RETAIL AND THE NEW NORMAL PHENOMENON
A new macroeconomic context with constantly evolving consumers: AI as the solution
Conad Centro Nord (one of our customers), using Delphi—an AI-based platform for retail point-of-sale management—achieved a 46% reduction in stockouts while simultaneously lowering inventory holding costs by nearly 9%."
The past four years have been quite exceptional for the grocery retail sector. While data from 2020 generally show an increase in sales value, a closer look reveals that, while during the COVID period volumes grew by up to 5.1%, they dropped by 1.7% in 2023 due to inflation. This market transformation requires close monitoring, as when consumer behaviors shift radically, almost overnight, the implications for retail can be dramatic.
Now, after years of instability, there is talk of returning to a "new normal." But what does this mean for the grocery retail sector?
Although the lesser impact of inflation and the resulting decrease in prices is currently a consistent trend, this doesn't necessarily indicate a return to pre-pandemic conditions. The past four years have taught us that consumption patterns can change suddenly, making it essential to adapt proactively rather than reactively, in order to gain a competitive edge in the market," explains Giulio Martinacci, founder of Tuidi. There’s no room left for those who believe they can remain indifferent to external factors affecting the evolution of the industry; ignoring how these truths impact daily processes could be a fatal economic mistake.
New Consumer Needs, Rising Production Costs, and Technical Challenges: Retail's New Reality
As we move into the 2024 season, some of the major challenges facing the grocery retail sector include:
- New Consumer Demands: Buying behaviors have shifted, and consumers have become more demanding. They now expect greater variety in product assortments, with new product lines like gluten-free, vegan, and high-protein options. For example, since 2019, the assortment of different types of milk has increased by 18% across various customers (according to our internal data).
- Rising Production Costs: One lesson from 2023 is the importance of controlling excessive costs, which significantly erode margins. Without effective cost control, margins shrink, limiting the ability to offer promotions or price reductions. This lack of flexibility makes it difficult to find new pricing strategies.
- Technical Challenges: In a world where understanding the consumer and identifying economic inefficiencies relies heavily on data analysis, it's crucial to have the right technical support. However, many Italian companies face a talent shortage in the IT sector, with 6 out of 10 companies reporting difficulties finding the right specialists.
Rebuilding from the Consumer with Artificial Intelligence: Delphi's Predictive Power
The grocery retail sector is at a turning point. While retailers have traditionally prided themselves on their selling skills, it's now imperative to enhance this role with data-driven insights. How much to buy? At what price to sell? Which products to display? These questions need to be answered with scientific precision. Giulio Martinacci emphasizes the importance of shifting the focus to demand forecasting with a new approach, starting directly from the point-of-sale receipt.
In the past, the lack of proper calculation tools limited the understanding of consumption changes, and decisions were based on simple statistical averages. In 2021, for example, when we worked with food chains, the management of fresh products (particularly fruits and vegetables) and promotions relied heavily on human experience, as there were no adequate tools to counteract the unpredictability of daily consumption changes.
Today, that challenge has been overcome. Companies now recognize the importance of investing in data management (often referred to as the "new oil"), data processing (the deep, customized analysis of customer datasets), which could be considered the "fuel" of a business, and artificial intelligence, which serves as the engine that fully leverages these resources.
What is Needed Moving Forward?
A paradigm shift: it’s no longer the human expert who determines the impact of events based on experience. Artificial intelligence does this by reading, analyzing, and interpreting millions of data points and hundreds of variables that affect sales daily in unique ways.
We have a particular case study," says Giulio Martinacci, "that highlights why AI outperforms traditional manual parameterization in terms of accuracy compared to commonly used statistical averages. For one client, lemon tea sales increase by 61% when on promotion. However, this figure turns negative (-44%) if the introduction of a new brand of tea causes a cannibalization effect. What if this occurs during the summer? In this case, sales increase by 76%, but if bad weather, such as rain, sets in, sales drop by 30%. The data and results described are representative of specific cases analyzed, but different scenarios could lead to varying outcomes, underscoring the importance of customizing predictive models to the unique characteristics of each context.
In a retail environment with tens of thousands of SKUs per store, where market conditions change daily, leading to shifts in customer preferences, and constant assortment turnover, trying to manage everything without a scientific approach is the root cause of economic inefficiencies. Today, there is finally a solution. The only source of truth is the receipt: when and what the consumer wants, in what quantity, and at what price. Building a system of accurate forecasting based on this information is crucial.
In this context, we developed Delphi, an AI-based platform that works with real-time data streams to anticipate consumer needs, starting from point-of-sale receipts, in order to better understand their preferences. This approach optimizes the entire supply chain, creating new collaborative relationships between stores and distribution centers," says Giulio.
Grocery Retail Moving into the Future: The Case of Conad Centro Nord
Even in highly complex organizational settings, innovation is possible and beneficial. The complexity inherent in a large organization like Conad Centro Nord can often hinder the adoption of new technologies. However, the retail giant challenged this status quo by embracing Delphi and artificial intelligence as a catalyst for change, putting the customer at the heart of its initiatives.
“We started by managing promotional products, trying to prevent randomness caused by external factors by studying different promotion types and store locations (from Lombardy to Emilia), and understanding the impact these promotions had on other products, such as cannibalization or spillover effects. Creating custom predictive models for each store led to tangible results: Conad Centro Nord saw a 46% reduction in stockouts of promotional products, while also reducing warehouse inventory costs by 8%,” says Stefano Elli, Director of Innovation at Conad Centro Nord. “To appreciate the impact of our new demand forecasting approach, just look at the results from promotional management. Delphi has elevated our organization: not only do stores benefit from more accurate stock management, but the central office has become more efficient in logistics by avoiding sudden, unexpected requests from stores. We’ve been on this path for years, and AI has been extremely helpful in understanding daily consumption patterns and predicting future shifts. That’s why Conad Centro Nord has placed Tuidi's demand forecasting at the core of its decision-making processes.”
Technological Innovation in the Discount Sector
It’s not just large grocery chains adopting this new approach to distribution. Regional and provincial discount retailers have also realized the potential of demand forecasting at the point of sale. When combined with increased operational flexibility, the result is a game changer. Take, for example, a retail chain in Puglia with 15 stores. Using artificial intelligence, they forecast daily demand for each product at each store, allowing them to determine how many units of each product should be on the shelves.
What happens, for example, when some stores have excess stock while others face shortages? With precise knowledge of future needs, stock can be redistributed between stores. Although transferring stock between stores is generally considered economically inefficient due to high logistics costs, this retailer has found that, thanks to Delphi’s daily sales forecasts, the accuracy is so high that it’s possible to manage not only fresh produce but also to transfer products, like meat.
This demonstrates the power of AI-driven demand forecasting: solving an age-old problem in retail, like buying in bulk and selling in individual units.
The conclusion is simple: knowing the daily demand for each product at each store enables retailers to make informed decisions that impact purchases, sales, logistics, and marketing, all based on the real-time preferences of the end consumer.
“Conad Centro Nord is a pioneer in this regard,” concludes Giulio Martinacci. “With the development of Delphi, it has proven that it’s now possible to avoid being caught off guard by unexpected market fluctuations, generating significant economic optimizations that have dispelled any doubts about the introduction of technological innovation, even within a large company.”