Margins and sell-through rates under control: automatic markdown optimization based on price elasticity, demand forecasts and current sales and inventory developments.
It has been proven that price is the most effective lever to increase earnings and sales in retail. Mark-downs are an important instrument to counteract falling demand and to increase the attractiveness of a product. Every day, retailers are faced with the task of setting the optimal price for each product at the right time. Due to huge amounts of data (big data), a multitude of influencing factors and a volatile market environment, however, price management is becoming increasingly complex and the manual entry and processing of all data and information is no longer manageable.
aifora Markdown Optimization
Our Markdown Optimizer forecasts the ideal pricing strategies for each product based on current and historical transaction data and inventory information. Based on demand forecasts and price elasticities, the ideal markdown level, location and timepoint are determined. By combining inventory and profit-oriented pricing (= lifecycle pricing), the software is able to generate the optimum yield at any time of the season. For this purpose, internal company data is extracted and combined with further external data. The algorithms used react to changes in real time and learn continuously (machine learning). In addition to the algorithm-based forecast, aifora takes into account business rules, budget targets and planned promotions. The integrated workflow engine supports the internal decision-making process and the implementation of pricing in brick-and-mortar retailing.
Key Feature 1 – Mathematical Models
In contrast to many other solutions, the aifora markdown algorithms are based on statistical models. This approach makes it possible to map and control the company-specific influencing factors in a mathematical function. This allows the concrete probability of occurrence to be calculated and displayed for each forecast. Blackbox decisions are thus excluded and the necessary transparency and user acceptance is ensured.
Key Feature 2 – What-if Scenarios
To enable the user to simulate different price strategies, we offer the possibility to perform what-if analyses. In this way, the interdependencies become clear and users can incorporate their specific knowledge into the forecast. The system predicts and evaluates the impact on the defined targets, which makes these scenarios an ideal decision support for the users.
Key Feature 3 – Clustering
Product, store and country clusters are formed on the basis of a behavior-based process. It is not the existing attributes in the master data that are decisive for cluster formation, but the similarity of the entities in sales behavior. Sophisticated algorithms first determine the optimal number of clusters in the historical data and then assign the current products to the respective clusters in a dynamic process.