The consumer landscape for supermarkets has changed rapidly over the last few years thanks to the explosive popularity of trends such as e-commerce, customer-centric floor plans, and food-court style seating. Each of these trends highlights an exciting glimpse of the future.
However, as supermarkets evolve in response, the need for improved operational efficiency is still a hot topic to date. One of the technological investments that supermarkets could consider is in improving their employee scheduling as creating more efficient schedules offers supermarkets the opportunity to reduce lost sales, minimize labor costs and improve employee happiness.
At Widget Brain, we’ve identified three key employee scheduling challenges that tend to block supermarkets from achieving that higher level of efficiency:
Forecasting based on different demand drivers per store
Creating shifts that address the different demand drivers per store
Prioritizing and assigning tasks within shifts
The question then arises; how, exactly, can these issues be addressed to make supermarket scheduling more efficient? We’ve already briefly touched upon the role of automated employee scheduling in our recent blog; Effective People Operations Strategies to Surpass Customer Expectations. Let’s now dig deeper into how AI-driven scheduling can help specifically per challenge.
Forecasting for different demand drivers per store
Supermarkets have different departments that each drive different demands for work, tasks, and required skills. For example, the bakery sells fresh bread and pastries, which requires employees with the skills to operate an oven. The number of cashiers, however, is not dependent on how much bread is sold in an hour, but on the number of transactions completed. This can be broken down even further by defining the demand per time of the day. Bread may be more popular in the morning while deli meats are sold more in the late afternoon. These patterns are unique per store and location. Making this clear distinction allows supermarkets to staff better against peaks or slower periods across all departments.
To make granular forecasts like these, we have to go beyond manual forecasting and leverage AI technology. Using past POS data or footfall will not suffice as it overlooks the impact of local weather, events, and other factors. Since different demand drivers each have their own data patterns, they also require different forecasting methods to get the most accurate results. That doesn’t mean historical averaging can’t be the right method, some of the time. The method can differ per demand driver and time period. AI-driven forecasting, or what we call Hyperlocal Forecasting, will allow you to achieve the most accurate results whilst automatically picking the right forecasting method for each unique location.
Creating shifts that cover the different demand drivers per store
Now that you know what your demand looks like per demand driver and the required headcount (as per labor standards) has been defined, you can create shifts to cover that demand. Correctly staffing against demand prevents excessive over- or understaffing and has a significant impact on your service levels and labor costs.
The process of creating shifts against demand often happens suboptimally as we tend to rely on “rule of thumb” approaches. Rule of thumb approaches mean that we take one day, see where demand is open, and create shifts to cover that demand. Then a check follows for any possible rule violations around required breaks and minimal and maximum shift lengths. Once shifts for day one have been finalized, you start again for the next day. This standalone and isolated approach is prone to errors and suboptimal results because unique situations and demand patterns can easily be overlooked. This approach results in unfair schedules for your employee's breaks that are scheduled during peak hours, and schedules that adapt poorly to unique situations.
It’s significantly better to take a holistic approach and use advanced algorithms to calculate scenarios with room for customization. This approach allows you to automatically consider all time periods, roles, skills, and relevant business rules. By putting different weights on different solutions you get shifts that cover the demand in a way that is exactly tailored to your preferences. For example, if a business goal for your supermarket is to minimize lost sales, you can prioritize accordingly and create shifts that fully cover the demand curve with the premise of more and longer shifts and preventing understaffing. The more people you have, the smaller the chance that shelves are empty, that employees have breaks at the same time and that customers have to wait. Perhaps the most promising part is that though you may be overstaffed, you can still create the smallest possible number of shifts considering shift lengths, minimizing overtime and all other factors, with a holistic and algorithmic approach. This approach is flexible when adding preferences per supermarket, per location, and per employee type. It also automatically takes employee preferences into account, creating fair solutions for all stakeholders involved.
Prioritizing and assigning tasks within shifts
To create an even better customer experience, you also have to drill down into tasks, based on their priority levels. It can be safely assumed that nobody benefits from unswept aisles, but empty shelves are a more important point of attention at any given time. There are two main challenges when it comes to scheduling tasks:
When it comes to prioritization, you want to make sure that the tasks impacting service level are completed first. Some tasks have to be handled within certain time windows, like unloading trucks. Other tasks are more demand-driven, like empty shelves that require immediate replenishment.
Grouping tasks is one method that AI uses to solve this complex puzzle. Algorithms can automatically consider all tasks and their respective handling times, time windows or demand, and the required skills to perform those tasks. To group the tasks, algorithms will also consider travel time, avoiding employees having to walk from one side of the supermarket to the other to perform the tasks, which in turn helps to improve productivity and reduce idle time.
Once these priority groups are defined, which employees do you assign to which task group? Every group has different required skill sets, so now the challenge is to match those requirements with the skill sets of your employees. Similar to creating shifts, the “rule of thumb” approach will lead to suboptimal results. That’s why we’d argue again to look at it holistically. This algorithmic way allows supermarkets to complete all high-priority tasks when needed and to assign employees who are able to do so. Perhaps most important is that this approach can fairly distribute employees to tasks based on their preferences and availability, to make sure not only operations go smoothly but your employees are productive and happy as well.
Start automating your employee scheduling today
AI-driven employee scheduling is the answer to complex forecasting and employee scheduling challenges. Automating this process is especially beneficial for supermarket chains with a large employee pool and several locations as it helps to minimize lost sales, improve employee happiness and fully comply with local labor laws at the lowest costs, in just 15 minutes.
Widget Brain can already help you get started with the data from your POS or WFM system. Using this data, supermarkets receive schedules tailored to their strategic business goals. Request a demo down below to learn how AI-driven employee scheduling can benefit you and to get a full understanding of our solutions.