working.
• We use "Prophet," a forecasting algorithm developed by Meta (Facebook). It's a powerful tool designed specifically for business forecasting and excels at finding patterns in data. 

  • What it Analyzes: The model processes one full year of your historical appointment data. For every hour in the past year, it maps the relationship between two key figures:

    • The number of bookings performed.

    • The number of staff required to service those bookings.

  • How it Finds Patterns: A key strength of Prophet is its built-in ability to detect multi-layered seasonality. We have configured it to specifically analyze:

    • Yearly Patterns: (e.g., if spring is traditionally busier than summer)

    • Monthly Patterns: (e.g., end-of-month rushes)

    • Weekly Patterns: (e.g., if Wednesdays are busier than Fridays)

    • Daily Patterns: (e.g., peak morning hours vs. slower afternoons)

  • What it Predicts: Based on all these learned patterns, Prophet forecasts the expected number of bookings and the required staff for each hour of your desired future period.

  • For more technical detail on the core algorithm, you can visit the official Prophet site: https://facebook.github.io/prophet/

• In addition to the 1-year forecast, a secondary logic layer was implemented. This new rule provides a "reality check" based on your most recent business activity.

  • What it Does: It analyzes the past 3 months of data to identify any significant changes in your weekly schedule.

  • How it Works: For example, it checks the appointment volume for every Monday, Tuesday, etc., over the last 3 months. If it detects that one day of the week (e.g., Mondays) has consistently shown extremely low or zero bookings compared to other days, it flags that day.

  • The Result: This adjustment overrides the main forecast for that specific day, predicting zero activity. This ensures the forecast reflects your current business reality (like a new closing day) and prevents scheduling staff on a day that is no longer active.

Contents

• Accessing Labor Forecast
• Background Colors
• Forecast Definitions

Accessing Labor Forecast

Once the hidden setting is enabled, follow the steps below to access it:

1. Navigate to your schedule
2. In the left panel, select the "Hours" button
3. You will be directed to the Custom Hours user interface. The forecast data will be visible at the top of the page
4. You can do the following:
   • Filter by a specific location
   • Filter by business hours or all hours
   • Filter by a specific day
   • Choose between the day or week view
   • View data for staff or booking forecast

Background Colors

Forecast Definitions

For more information, contact us at support@mytime.com.