Learn more
Data Strategy
Learn more
A major international retailer with online and offline stores faced a challenge from competitors. As a market leader many countries, in a position to define products as well as price and quality standards, our customer needed to decide which prices to adjust and how the adjustment might affect future demand or profit. The coexistence of different players, markets and objectives made the task especially challenging.
Over the years, processes using tools ranging from Excel to custom-built AI platforms had been established in each market to determine price modifications. Most of these processes profited from the long experience of sales experts, who are responsible for the pricing, but lacked data to support the decision process. The client wanted an overarching price scenario tool for the six most important markets, which would be potentially valid for all global markets. The client approached FELD M for support in developing an intuitive tool to pursue a data-driven approach.
We started out with a data strategy workshop to define the overall scope from the central market’s perspective. This was followed by interviews with sales leaders from the six markets to collect their user stories and get an overview of the online and offline sales data. Together, these data enabled a forecast of the effect of price adjustments for product demand and profit as well as an estimation of the cross-media effects.
The client asked for delivery of a first MVP including the data architecture, data integration and development of a customized forecast model within six months. Based on the gathered data, we developed an initial concept for a price scenario tool including a model to predict different outcomes of price adjustments. In parallel, we started building up the data infrastructure in Azure and integrated online and offline data including millions of receipt lines per day into a huge data archive. Our concept was challenged by the stakeholders and iteratively improved and extended according to their feedback until all requirements were covered.
Within a short timeframe, the client received a customized price scenario tool, which was continuously evaluated by the sales leaders until rollout. To ensure availability in each market, the frontend was based on Excel and connected directly to an SQL Server in Azure. Our tool was able to predict the effect of price adjustments for a specific product as well as substitution and cannibalization effects between products. The architecture can be easily scaled to further markets with low costs within Azure. Our forecasting model (R in combination with SparkR) calculates the effects of price adjustments for more than 10,000 products within seconds.
What is a case study? We use the case study format to present our customer projects as examples. Our case studies on topics such as Data Integration or Data Strategy and Advanced Analytics aim to highlight the challenges and problems faced by our customers and the solutions we have developed. You will learn about the approaches we use to support our customers as service providers and partners and how we have jointly achieved the defined goals.