In this article, we highlight 10 myths about applying optimizers in workforce scheduling. It’s clear from these myths that there is a big potential to use workforce optimizers. These optimizers are able to improve employee engagement and efficiency simultaneously, as many of our customer cases prove.

Myth 1: An optimizer needs a definition of a “good” roster. But this is impossible to define.

An optimizer does need a definition of a good roster to be able to validate and compare rosters. Different stakeholders have different views on what constitutes good planning. The HR-manager wants rosters to comply with labor laws to prevent violations. The team leader will focus on covering all tasks and shifts. Employees want their preferences respected. The financial manager wants to minimize the costs associated with temp workers.

The good news is that all the examples above can be measured and accounted for in a roster. Organizations that want to start using an optimizer typically kick-off the project by bringing stakeholders together in a workshop, where they define the most important priorities for schedule optimization. They evaluate the quality of the current rosters. They can also apply machine learning on existing schedules to understand their accuracy and quality.

Tuning the workforce scheduling optimizer enables the organization to assess the quality of each scenario measured against the current roster, using the various criteria. This provides direct transparency and insights. Though the optimizer might not be perfect nor create a perfect roster for every stakeholder, the quality is much better compared to manually created rosters. The question is: are you ready for the level of transparency provided by the optimizer and are you ready to measure and improve?

Myth 2: Our schedule is made in a few hours, so an optimizer has hardly any benefits.

The focus of implementing an optimizer is not to decrease the planning effort but to create better schedules. Suppose you have a centralized planning process, with 5 planners creating a roster for 250 FTEs. This means that if you reduce the planning effort by 50%, only 2.5 FTEs will be necessary to create the plan. The remaining time can be spent on other work. But if you create just 2% extra efficiency in the roster, which is on the safe side, this will save you 5 FTEs, which is twice as much. The focus is to optimize the schedule, rather than to automate the planning process. Planning time will decrease in general, but this is just a bonus.

Myth 3: Our planners know all the wishes and circumstances of our employees. The optimizer does not know these, and they are not documented.

Planners certainly know your employees’ wishes and circumstances. But depending on the size of your organization, it may be impossible for them to know them all. It’s also likely that some employees are less assertive about specifying their preferences. And who makes the roster when planners are on holiday? When employees can enter their preferences on a system on their own, everybody’s needs are equal.

An optimizer can take more preferences into account when creating a plan, compared to the manual planner. Simply because it can take much more complexity into account and validate millions of options within seconds. We have seen cases where employee satisfaction, based on work-life balance, increased from 4 to 6.7 (on a 1-10 scale) after implementing the optimizer, and that sick leave percentage went down by 2 percentage points.

Myth 4: We are planning in cyclic rosters, so an optimizer is not relevant for us.

This might be the case. However, due to labor rules, aging workforces and changing employee preferences, more and more employees prefer a personalized roster. The demand for labor also changes over the week(s), even in factories and logistics. This means that leave requests must be filed quite far in advance and cannot always be confirmed. Our experience is that even when using cyclic rosters, there are sufficient opportunities to create better schedules, where optimizers play an important role. Notice that optimizers can also balance regularity and other factors, like costs and employee wishes. When employees are planned for a certain day (e.g., tomorrow) the next question is which employee works at which workstation. To optimize this, certificates and capabilities play an important role, but so do personal wishes, built up experiences, etc. A workstation optimizer can make much better and faster decisions to support this.

Uncertainty is a well-known challenge in various industries, like logistics and e-commerce. It’s nearly impossible to create a fixed schedule for the next month; even the next day has uncertainties.

A logical first step is to create a good forecast, based on historical information, interaction with sales or capacity management. Based on what is likely going to happen from a work demand perspective, you can create an indicative roster. You can also define some standby shifts or arrangements with working agencies. These insights help you avoid stress and uncertainty in the planning process. It can help you create better arrangements with your own employees and working agencies. For last-minute adjustments, an ad-hoc planner or rescheduling optimizer is very useful.

Myth 6: We tested the optimizer, but it did not create the schedules we were able to create manually.

This is a well-known complaint. Some companies even use the difference between the manual and optimized roster as a measure for quality. However, the question is whether the optimized schedule should look like the manually-created schedule. The objective is to create better schedules, not to repeat historical mistakes and keep the current quality. From a change management perspective, you want to have a seamless transition and preferably very few changes. We understand this. Yet, it’s better to think about how you can best manage the change and the transition, rather than how to keep the schedules the same.

Myth 7: We tested the optimizer, but it did not solve all our problems.

An optimizer is a great tool to improve the roster, but please consider that:

- A planner is still needed for the final touch.

- Data is key and needs to be correct, so must the rules and the settings of the roster.

- Change management is key since an optimizer makes different decisions compared to the planner.

Therefore, it is important to review your definition of a good schedule, examine the data and the rules, create a baseline to understand the current situation, and slowly but steadily try to get used to the optimized roster. Then, let the planner do the final touch.

Myth 8: Centralized roster optimization is old-school, we apply self-rostering, which is the way forward: the wishes of our employees are the starting point.

Self-rostering is a great way to involve employees in schedule creation. You don’t need to create a large number of rules of what’s desired and what not, since employees can plan their own schedules. That is, assuming you respect the required team sizes and labor rules. However, after the first round of self-rostering a second round must follow in order to solve all the over and under capacity issues. If this does not result in a feasible outcome, an optimizer is still required in the third round to solve the remaining bottlenecks.

If team sizes are relatively small, self-rostering can be a great way to create your schedules since it supports team spirit and collaboration. But if the team is very large, and there is limited experience in advanced rostering techniques, the change towards self-rostering can be too big. Our experience is that introducing self-rostering in traditional industries is a bridge too far from a change management perspective. Often, these companies go back to centralized planning and scheduling.

Myth 9: Setting up and maintaining optimizers takes a lot of time.

Regardless of whether you want to use an optimizer or not, good planning results start with understanding the characteristics of your planning process, the goal of the organization and how you can influence certain parameters. These steps are necessary for the implementation of any workforce scheduling system. Depending on the circumstances, the analysis of the current situation and the business rules, this may take some time. In fact, this takes probably most of the time, compared to the time to configure, validate and fine-tune the optimizer. The better you get to know an organization and the more often you participate in this process, the faster you’ll get at it. It’s important to train the functional administrator so that the right support can be delivered internally to deploy the solution over various departments.

This is an interesting argument but it’s always good to check the quality of the current roster. For example, are there any labor law violations? Are the agreements with working councils, unions, and others respected, including timing of publication? It’s also good to check how good the actual work demand was compared to the number of scheduled employees. And similarly, how many extra employees (e.g., from working agencies) were hired, compared to the available people within the own company. Finally, are employees really satisfied with their roster, their work-life balance, their ability to exchange rosters, personalized communications, etc.?

Our experience is that big improvements are possible when you apply optimization. Combine optimizers with forecasting, ad-hoc planning during the day and personalized communications for a winning workforce schedule. Depending on the maturity of the planning process, quality can be improved considerably. We have seen time efficiency and budget improvements of 2 to 5%. This means that there is a strong business case to invest in professional workforce scheduling software. The benefits for employees are also clear: optimized schedules improve engagement and retention.