Modern CPG Merchandising Takes Data, AI, and Meeting the Moment
Consumers can be fickle creatures, lending their loyalty to brands with the lowest price, the best deals, or simply the hottest-ticket item going viral on social media.
It’s an everything, everywhere, all at once landscape that makes it nearly impossible to pinpoint trends and jump on them before it’s time to move onto the next big thing.
While advancements in consumer insights have provided CPG brands and retailers with a looking glass of the drivers behind shopper behavior and trends, the image can be sharpened and personalized using merchandising tactics that fuse data and AI to truly embrace the why, the how, the when — and even the will do — of consumer actions and motivation.
It's about meeting the moment no matter where and how it's happening.
The Impact of AI-Optimized Merchandising
AI has turned merchandising into a real-time, shopper-driven process instead of something that’s reactive and based on outdated reports, Cem Kent, CEO of Harmonya, tells CGT. “Instead of just looking at historical sales data, AI surfaces patterns in consumer behavior, sentiment, and product attributes to help brands make better decisions.”
While AI can be implemented at all levels of merchandising, Holger Harreis, co-leader of McKinsey's global data initiatives, says the primary areas of application for artificial intelligence or GenAI are typically in product, e-commerce, and pricing and promotions — and the latter is where he’s seeing priority use cases that are delivering value.
“Applications include regular price optimization, running promotions that take into account local effects, assortment and availability effects, and markdown management,” says Harreis. “They also all get linked more and more with another top priority use case around personalized promotions."
Here he says it's really about finding out a customer's personal circumstances, such as items they may have purchased, personal interest, and size consideration.
The benefits typically translate to cross-selling and upselling, with revenue increases of 10% to 20% (or a 1.5 to 2 times improvement in conversion rates), says Harreis. Common use cases include the implementation of chatbots in e-commerce, which can be piloted and deployed in just a few weeks, and can help consumers in their buying journey, “guiding them to appropriate categories, sizes, and styles.”
Operationally, AI gives merchandising a speed boost — helping brands react to trends in real time rather than waiting weeks or months, says Kent. It’s also a boon to inventory processes, keeping assortments varied and relevant while also getting ahead of stockout or overstock challenges. AI systems can fortify demand signals, keeping companies informed of competitor moves, seasonal shifts, and even social media-driven trends.
Within inventory and assortment, there’s often 15-25% in savings due to reduced logistics costs, and 5% to more than 10% in revenue uplift due to the availability of merchandise being optimized for demand, which yields higher sell-through levels, per Harreis.
Related: Learn how brands like PepsiCo, Reckitt, Estée Lauder, L’Oreal, and more are modernizing their consumer data strategies with generative AI, data clean rooms, and loyalty programs, as well as exploring the potential of retail media as a way to leverage first-party data.
Get Away From the Guesswork: How to Modernize Your Consumer Data Strategies
Taking the Leap: The Southern Glazer’s Journey
While the benefits sound difficult to pass up, retail and CPG often find themselves behind the AI game compared to outside business markets. BevAlc, for one, is notoriously difficult to get buy-in for, says Alan Wizemann, chief digital officer at Southern Glazer’s Wine & Spirits — a primary reason the company sees it as an opportunity to embrace AI and leapfrog other industries.
Southern Glazer’s has been a proponent of unlocking value through insights, leveraging over 100 different data points to make recommendations on what its customers should carry to complement their businesses and complete their assortment, says Wizemann.
One strategy has included keeping a focused eye on inventory through AI tools that analyze sales data, online inventories, point-of-sale transactions, and then alert customers when they might need to reorder. “We have captured some key learnings about how product velocity changes and how online inventory can be a beacon to what is happening in-store,” says Wizemann of a pilot it rolled out.
The company couples this with a market intelligence approach that integrates multiple AI agents and machine learning technology to examine customer strategies and experiences from initial order to delivery and merchandising to develop a deeper understanding of behavior and create hyper-personalized supply chain strategies across its network of customers.
Progress in this space comes from a no-holds-barred, meeting-the-moment approach at Southern Glazer’s.
