To stock or not to stock or how much to stock is a perennial conundrum every fast-food chain or Quick Service Restaurant (QSR) faces each day. Striking a perfect balance between demand and available stock is easier said than done. While under-stocking is a sure-shot recipe for losing customers and brand reputation, overstocking ends up locking up operating cash, taking up space, and often result in expired or obsolete stock that strains the bottom line. In other words, money is now stuck in the inventory, when it could have been returned as free cash flow or invested towards growth.
These global food outlets generate and accumulate large volumes of data about their inventory and supply chain. The complexity and volume of data make it challenging to apply traditional BI and analytics to gather insights. Advanced technologies using Artificial Intelligence (Ai) and Machine Learning (ML) can enable QSRs to drive tremendous efficiencies into their supply chain operations, inventory planning and projections, and forecasting future demands.
A typical QSR has thousands of suppliers who provide food, equipment or ingredients, packaging, and many other supplies to run their operations. Managing such a vast set of suppliers and supplies is always daunting. High freight costs, uneven stock distribution leading to severe stock volume imbalance across stores, and inadequate optimization of their current SKU size from warehouse centers to city stores put tremendous strain on their supply chain network.
How can a QSR break away from fixed weekly refills across all categories? In the case of logistics, how can these outlets go beyond their shipping schedule and optimize shipping based on what is being shipped and where? Interestingly, the data to address all these challenges is already available but lying scattered in disparate and distributed systems. Also, the available data is too complex for humans to comprehend, analyze, or draw insights. To make matters worse, the QSRs often lack the right tools or technological expertise to translate raw data into invaluable business insights to predict and plan effectively.
Businesses today are drowning in data. The question often is how to extract insights from real-world messy and unruly data from disparate sources that have been accumulating unceasingly over the years. The key is to identify relevant data that matters, enrich it, and then place a powerful layer of intelligence on top of this chaos. This intelligence enables us to remove inefficiencies, increase productivity, help deliver higher revenue, and pick out potentially wasteful expenditures.
Take for example our client, a leading global QSR, that owns multiple fast-food brands and a chain of stores worldwide. ElectrifAi helped the company harness the power of Consequential Ai and ML to reduce costs and risks and improve operational efficiency in their supply chain operations.
Here are two ways we helped our QSR client to enhance their forecasting capabilities, manage demand and supply unpredictability better, and drive better efficiencies into their inventory and supply chain operations.
The client’s existing inventory strategy, as most companies do, is to set a static number on Safety Stock. In reality, the demand and supply variability will never be the same. Hence, the risk of having too much or too little Safety Stock stands exposed in such dynamically changing situations.
As we can see in the figure below, inadequate Safety Stock results in a backlog or out-of-stock scenario, which in turn impacts a company’s revenue.
Traditionally, there are several formulas to compute Safety Stock based on Service level and Z Score. However, these formulas assume that the demand variance follows a certain statistic distribution, which unfortunately is not always the case.
Safety Stock is all about placing a hedge against the risk—either demand risk or supply risk. The risk can be quantified by probabilities, which can be identified and recommended by AI and ML technologies. ElectrifAi’s ML model has the capabilities to identify your demand pattern and supply pattern, and thereby recommend the optimal Safety Stock based on each individual SKU’s history data. Not just that—as new data comes in, our ML model continuously learns itself and improves.
Through this engagement, we enabled the client to achieve an average 40% Safety Stock reduction, while maintaining the service level.
The client’s supply chain network setting was very complex with 30+ distribution centers and 10,000+ store locations. As a result, suppliers had to negotiate a complex network of routes to reach their destination, making it difficult to gain efficiency and mitigate rising freight costs without the help of ML.
ElectrifAi, in extensive collaboration with the client, analyzed the existing supply chain networks and established the metrics and goals. We leveraged the optimization algorithm to recommend the best routes with the right SKUs for the assigned shipment destinations. We also introduced new nodes (optimal-sized logistic centers) to the chain to rebalance the network and reduce costs when the demand was high and shipping distance was long.
Successful outcomes were predefined with metrics and key success criteria by the client. And the results exceeded expectations, with costs reduced by 7-8% across different cities. We achieved this by optimizing the distribution route and the distribution volume of supplies from existing warehouses to city stores. We also mapped out city-wise warehousing sites or locations to ensure better route optimization and lower operational costs.
Thanks to Consequential Ai, our QSR client was able to reap the following benefits:
• 7.73% reduction in freight costs by recommending switching SKU deployment (SKU-DC combination) based on demand pattern of the region.
• 13.58% additional reduction in freight costs by adding a Distribution Center in a particular city.
• Overall optimization of the supply chain network by better alignment of DC to store cities and a reduction in both number and distance of routes.
ElectrifAi: US’ leading ML solutions provider
ElectrifAi is all about solving high-value business problems for the C-suite at the Last Mile. We call this Consequential Ai, leveraging years of deep domain expertise and pre-built Machine Learning solutions to quickly drive top-line revenue growth, cost reduction, and operational efficiency. We work with Global 2000 enterprises, including several Fortune 500 companies, in a core set of verticals. Our clients see results in 6-8 weeks, transforming their data into a strategic weapon to drive enterprise value growth and profitability.
Our solution does not require investment in a new platform or infrastructure. Instead, we leverage the data existing in your system to power the ML models to deliver business outcomes.
We are the last-mile solution that sits on the top to solve specific business problems and bring about savings. Contact us to learn more!