The cognitive retailer

How 3 retailers leveraged data and AI across merchandise planning, forecasting and distribution

Retailers find themselves in a time of extraordinary disruption and opportunity. By definition, that means that historical data—the buying and behavior patterns on which demand forecasts, assortment and allocation plans are typically built—are of little help in merchandise planning, forecasting and distribution decisions. Perhaps as importantly, data gathered throughout this past year will reflect the industry’s highly unusual circumstances, which will impact the accuracy of future plans.

To operate in today’s environment and ensure their place in the future, retailers need to not just adapt, but also act—by which we mean shift from a reactive mindset to one that is proactive, predictive and, ultimately, prescriptive. In this paper, we explore how retailers can take control of their future by advancing their digital capabilities, including the use of data, analytics, AI, automation and cognitive computing, in order to better anticipate and respond to trends, shape and influence events, identify exceptions and outliers and leverage timely insights to drive standard decision making.

Creating the cognitive retailer

Retailers face two basic demands: 1. Ensuring they have the items that consumers want in stock at the right level; and 2. Ensuring those items can quickly reach consumers through a variety of highly-localized distribution channels.

This is a complex task at any time, but especially so in the face of disruptions as retailers temporarily close stores, operate with limited hours or reduce services due to health and safety concerns related to global or local events.

“It’s no longer just a matter of having the right products on our shelves, it’s about delivering an experience to a specific consumer,” says Lindsey Mazza, Global Consumer Products and Retail Domains Lead, Capgemini. “In omnichannel operations, data needs to guide every decision the business makes, from what to stock, to where to hold inventory, to how to fulfill orders. Even functions that are not consumer-facing, like warehousing, must be reexamined in light of this increased need to reach consumers when and where they are. A fully integrated omni-channel operation will improve consumer experience while keeping supply chain costs low.”

In this landscape, retailers must leverage data-driven intelligent automation applications, such as artificial intelligence (AI) and machine learning (ML), to modernize their existing planning, merchandising and logistics capabilities. Initially, these tools help streamline processes and analyze a bigger data set more quickly, thus giving the organization both time to respond to a situation and the ability to do so more effectively.  At the most advanced level, the systems can be self-correcting, self-improving and self-optimizing—meaning that the business can automate responses to pre-determined scenarios. Responses can also be a lot quicker as machines can process significantly more data than a person. Finally, the number of scenarios that can be accounted for is considerably higher, resulting in more accurate decisions.

For example, in demand planning, AI and ML can be used to review historical patterns, select best-fit statistical models and draw on a variety of inputs and forward-looking variables, such as promotional grids, sell out data and environmental factors, to create more accurate demand forecasts with less manual effort. As organizations continue to build maturity, they will also be able to support a planning system that is autonomous and touchless—meaning that it can identify disruptions, shortages, transportation delays and any other number of issues and then automatically re-plan to account for that event.

Benchmarking your organization’s digital maturity

Most retailers use what’s considered descriptive or diagnostic pre- and in-season planning today; planners are aware of what is happening in the moment and why it happened. Depending on how long the insight is relevant, planners are able to respond to trends or events that are relatively long-lasting. While many retailers are exploring the addition of predictive or prescriptive capabilities, they are doing so on a smaller scale, using data sets within select areas of the business but not yet across the entire enterprise.

However, as we’ve seen with COVID-19, even in events that span weeks or months, the ability to react can be hampered by any number of issues: stock outs, supply chain disruptions, delivery personnel shortages and staff availability to name a few.

While many retailers managed to adjust operations in response to COVID-19—be it in the form of new sales channels, delivery methods or management of inventory—the changes may not be sustainable.

“In the quest to serve quickly, organizations may create a solution that prioritizes speed over long-term success,” notes Chew. “In order to create that foundational responsiveness, retailers need to design with intent, thinking through how each function—from planning, to assortment, to logistics, to delivery—is connected and unified. The goal is for the business to shift in tandem, not manage disruptions in isolation.”

In order to advance on the digital maturity curve, retail organizations must first understand where they are—and determine what foundational capabilities must be enhanced in order to move forward. While every organization’s needs are unique, we have developed the following benchmark tool that uses four main criteria to help organizations assess their existing capabilities.

Predictive > Prescriptive > Cognitive: 3 stages of digital maturity

1. Predictive: Anticipating future trends and event

As organizations compile a more robust data set from which to draw, organizations approach the Predictive phase. In this stage, it is possible for the retailer to anticipate trends and events—and how they will affect the business—based on data and analytics.

“As the algorithms improve and as more data is entered into the system, the business will be able to see patterns that signal future events,” says Mazza. “Once again, with this information, organizations have the ability to respond, fending off challenges or seizing opportunities.”

For example, this past summer as consumers tired of being indoors and were reluctant to take long trips away from home, local travel increased as did the sale of outdoor recreational equipment in the U.S. However, many retailers were caught off-guard by the sudden demand since planning and merchandising decisions were based on historical sales data.

But with the power of AI, these companies could have considered a wider set of inputs, such as: social media keywords for camping, hiking or outdoors; weather reports indicating that temperatures would be unseasonably warm; a downturn in air travel; and even disruptions to school and work schedules.

Taken together, these inputs could have helped retailers realize that seasonal demand would far exceed that of the previous year. Based on these insights, organizations would then have a chance to respond by adding more stock, moving inventory around based on regional demand or creating promotions or offers in key markets.

