Anyone who has worked in retail knows the anxiety that attends workforce scheduling for both manager and employee. The manager wonders “Will my employees show up at the right times?” The employee wonders “Will I be scheduled for inconvenient times? Enough hours? Too many hours?”
The ability of Uber and Lyft to attract drivers, despite average hourly wages that, according to one study, place them in the bottom tenth of all workers, lies in drivers’ ability to control their own schedule. If the companies face a shortage at a particular time, they just temporarily boost fees paid to drivers. The adjustment costs little in overhead. Forecasting algorithms (based on large amounts of data) allow pricing to anticipate demand. Uber and Lyft revolutionized an industry by recognizing that work schedule flexibility is worth a lot to employees. You would think other industries would try it as well.
What is known as “workforce management software,” or WFM software, is becoming increasingly automated and analytical, but scheduling remains just one component of what it does, and a relatively small one at that. Moreover, WFM is thoroughly evolutionary, not revolutionary, in its approach to shift-based employment. What used to be done by hand is now increasingly automated, but even that automation is proceeding slowly – according to the August, 2019 Gartner Report “Market Guide for Workforce Management Applications,” in 2025, 60% of large enterprises with hourly paid workers and variable demand for labor will still be using mostly manual (non-automated) scheduling approaches.
Still, WFM solutions are an active sector – Gartner reviewed 21 firms providing WFM services. Plus, there are players so large (e.g. Walmart) that they grow their own solutions.
What do these WFM providers do?
- One says it has an algorithm for automatically approving, and even suggesting, swapping of workshifts.
- Another uses voice recognition and text processing to handle employee requests for last-minute schedule changes.
- Several provide labor supply and demand forecasting (including some that are quite granular as to time and location).
- One predicts employee propensity to quit.
What might they do?
The above functions are useful, but limited in scope. The shift swapping algorithm does not determine the shift schedule in the first place. Some software does suggest shift schedules, but it rests upon the ability to assign employees to shifts at will.
Granted, extending flexible employee-determined scheduling throughout the service economy is a much more complex problem than that faced by the ride-sharing companies. Uber and Lyft regard their drivers as contractors, not employees, so they have avoided a host of laws and regulations pertaining to employer-employee relations (though California is seeking to change that). They also do not try to set driver schedules. Their only workforce “unit” to schedule is drivers, not a variety of employee types. So their scheduling problem on each of these fronts is much simpler than that of large shift work employers. Still, there are some functions not yet on offer that could “move the needle” on enterprise workforce scheduling to make it more employee-centric, and provide non-monetary compensation to employees.
- Provide an incentive-based system (cash or points) to build an automated quasi-market for scheduling within an organization – the system offers incentives for hard-to-fill schedule slots, employees could bid for desired time off or preferred shifts.
- Treat schedule requirements (e.g. you must work shift X) as a part of the system when the auction fails, not as an apriori element; offer compensation when this happens in a non-market way.
- Make scheduling a large-scale optimization problem, with employee satisfaction a significant component of the objective function to be optimized.
Sound impractical?
Think about the airline overbooking auction system, by which airlines must bid, auction-like, to induce passengers to give up seats when a flight is overbooked. The market-based scheme ensures that almost all transactions are voluntary, and enhance the welfare of both parties. When Julian Simon, an early pioneer in resampling statistics, first proposed this scheme back in 1977 (see his paper in The Journal of Transport Economics and Policy), it was roundly derided as unworkable. Regulators eventually forced airlines to adopt it, and it is now a smoothly operating component of airline operations. Mobile devices, plus machine learning and optimization algorithms drawing on auction economics offer significant possibilities to enhance worker satisfaction based on giving them better control over their schedules, while at the same time meeting organizational needs.