The four steps to infuse artificial intelligence into merchandise planning

    

In recent news, Nike made a shocking announcement that they were acquiring Celect, a predictive analytics and demand sensing software company, to fuel its consumer-direct offense strategy. The goal for Nike, despite their brand innovation and trendy fashion, is to better serve their customers on a more personal level by leveraging an AI platform to become hyper-focused on consumer behavior. This point of differentiation was underscored by Eric Sprunk, Nike, Inc. Chief Operating Officer, who stated:

“As demand for our product grows, we must be insight driven, data-optimized and hyper-focused on consumer behavior.  This is how we serve consumers more personally at scale.” 1

For most, this news spurs a “call to action” to make their assortment planning and consumer buying decisions more analytically sophisticated, blending art and science, to answer the age old question – what, where, when, and how to best satisfy my consumers’ demand while maximizing my profitability potential? 

One of the key reasons Nike acquired Celect was to improve the Assortment Planning process.  This foundational practice is the cornerstone of any organization to improve purchasing of products, align to latest market trends, and ensure the purchase quantity meets consumer demand. With an omni-channel focus, more retailers are relying on planners, buyers, and allocation teams to grow endless aisle experiences, localize assortment decisions, and understand the role of each category seamlessly.

At Antuit, we value the power of AI and machine learning to transform business processes like Assortment Planning.  To achieve this, we infuse analytics into the planning process for our clients to provide an automated and scalable solution in four simple steps:

  1. Perfecting demand enables our clients to identify the ‘sweet spot’ or best timing that maximizes consumer demand by addressing the “when do we purchase” question. Perfecting demand also identifies significant gaps in the assortment mix leveraging market data and historical sales to begin shifting the plan to better align to consumer shopping behavior versus one-off market trends.

  2. Intelligent Clustering leverages various clustering techniques like DB Scan, K-means, and Doughnut to analyze the demand patterns observed in each location. These demand patterns help define store clusters in a volume independent approach that identifies why similar category sales patterns occur across high volume and low volume stores. For example, stores in eastern Atlanta (small volume) may behave like stores in northern Los Angles (high volume) due to younger shoppers, middle income, price conscientious, and suburban centers that drive market trends for a specific commuter pant for daily work and leisure wear.

  3. Assortment Profiling & Assortment Optimization exposes the most significant attributes of each product that drives consumer shopping behavior and category performance within each cluster. Using consumer decision trees (CDTs), we can define the rank of attributes to better inform the buyer and planner to what role each purchased product will play within a cluster. Using techniques like CHAID, we show how the CDT meets demand to drive revenue, maximize sales, or earn profits based upon the mix of products for each shopper persona by cluster. Optimization evaluates the outcome of each CDT to determine the best choice count to maximize each goal while minimizing risks to inform how much is too much to carry for each cluster or location.

  4. Assortment Plan & Allocation Profiles provides the final automated output that merges all the analytical insights into a pre-populated assortment plan and allocation profile that aligns buyers’ budgets and allocators’ space constraints. This will assort and distribute the depth and breadth of each choice to the best store(s) to maximize profits while aligning to consumer demand.

This four-step approach is exactly why companies, like Nike, are bringing advanced analytics in-house. Actionized, a retail advisory firm and partner of Antuit, leads this change management process to effectively operationalize these steps to give companies a “Customer First” approach to product development and alignment to “Voice of Customer” locally.  These consumer centric insights enable Merchandising and Planning to make more informed decisions that reduces disparity between company strategy and consumer demand.  Eric G. Prengel, Managing Partner, Actionized, articulates this point clearly:

“Without the use of AI based analytics infused into your Assortment Planning Process, it is impossible for a merchandising team to identify what is truly important to a customer and what trends are driving their decision making. It’s time for retailers to add insight to their already complex design and buying processes’”

By automating and scaling this process for our clients, decisions are made faster with less risk across millions of decisions in a matter of seconds, placing the art of decision-making back into the hands of those who love buying and planning most.

1 Nike News, “Nike, Inc. Acquires Data Science and Demand Sensing Expert Celect”