Overview
LTP is an analytical-driven management consultancy company combining advanced analytics with business expertise. They have proficiency in supply chain network and facility design, inventory and replenishment, resource allocation, etc. LTP carried out the checkout areas planning project in stores for Sonae MC.
Sonae MC is the leading food retailer in Portugal with more than 1,300 stores, 36 thousand employees, and a 5,4 million-euro turnover in 2021. These stores are divided into 3 groups that LTP represented in their model. It included malls, supermarkets, and convenience stores in urban areas.
Problem
Service in checkout areas is very important as crowds of customers accumulate at the exits to pay for their goods. Therefore, retailers should take care of customer satisfaction at the last stage of shopping.
Sonae MC wanted to improve service quality and customer satisfaction without increasing operating costs. In addition, they aimed to understand how to leverage the wide variety of checkout solutions available in the market.
The goal was to determine the ideal number of checkouts of each typology to be available at each store, in order to fulfil the demand and expected service level. Another aim was to identify and prioritize the stores in greatest need of reconfiguration of the checkout areas.
The trade-off between costs and customer satisfaction increased the problem’s complexity. The cost of each checkout could be divided into fixed costs such as equipment, software and maintenance, as well as operational costs (cashiers). Customer satisfaction could be evaluated by the queue sizes and the average checkout time.
So, LTP needed to simulate several scenarios to find the ideal position in terms of the trade-off between retailer’s costs and customer satisfaction.
Solution
For this purpose, the consulting company built a multimethod simulation model in AnyLogic which included agent-based and discrete-event approaches.
In AnyLogic, it was possible to simulate any configuration of the checkout area and specify the characteristics for different scenarios. It was also possible to use different Java functions in order to simulate real-world process features as accurately as possible.
The checkout modeling required understanding the various stages of the process. The flowchart below shows the stages of customer service at checkout.
AnyLogic simulation provided both detailed (customer’s time in a checkout area) and holistic views (performance in a year). The model allowed LTP to test, in a virtual environment, checkout configurations that currently do not exist. Also, it was possible to foresee KPIs.
In addition, AnyLogic software gave LTP consultants great flexibility for the solution’s development.
Results
With the help of the AnyLogic model, LTP was able to get key results as well as retrieve some valuable insights for the retail business. The insights were, for example, stores’ stress test, scalability and personalization, scenario exploration, etc.
Testing of alternative scenarios revealed opportunities to decrease costs and improve customer service in checkout areas. The company switched some customers to self-service checkouts, and due to this, queues were reduced. As a result, fixed costs were reduced by nearly 15% and operating costs were decreased by almost 12%. In addition, the quality of service was improved.
The developed model provided the retailer with analytical support regarding the checkout dimensioning process. The simulation-based methodology ensured better support for decision-making thanks to the ability to test multiple scenarios before the real transformations.
Using AnyLogic, it was easy to identify improvement possibilities in the checkout process. Additionally, LTP was able to evaluate new and innovative checkout concepts for the retailer.
LTP plans to conduct a more profound study regarding customer preferences for each checkout typology, incorporating new market tendencies and shifts.
The case study was presented by Diogo Miranda, of LTP, at the AnyLogic Conference 2022.
The slides are available as a PDF.
