Tanda’s Predictive Rostering tools can help businesses:

The key Predictive Rostering features include generating a forecast for your optimal headcount, automatically generating schedules and assigning shifts, and allowing roster managers to see where they are over and understaffed.

In this article, we will cover:

Tanda can automatically provide recommendations on how many staff you will need at each point in the day. We calculate these by analyzing historical schedules and incorporating external signals such as weather, foot traffic, holidays, demand data (e.g. sales), and other demand factors. Tanda will then generate datastreams to drive the auto generation of rosters.

To enable this option head to Settings > Roster Settings > Predictive Scheduling:

* When first enabled, please allow up to 5 minutes for the set-up to complete. Tanda will generate datastreams from our AI/ML service to drive the predictive rostering looking like the below:

Each datastream relates to one of your teams and allows Tanda to autofill your rosters. You don't need to do anything with these datastreams, they are fully automated.

This option uses productivity ratios (“Roster Ratios”) to calculate the optimal headcount for each team. This approach is appropriate if your business has a clear mapping between demand signals and staffing requirements (e.g $150 sales per hour = 2 employees). There should also be minimal other factors that influence productivity ratios and no lead/lag between staffing requirements and demand signals (e.g prep time).

To enable this option you will need: input data imported into your account, rostering ratios entered for each team, and the necessary cognitive settings configured. See this help guide for further information.

In Workforce/Tanda each location has a business hours field. This field lists the hours your store is operating and should include any hours which employee(s) may be scheduled.

To ensure the predictive scheduling feature works properly, make sure these hours are set correctly. If in doubt, set them to the default of 4:00 AM - 4:00 AM. You can make this change by navigating to Workforce > Team > Edit Location (show advanced options).

Once the above steps are completed you can view the recommended headcounts within the rosters screen.

Weekly View:

When in weekly or bi-weekly view you will see the recommended and rostered hours below the date (if this is your pinned stat, click on the stat indicated to change this). The below example indicates you are likely overscheduled by 0.25 hours.

Day View:

When viewing the schedule on day view you have two options. The first is the table view, which for each team shows you: projected demand signals (sales etc), recommended staffing plan, and actual staffing numbers. Blue shaded cells signify over-scheduling and red shaded cells signify under-scheduling.

The second option is the graph view which shows the same information but displayed in a different format.

If your business has data streams (e.g Sales, Orders, Transactions etc) connected in Tanda, rostering managers can view, edit and analyze these forecasts in our Predicted Demand modal. Editing these forecasts will contribute to each team's recommended headcount numbers. So for example, if you believe the sales projects and recommended staffing level for next Friday is off by 18%, you can easily rectify this by editing your demand, as demonstrated below.

Once you're happy with the above forecast and recommended headcount, the next step is to build and fill your schedule. If you are viewing an empty schedule you will be prompted with the option to build and fill your schedule in one step.

Clicking this will build a complete schedule for up to two weeks. Once built you will also have the option to fully or partially revert the schedule build.

If you are either viewing a partially completed schedule or in a non-weekly view , you can build and fill schedules using the buttons found on the left tab of the schedules screen.

Note: If the “Build Empty Roster” is disabled it may be due to one of the following scenarios: you are attempting to schedule too far in the future, the team(s) you’re building for don't have sufficient historical data, or there is an issue with step 1 of the set-up process. Please contact support if you cannot get this rectified.