Problem
HAVI is a $5 billion global company and McDonald’s long-time supply chain and packaging partner. They provide services for supply chain management, packaging, logistics, recycling and waste. When McDonald’s wanted to build on the success of the All-Day Breakfast launch by expanding all-day availability to more menu items at 14,000 restaurants, they encountered a number of challenges (menu complexity, new equipment needs, space constraints).
McDonald’s business objective was to equip and staff its kitchens to obtain the best financial yield possible for the menu expansion. Working with HAVI, a simulation model was created that reflected the enormous complexity of the supply chain and operations across the 14,000 restaurants. The value of the model was to enable more informed decision-making for equipment purchases and staffing.
Read also: Developing Disruptive Business Strategies for tips on using simulation to predict the future with qualitative scenarios, historical data analysis, and spreadsheet modeling.
Solution
HAVI employs an iterative hypothesis-driven process for its simulation and analytics, balancing data with human experience.
To meet McDonald’s requirements, the model considered:
- Regional preferences
- Menu complexity
- Cooking space
- An increase in the variety of items being cooked simultaneously
- Equipment and staffing
Using AnyLogic, these requirements could be met and simulated, along with the spatial constraints and the wide variety of equipment and labor configurations. Decision variables of the simulation model included:
- Equipment type
- Equipment space requirements
- Labor needs
- Demand rate
- Batch size
- Product mix
- Store layout
- Drive-thru demand
On the output side, it was vital to measure the customer experience. As a result, factors such as service time, product freshness, and waste, among other service metrics, were also included in the model.
Finally, according to the rigor of HAVI’s analytics process, the model was subject to validation and calibration, including trials in McDonald’s test kitchen. The resulting model captured the necessary metrics and provided simulation comparable to the real world. In short, the AnyLogic simulation assisted in the decision-making process, providing McDonald’s with the best financial yield for the desired menu expansion.
The power of agent-based modeling in AnyLogic allowed the nature of the system to be captured as it is in the real world. The characteristics and parameters of equipment, labor, and the environment they operate in, can be modeled as necessary and custom objects developed for re-use.
HAVI chose to use AnyLogic simulation due to its support for multiple modeling methods, with agent-based, discrete-event, and system dynamics working together inside one system for the most holistic and powerful results.
Outcome
The AnyLogic model delivered results for a variety of demand profiles and restaurant configurations. This enabled HAVI to provide tailored recommendations.
These recommendations covered equipment needs and cost estimates for meeting customer service level thresholds in various scenarios. The benefits of which produced proposed equipment cost avoidance and optimized cost tradeoffs for labor and equipment.
Without AnyLogic simulation modeling, time and cost constraints associated with exhaustive physical tests would have prevented tailored recommendations.
Project presentation by Nate DeJong, HAVI