“We have developed a strategy that not only leverages advancements in the AI space, either with new models or processes, but can take advantage of them quickly without changing our overall approach or needing to adopt new tools,” says Wizemann. “It has taken us over a year to not only design this strategy and platform, but it has also allowed us to introduce these technologies internally as tools to help our employees reduce the fear of these new capabilities and provide access and training for these tools for use in their day-to-day work.”
What companies often encounter are challenges related to jumping into AI without first prepping the workforce or allowing for the unknowns of new technologies, he adds.
Getting Over the Hump
Implementing AI isn’t without challenges, says Mat Brogie, CEO of Repsly. Key issues typically revolve around the quality and consistency of data, and not having reliable systems through which they can integrate multiple sources.
AI is only as good as the information it’s working with, Kent emphasizes, and a lot of brands and retailers are still dealing with messy, fragmented product data that doesn’t actually reflect how consumers shop.
“If your product attributes aren’t accurate or detailed enough, AI won’t be able to surface the right insights,” he says, adding that another challenge is adoption, as a lot of merchandising teams are used to making decisions based on instinct or legacy category management practices. AI insights can feel unfamiliar and lead to resistance in trusting data over experience.
Read more: Stanley Black & Decker Bolsters Data Quality With PXM Investment
“Also, ensuring that the insights provided by AI models are clear and actionable for merchandising teams can be difficult, especially in critical areas like pricing or inventory allocation,” notes Brogie.
AI in Action: Personalizing Pricing and Placement
How can brands and retailers work together to create a product and promotion journey that gets the consumer engaged? And how can companies bring that same level of attractiveness to the e-commerce experience? A look at AI implementations in the field points to varied use cases.
Amazon: Sudip Mazumder, SVP and retail industry lead for Publicis Sapient’s North America business, looks to Amazon as a leading example of an AI ecosystem that drives sales with personalized product suggestions based on browsing history and past purchases. “They employ dynamic pricing, adjusting prices in real-time based on demand and competitor activity, while anticipatory shipping minimizes delivery times by predicting customer purchases and positioning items closer,” Mazumder tells CGT.
Hershey: The Hershey Company is using shopping behavior analysis and promotion feedback, intersecting the data streams to inform customer investment and promotional dollar strategies. To achieve this, it has built a proprietary solution that is currently active in 40% of the company’s U.S. portfolio and has resulted in 15% to 20% of Hershey’s digital transformation savings in combination with its new ERP platform.
Predictive analytics plays a major role, says Brogie, with analytics helping brands proactively manage their inventory, avoid unnecessary markdowns, and spend promotional budgets more efficiently.
Harreis says implementations can extend to generative AI as well, used for hyper-personalization purposes to present products to consumers in different contexts depending on their needs. For example, for in-app or social media promotions, companies might use GenAI to generate sports-related visuals for a pair of shoes if the consumer is interested in trail running. However, another shopper might see that same pair of shoes as a stylish addition to their urban wardrobe if that seems more fitting.
“This personalized content could be created instantly using GenAI, tailoring the presentation to each customer's preferences,” says Harreis.
In what might feel like a multi-channel game of Where’s Waldo, there are ways to bring the consumer to brands rather than wait for them to engage on their own. Online, AI tools can facilitate visual search and product discovery, allowing consumers to find items using images rather than text, says Mazumder. While in-store, it might power interactive digital assistants that answer questions, provide directions, and offer personalized product information.
When consumers do arrive in front of a brand, however, strategy matters. Once at the shelf, what journey will they be presented with?
“Instead of lumping everything into broad sections, AI can suggest cross-merchandising opportunities based on real buying behavior — like placing plant-based snacks next to functional beverages because people who buy one are likely to buy the other,” notes Kent.
Many companies are leveraging these technologies to create more value for their consumers and customers as a key driver to loyalty, adoption, and growth, says Wizemann. But that doesn't mean it's a level playing field.
"Companies in the space are trying to balance speed to market and cost, and they aren't winning," says Wizemann. "They are either set up incorrectly or lack investment possibilities. They are struggling to make pilots work at scale or to bring in the talent and technology needed to advance beyond the ideas that could move their business forward."
Advancements in this space will undoubtedly continue, but the companies more likely to win out will need to carefully select use cases, align them with business strategy, and focus on areas where they can have the greatest impact, suggests Harreis.
