As competition in the market increases from the number of competitors to the number of products being rushed to market, a challenge many large retailers face is knowing what to order, when to order and how much to order meanwhile being keenly aware of the penalty when a product isn’t available. This ordering dilemma is on full display in the grocery business where knowing what, when and how much can mean the difference between a profitable month or not.
Fresh food products can account for 30% to 40% of a retailer’s revenue. Having too much can cost a company in terms of spoiled or expired items that have to be thrown away. Having too little may impact a retailer’s customer base whose loyalty to a store may run thin after a few futile trips to an empty refrigerator bin.
Given the razor thin margins in this hyper-competitive environment, retailers are turning to Artificial Intelligence (AI), more specifically Machine Learning (ML), to be able to better predict the what, when and how much. Rule-based systems have become ineffective. The ‘if this, then that’ rules may be sufficient when managing stable, non-perishable products (e.g., tires), but fail dramatically when it comes to perishable products (e.g., food).
The Fresh Food market is a difficult space to compete in when your inventory has a short shelf-life and the demand for each item varies given the time of the year, weather, and other external factors. Add to this volatility the unexpected delivery times of getting a product onto the shelf and it becomes clear that a more intelligent prediction system is needed.
Using Machine Learning in a high-transaction business with thousands of Stock Keeping Units (SKUs) and a variety of external variables is a natural choice to calculate when to replenish stock. Being able to predict what to have on the shelves based on historical sales, seasonal trends, holidays and unexpected events (e.g., hurricanes) can improve a company’s profitability by reducing waste.
On a humanitarian note, Microsoft, Amazon,and Google have begun to work with international organizations to predict when famine will hit a developing nation using ML. The Famine Action Mechanism (FAM) will alarm when a food crisis is predicted which in turn will trigger donation drives and the appropriate actions to get food quicker to those countries. “If we can better predict when and where future famines will occur, we can save lives by responding earlier and more effectively,” said Brad Smith, President of Microsoft.
Here are some interesting and eye-opening facts about food wastes from the Food and Agriculture Organization of the United Nations:
AI is in a position to help retailers reduce food waste to increase their profitability while on a global scale it will help ameliorate famines and starvation in less developed countries. AI can help feed the world more effectively!