Assortment optimization (AO) refers to the problem of selecting a set of products to offer to a group of customers to maximize the revenue added when customers make purchases according to their preferences. Assortment optimization plays a very important role in increasing profits, especially in the retail sector, due to limited shelf space, high stock costs and product diversity.
KoçDigital builds models to provide the most efficient product mix that maximize the profits at every location based on analytical analysis of customer behaviors and historical sales data. While creating the models, we consider factors such as customer, location of the stores and behavior of the customers at that location beyond sector specific attributes.
Benefits of Assortment Optimization In an environment where costs are on the rise and customer habits are constantly changing in parallel with technology, assortment optimization becomes crucial for companies providing many benefits. It enables
- Placing the optimal products on the shelves by using the shelf space more efficiently
- Preventing lost sales and thus increasing customer loyalty
- Maximizing profits by optimizing inventory cost and turnover
- Adapting to changes of purchasing behavior and market conditions quickly by analyzing the product mix and store structure through the model.
Establishing the right product assortment and getting the benefits brought by that product mix in a correct structure and to ensure its continuity is an essential task for AO.
KoçDigital’s Approach to Assortment Optimization Understand the business and the underlying business process is the most important step in addressing AO problem correctly. Model outputs strengthened with the correct set of business rules make the businesses themselves more efficient (one of the most important KPIs) by enabling making better decisions.
Generally spreaking, improving the way we operate is much more complex than improving the metrics. This complexity can only be overcome when more flexible perspectives are provided by working closely with business units. By this way, teams can establish specific business rules and address special cases at different levels such as location, product group, etc. in the modeling process together.
In addition to the benefits of assortment optimization, we provide the following benefits to customers:
- Obtaining the optimal product variety according to the customer behaviors in the relevant location by making optimizations specific to each store location and each product group.
- Discovering the potential sales of products that have not been sold in the relevant location before with estimation methods.
- Ensuring the automation and continuity of the model through end-to-end (model installation + model integration + model maintenance and development)
with the models we create.
Making Sure that Benefits Last After the set up, integration and the creation of the decision support system, we take the benefit we provide to our customers to the next level. KoçDigital’s success criterion is to ensure the continuity of the benefit that the customer receives. In order to ensure this, and to keep the model success at the highest level possible in current conditions, our Continuous AI team continues to work and review the model regularly. With Continuous AI team:
- Keeping the model responsive to new requests in an ever-expanding environment, regularly monitoring and making necessary improvements.
- Detecting and intervening data errors/changes that may occur over time
- Improving the model to adapt to changing market, industry and customer business conditions and consumer behavior
Through this perspective, we transform our models into ever-evolving tools that meet the changing needs of business units, rather than restricting them to one-time studies. In the projects we have launched so far with our customers, we continue to improve our models to achieve the continuity of the model with the highest possible customer benefit.
Atakan Kızıltan — Specialist Analytics Consultant, KoçDigital Serhan Bayram — Specialist Machine Learning Engineer, KoçDigital