2. Prescriptive: Shaping future trends and events

Organizations with high digital maturity are able to understand, anticipate and respond to trends and also shape them. We consider this ability the hallmark of the Prescriptive phase.

In keeping with our outdoor equipment example, retailers at the Prescriptive phase could “beat the trend” so to speak—essentially reaching out to highly targeted audience segments with promotions and offers to jumpstart the customer journey. With the right insights, the retailer could also develop content that would strengthen the customer relationship, such as downloadable maps of nearby national parks or biking trails.

Finally, retailers can also take this information and consider how to best address the market need. For example, stock of tents and other large items were limited this past summer, which meant that even if retailers sold every item, their revenues would be capped based on their inventory limits.

“In these cases, it’s important to think outside the box—or in this case, tent. We saw some quick-thinking retailers launch a rental or sharing model, which allowed them to serve a wider group of customers with limited inventory,” said Mazza. “Meanwhile, customers who may be new to the outdoor world could try camping without investing in a big-ticket item.”

At the same time, retailers who pursue such strategies must continue to be guided by that north star metric of health and safety. In this case, the retailer needed to ensure proper cleaning and disinfection of the equipment and actively communicate that process to the customer.

3. Cognitive: High-level reasoning that replicates the human thought process

Cognitive is the most advanced stage in our digital maturity curve.

Though cognitive is sometimes used interchangeably with AI, in scientific terms the difference between the two is quite distinct. The main difference is that AI is an augmentation tool meant to support human decision-making; cognitive computing, on the other hand, is an advanced technology that mimics the human thought process. The most prominent example of cognitive technology today is chatbots or other customer service channels, which are programmed to reason as humans do to respond to customer questions and issues.

“In the Predictive and Prescriptive phases, AI is a tool that helps people make better decisions, faster and with improved outcomes,” explains Mike McCullough, Supply Chain Lead, Capgemini Invent, “Cognitive computing is a capability that is almost akin to human reasoning that can alleviate some routine and repetitive tasks. The two technologies are similar in theory, but different in practice.”

Looking beyond AI-based chat features, cognitive capabilities can be used for a variety of retail planning capabilities.  Leading retailers are using automation to mimic decision making on stock intake, suggesting assortment extensions and edits and determining commercial benefits of expedited transport options.

Becoming cognitive: Building incremental value while accelerating the data strategy

Retailers that want to advance their data strategy should keep in mind that digital maturity is a journey they should jumpstart now.

“AI systems are based on algorithms that learn as more data and events are added, which means that they become more powerful over time,” says Chew. “This is why it’s so critical for retailers to start their journey now—because building digital maturity takes time. Delays now could mean obsolescence later.”

But while maturity cannot be achieved overnight, results can. Organizations that begin their journey can expect to see incremental gains at the outset of their program. And, because the digital solutions will become more precise and accurate over time, the impact will grow.

A focus on adoption of AI technologies by users is a critical step that cannot be overlooked.  Holly Nurse, Management Consultant, Capgemini explains: “Bringing real value through AI means incorporating both the art and the science, the knowledge and decision-making of the planner/merchandiser and the cognitive technology to provide insights, capturing outliers and trend spikes.”

Programs highlighting the importance of the user succeed in gaining trust and securing adoption of AI methods, driving business benefit. The journey to AI isn’t a person one day and a computer the next, but a partnership allowing strategic decision-making by our talented retail teams, enhanced by valuable insights.

“The ability to predict and shape trends is a capability that organizations need to build through the constant gathering and continuous analysis of data,” says Mazza. “In the meantime, the investment will pay for itself as the business is able to redefine core areas of planning, merchandising and distribution in a way that optimizes inventory, staff and other resources.”

The Solution: Smart Retail Planner from Capgemini

 Capgemini’s Smart Retail Planner is a holistic approach to addressing the two most pressing needs facing retailers via automating the full value chain: ensuring they have the right stock on hand and that they can quickly distribute it to customers across a variety of channels.

Leveraging our deep content expertise in retail merchandising and supply chain transformation, insights and analytics, artificial intelligence, Capgemini partners with some of the world’s leading retailers to help them adapt to the evolving landscape, deepen their understanding of their customers’ behaviors, and optimize their merchandising and supply chain functions, using data and analytics in a scalable, cloud environment.

Capgemini’s Smart Retail Planner is based on a proprietary approach that allows retailers to jumpstart results in as little as eight weeks and create a pathway to becoming a cognitive omnichannel retailer.


To learn more about how your organization can accelerate its journey to cognitive, reach out to our authors today.

Lindsey Mazza, Global Consumer Products and Retail Domains Lead, Capgemini
Joyce Chew, Smart Retail Planner Delivery Consultant, Capgemini

Jan Kazen, Vice President Consumer Product, Retail and Distribution, Capgemini Netherlands
Gabi Ledesma, Vice President, Global Retail and Consumer Goods, Capgemini
Sébastien Neyme, Director Supply Chain, Capgemini Invent France
Mike McCullough, Supply Chain Lead, Capgemini Invent, North America
Holly Nurse, Management Consultant, Capgemini Invent

Capgemini_The Cognitive...

File size: 1.18 MB File type: PDF

Meet our experts