From Famine to Feast: Grocers Are on the Edge of a New Data Bonanza

A person pushing a shopping cart in a grocery store in NYC
Credit: Andrew Kelly - Reuters /

Mindaugas Eglinskas, co-founder and CEO of Pixevia

Grocery retailers have long recognized the value of data in understanding their customers better and improving operational performance. However, they have faced challenges in either not having enough usable data or having so much that they don't know what to do with it. It has been a case of famine or feast in terms of data availability.

Now there is an emerging source of data that not only solves the famine experienced by some retailers but also holds the potential to create a feast of data that gives much deeper insights into customers’ in-store decision making and to transform how retailers operate.

Seeking out data

Today, grocery retailers leverage data collected directly from customers through interactions such as POS purchases, loyalty programs, and household surveys to create personalized experiences and recommendations. This kind of data is more reliable, accurate, and compliant than that collected through external sources such as cookies, web browsing, and advertising. After all, the latter doesn’t tell you much about what’s happening in physical stores: real stores only have real cookies. 

Grocers are avid users of data to optimize their inventory, pricing, promotions, merchandising, and customer service, as well as to make progress towards building trust and loyalty with their customers. However, conventional data sources only tell retailers what products are purchased, not how the products were discovered. 

Conventional methods of trying to understand purchase discovery decisions are often crude. Standing on the store floor to observe purchasing patterns is time-consuming and likely to be small-scale, snapshot, and anecdotal. Consumer surveys or ethnographies are self-reported and point-in-time. Retailers are left without the sample size and objectivity necessary for real insights.

Autonomous stores open the lid on a new data trove

The emergence of AI-powered, autonomous stores, where customers can walk in, pick up what they want and walk out again - all with cashierless checkouts - changes the game. Autonomous stores leverage video and sensor technologies and AI to understand what the customer is doing, what they have chosen and how long it took to choose a product. With AI-enabled analysis of in-store video/sensor data, retailers can go beyond POS information and gain access to a scalable, data-driven view on product interactions across the store.

How data analytics can transform the shopping experience

With this new source of data, retailers can understand so much more about shoppers' decision-making process. Analytics applied to the data can discover traffic and path-to-purchase trends across the store and by shelf, fixture, and product category.

It can provide retailers with new insights on how products are considered, picked up, returned to the shelf, and purchased across their autonomous stores. It’s early days, but in time AI will learn how to leverage this information to help retailers and manufacturers improve product assortment and placements, drive operational efficiencies by refining inventory levels and shrinking store waste, and enhance the store experience for consumers.

Labor efficiency, for example around restocking shelves, can be tracked alongside traffic and consumer engagement trends, optimizing staff scheduling in busy periods. Out-of-shelf or planogram noncompliance alerts can be raised if such cases are detected. Data analytics can also help chains ensure uniformity in operations across all locations and maintain the brand standards that customers trust. 

All this is possible and, although it is an emerging science, the raw data now exists to make progress.

Retail’s progress report 

Several retailers are already utilizing data analytics to increase efficiency and customer experience. For example, stores leverage video and sensor technologies to understand what the customer is doing and what they have chosen. Analytics have helped the retailer to understand customer behavior and preferences in real-time. This information allows them to optimize product assortment and placements, build basket sizes, and streamline operational efficiency.

To give a quick example - a shopper spends 5 minutes near the sauces section with pasta in the shopping basket. They pick up and put down four different ones, but in the end, they don't choose anything. Most likely, this category in the store needs special attention, and extra SKUs should be added here, especially if pasta is a hot seller in this store, and there is a pattern with pasta buyers. For different categories, it's possible to calculate average decision time, based on items in the basket or other variables.

When you know the situation on the shelf in real-time, you can better understand the customer's choices. For example, customers choose to buy less-known water brands only if you run out of the well-known ones; otherwise, they stick to the main brand. The decision time is longer when the preferred item is absent and quicker when they instantly see what they would like to buy. It can seem like the store still sells other brands, but it can be only when the main ones are absent - it's better to have a bigger stock of preferred customer brands.

From experimental to mainstream

If this sounds like a far-away future, according to PwC the global number of autonomous stores is expected to grow at 91% CAGR and may reach $400 billion in transactions in 2025. The technology addresses several strategic issues facing retailers, including the scarcity and high-cost of labour, customers’ declining patience with checkout queues, expensive city center rents, and excessive shrinkage in city centers. 

As shown here, it also generates a new strategic source of data that is spreading rapidly through retail estates. It is essential for retailers to learn how to apply data analytics to this new source of data and deploy the insights it is creating. The leaders in the industry are already doing so and gaining an advantage in the market.

About the author:

Mindaugas Eglinskas is the founder and CEO of Pixevia. A former assistant professor of machine learning and robotics at Vilnius University, Eglinskas was working on real-time payment systems at the Central Bank of Lithuania in 2016 when he saw the opportunity for major improvement in the grocery shopping experience and in retail management using AI. With a core startup team he built an office-based MVP to test the technology in a simulated environment before Pixevia launched its own store in 2019. He believes technology can bring about a breakthrough to the retail experience and operations and aims to build Pixevia into a global technology leader to help achieve this.

The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.

Other Topics