How Artificial Intelligence Optimizes Merchandising, Pricing, and Replenishment for Omni-channel Retailers

    
“At each step of the process, planners have access to accurate information which becomes an invaluable tool. Companies that have imbedded AI into their systems have experienced up to 6% improved gross margin and a 10% increase in sell-through,”
Steve Gordon, SVP of Business Development for Antuit

Delivering great customer experience – by providing their customers with the merchandise they want, when they want it, from the channel of their choice – has quickly become the defining way to do business in retail. Consumers have an almost unlimited supply of shopping options. From anywhere in the world, most individuals can receive their purchases by the same or next day. 

This radical, fundamentally different approach to shopping has created new challenges for retail planners and buyers who must now determine how to plan, merchandise, price and fulfill across a complex, omni-channel distribution system.

This fluidity of product movement generates hundreds of millions of data points. However, many fashion retailers continue to use basic statistical modeling tools (some still based in Excel) to generate forecasts and determine demand signals for merchandise planning, product pricing and replenishment.  These antiquated tools need a lot of human effort, are predominately based on non-imputed historical data only, and are unable to process the hundreds of millions of data points needed to isolate components of demand such as seasonality and trend, honing in on a highly accurate depiction of the demand signal.

Today, artificial intelligence and machine learning are used to solve many of these problems for retailers while minimizing carrying costs, increasing sell-through, and improving profit margins. For the purposes of this article, let’s explore how artificial intelligence can transform single season planning, buying, pricing, and fulfillment.

Pre-Season – Planning and Buying
During the buying season, merchants must determine what and how much seasonal merchandise to purchase to achieve their financial forecast. Before the introduction of artificial intelligence, merchant planners relied on disconnected systems, poor demand signals, and unreliable historical data, attributes and causals to make planning decisions, resulting in repetition of past mistakes or committing new ones.

Artificial intelligence, imbedded into the retailer’s current systems, can process large amounts of online and offline data to generate precise demand signals. AI can isolate demand preferences and profiles, by brands, silhouette, color or size at store or store group level, thus allowing for merchants to place buys based on science rather than simply last year or intuition. AI has the ability to deal with poor, sparse and missing data, a common problem in retail, with efficient models that operate at different levels and use a multitude of causals to break down demand accurately into its key components. By modeling demand and demand drivers at a granular level, companies can obtain a tighter alignment of inventory with financial goals by enabling planners to determine the accurate buy and allocation of sizes, colors, and quantities across multiple stores and warehouses for both in-store and e-commerce inventory.

“Today, artificial intelligence is revolutionizing how companies optimize their assortment and allocation decisions. Never before have companies had the ability to efficiently model and predict demand and demand drivers using granular data at the lowest level,”
Yogesh Kulkarni, EVP of Marketing and Pricing Analytics for Antuit.

In-Season – Pricing and Fulfillment
The ripple effect of accurate pre-season buying and allocation reduces many problems by stocking the right product, in the right quantity, in each location at the beginning of the season. This initial allocation accuracy needs to be extended in-season by using pricing and fulfillment intelligently to avoid out-of-stocks and carrying costs.

Miscalculations for promotional, markdown and omni-channel pricing can cost companies millions of dollars in lost profit. The benefit of infusing artificial intelligence into lifecycle pricing is that these challenges can be greatly reduced or eliminated by optimizing their pricing strategy across products sold online, sold in store, or purchased online and shipped from in store. AI can also forecast omni-channel returns and take that additional inventory into account when recommending future markdowns. 

Additionally, many retailers will offer continuous promotions to drive store traffic and ensure a steady stream of consumer shoppers. This creates a huge challenge to calculate potential promotional lift on future promotions with any degree of accuracy. Artificial intelligence provides retailers the means to achieve an accurate picture of their baseline demand. This is imperative to test “what if” options to make the best decisions for the profitability of seasonal promotional markdowns and the overall impact on gross margins.

Replenishment has its own challenges. Artificial intelligence can be deployed across hundreds of millions of product and location combinations, consuming sales and returns data points in real-time, adjusting to provide in-season fulfillment recommendations. With increased efficiencies and accurate replenishment decisions, companies experience reduced carryover that can significantly impact profit margins and sell-through. 

Conclusion
Retailers who augment their systems with artificial intelligence and machine learning understand the fundamental importance of transforming their business to the expectations of their consumers today. Adapting an AI strategy helps them deliver excellent customer service by providing consumers with the merchandise they want, when they want it and from their preferred channel. Using AI-powered solutions, retailers can solve many of the challenges they previously encountered, giving them a market advantage over the competition.