Improving Retail Demand Sensing: A Hybrid Approach
Unlocking the potential of retail business by adopting a hybrid approach to demand sensing, combining AI and human input to improve forecasting, and boosting sales growth
Retailers of all sizes use demand sensing to determine market trends and consumer preferences in order to make decisions about inventory, distribution, sales, and other aspects of their supply chains. However, traditional methods are no longer as effective as they used to be.
Recent fluctuations in the economy, the war in Ukraine, and the ongoing global health crisis have intensified existing supply chain challenges, putting pressure on retailers to improve their demand sensing methods and business decisions. Those who have relied too heavily on outdated processes or solely on institutional knowledge and desk research have ended up with surplus inventory or selling through the wrong channels – or have been unable to meet consumer demand at peak times and lost market share.
Overcoming challenges using a hybrid approach
Retailers can overcome these challenges and increase the value of their business by adopting a hybrid approach to demand sensing, which combines artificial intelligence-backed sensing and predictive analytics with human-directed learning and customisation. Consulting giant AT Kearney has found that retailers who implement this approach improve the speed and accuracy of their forecasts, allowing them to move the appropriate amount of inventory into the correct sales channels at the right time so products are available when customers want them. Additionally, more precise methods of demand sensing have been found to improve forecast accuracy by 5-20+ percent, resulting in a 5-10+ percent reduction in inventory safety stocks and up to a 10 percent increase in sales growth.
To achieve these results, retailers must identify and track the most relevant data for their business, use technology that is flexible enough to scale and accommodate future uses, and assemble teams with the appropriate mix of skills. They must also have the support of senior executives to drive the adoption of a hybrid AI-human-powered retail demand sensing strategy throughout the business.
One example of a retailer successfully implementing a hybrid demand sensing approach is Amazon. The company uses a combination of AI-powered forecasting algorithms and human input from its merchandising and operations teams to make decisions about inventory and distribution. This approach has allowed Amazon to quickly adapt to changes in consumer demand and remain competitive in the retail market.
Another example is Walmart, which has been using machine learning algorithms for demand forecasting since 2016. The company also uses human input from its merchandising and operations teams to fine-tune the forecasts and make more accurate decisions about inventory and distribution. As a result, Walmart has been able to improve its forecasting accuracy and reduce inventory costs.
Insight into Demand Sensing
The current limitations of demand forecasting methods became apparent during two major shifts in consumer behaviour in recent years, from higher demand for at-home products to less demand and more interest in lower-priced goods. Even major retailers with advanced analytics teams that excel in forecasting were unable to predict the speed of these changes in consumer sentiment. When consumers concerned about inflation and rising interest rates switched from buying high-ticket items to basics and lower-priced goods, even the best-equipped retailers struggled to keep up. Inaccurate demand forecasts have several underlying causes, including:
- Inadequate sensing: Existing systems for sensing demand are difficult to update quickly to account for external factors such as a volatile market or interest rate increases. Some fail to take into account the most relevant indicators that could indicate a shift in demand. Additionally, the historical sales data and simple analytics tools and techniques that retailers typically use to predict future demand are not as comprehensive as they need to be.
- Limitations of existing forecasting applications:Companies typically use demand planning software with built-in demand forecasting features or modules. These applications allow users to convert forecasts into plans for inventory, stockouts, revenue, and so on. However, the forecasting features in existing applications fall short. They do not include new forecasting methods or address changing market dynamics or new data sources. Many also include generic indicators meant to cover all industries.
In conclusion, to remain competitive in the retail industry, it’s crucial for retailers to adopt a hybrid approach to demand sensing that combines artificial intelligence and predictive analytics with human-directed learning and customisation. By doing so, retailers can improve the speed and accuracy of their forecasts, and make better decisions about inventory, distribution, sales, and other aspects of their supply chain. This can lead to increased efficiency, cost savings, and ultimately, growth for the business.
Know more about our Top Ranked PGDM in Management, among the Best Management Diploma in Kolkata and West Bengal, with Digital-Ready PGDM with Super-specialization in Business Analytics, PGDM with Super-specialization in Banking and Finance, and PGDM with Super-specialization in Marketing.
