Demand forecasting is the kernel of stable food supply for every supermarket retailer. It becomes even more sticky when it comes to fresh food and perishables, as planning and forecasting demand accurately impacts the health of the supply chain, creating an irreversible dent on sales revenue, operational targets, inventory levels, cash cycles and upstream/ backstream activities.
Infact, a recent study shows that out-of-stocks account for $634 billion for lost sales globally, while overstocks create $472 billion in lost revenues due to markdowns.
With fresh produce accounting for up to 40% percent of the food retailers’ sales revenue and with the limited shelf life of just a week, perishables can pose a forecasting nightmare to demand planners with a need for granular and accurate daily forecasting.
Why Can Traditional Demand Forecasting Play Spoilsport?
Why Can Traditional Demand Forecasting Play Spoilsport? For years together, food retailers have relied on a myopic traditional demand planning and forecasting rule book which is limited in scope and inflexible as it is usually time bound and time lagged. Agreed, that this well-established technique has been somewhat helpful when dealing with stable and predictable supply chain trends.
However the post-COVID economy is anything but stable or reliable and the demand for ultra-fresh produce has been oscillating like a pendulum. A situation where grocery shelves have seen fresh produce flying off the racks within hours, needs a demand forecasting technique that is dynamic, frequent and most accurate to track real-time market trends.
Therefore, historical time bound demand and sales data which is guided by the long term forecasting approach may not be relevant to get accurate demand forecasts for fresh consumables.
Here are the 5 main challenges that food retailers need to overcome to ensure free flowing demand for their fresh food cycles:
1. Lack of Adequate, Accurate And Timely Demand Data:
Data is to a demand planner like gold is to a jeweller. Demand planning without data can be futile. Traditional demand forecasting falls short of this vital aspect as it stretches over a longer time period which can dilute the essence of real-time data tracking and near-term visibility. Further, traditional methods of stocktaking can be inaccurate and the data is often outdated. This makes demand planning even more difficult a task when there is a lack of visibility across inventory levels, supply chain levers and can pile up unnecessary excess stock levels or they can have stock outages if demand spikes for specific products.
2. Signal Vs. Noise:
Food retailers very often get misled by the slightest abnormal shift in demand, which if not explained can be treated as a signal for change in planning and forecasting models. Establishing the right cause and effect relationships between abnormal shifts in demand that may not repeat itself until the same conditions are met that caused the shift on a continuous basis, is critical to interpreting the trends appropriately and driving correct strategies to accommodate the change.
3. A Time Stretched Forecasting Horizon:
How far into the future do food retailers want to extend their forecasts, depends on the type of commodities and customer behaviour. For fresh produce, forecast accuracy is highest when done for the shortest time intervals as this diminishes in accuracy with long term future predictions. Food retailers should therefore choose shorter and granular data which is refreshed each time there is a change in customer behaviour or market trends, all of which lacks in traditional demand forecasting methods.
4. Confusing Correlation With Causation:
Often traditional demand forecasting can create an overlap between correlation of demand patterns and the causes of fluctuations in demand. At times, food retailers may use techniques to extract customer behavior patterns from correlations in demand shifts and link to external events. This may cause them to think that this is linked to demand shifts and believe it is the true cause of it.
5. Dealing With Product Markdowns:
Most food retailers use product markdowns for fresh produce to avoid waste, creating a risk of training shoppers to wait for the discount and training forecasting systems to buy. Legacy forecasting systems don’t account for price elasticity and miss out the actual price paid for the produce, especially when it has one day of shelf life left. Food retails wouldn’t want to replenish these fresh goods that are being marked down at large discounts, just to avoid wastage. This can impact accuracy of forecasts while resulting in out-of-stocks and large scale wastage.
Demand Sensing – Towards a Fresh Perspective in Forecasting
Traditional foresting systems need a facelift to improve accuracy and decision making when it comes to forecasting demand in food retail. This is where the promise of Demand Sensing steps in with a premise that- if the past is the best predictor of the long-term future, then the very recent past is the most accurate predictor of the very near future.
With Demand Sensing food retailers can do a great job in planning and responding to short-term changes in demand, as it is sensitive to sudden, immediate and real-time demand fluctuations that most traditional approaches fail to respond to.
It is time food retailers embrace a holistic and intelligent Demand Sensing and Forecasting solution to achieve a sound balance across the shop floor and top floor operations. The ELI Demand Sensing Module is your data-driven go-to planning and forecasting tool to accurately analyze and correlate demand insights and respond to fluctuating margins and inventory needs.
