Stock in. Stock out. Retail forecasting with AI

By September 11, 2019 FIND Papers
604FDA - The FIND Predictive Retail AI Blog

Stock in. Stock out. Retail forecasting with AI

The metrics of failing to forecast

$
0
B
US annual lost sales due to stock-outs
0
%
Frequency of stock-outs on store visits
0
%
Percent of Amazon revenue from stock-outs
9 min read

Forecasting demand is a crucial issue for driving efficient operations and meeting P&L targets.

September 11, 2019

This is especially the case in the fashion industry, where demand uncertainty, lack of visibility into historical data and seasonal trends typically coexist.

The effect of poor demand forecasting is stock-outs (too little inventory), overstocks (too much inventory), item obsolescence, damagingly poor customer experience, rush orders, inefficient resource utilization from replenishment to staff — all of which builds to a cacophony of disruption that bullwhips through the retailer’s organization.

This environment puts pressure on the retailer’s operations and marketing teams. These teams find themselves constantly in reactionary mode trying to put out the inventory fire-of-the-day as they are constantly confronted with too little or too much merchandise.  The effect is compounded with the number of locations, the seasonal nature of the apparel industry, and the requirement to sell through the existing inventory to make way for next season’s apparel.

To date, the industry’s response has been to offer deep discounts, open clearance centers, set up shop in outlet malls or in many cases burn or discard excess inventory into landfills.  The later solution is increasingly coming under intense scrutiny and is driving the industry towards becoming more sustainable.

Further troubling is the cost of lost sales due to stock-outs. Stock-outs are becoming an increasingly important issue for brick and mortar retailers with the acceleration of customer migration to online shopping and mobile. It is estimated that stock-outs cost US retailers direct sale losses of over $140B annually1. However, the loss compounds as shoppers, after a few poor in-store experiences, get comfortable to shop online.  This further erodes the customer’s connection to the in-person retail experience and leads to the ongoing loss of future sales. 

There are numerous examples of major stumbles by even the largest retail brands such as Walmart, BestBuy, and Nike to name a few. In the case of Nike, the company implemented a demand planning system without adequate testing. The result was ordering too many low demand SKU’s at the expense of ordering too few of the then wildly popular Air Jordan’s.  The company reported a corresponding loss of $100M. 

In this environment, the requirement for more accurate demand forecasting is evident. Case in point is when Nike began to work with Celect, an AI-driven demand forecasting company. Celect was tapped by Nike to address specific use cases and prove the effectiveness of AI in enhancing demand forecasting. Celect was just acquired by Nike. After raising more than $30M, Cruchbase pegged Celect’s post-money valuation between $100M to $500M.

FIND AI Whitepapers, Case Studies and Retail Analytics Tools

The proof is in the forecasting pudding.

AI is proving itself to be very effective in addressing apparel demand forecast planning. FIND is demonstrating how our planning recommendations would have captured up to $10M in unrealized gross profit on $130M in annual sales. Further, market timing is now aligning with the rise of AI. Apparel retailers are beginning to move from awareness to action — allocating resources to cleaning and visualizing their data as a first step to deploying more advanced analytics.

AI was front and center this August at eTail Boston with even large brands such as Newell Brands (think Rubbermaid, Marmot, and more) dedicating an entire presentation to their success in being able to simply visualize data. As such, FIND offers our clients a reporting module in addition to our demand planning. The ability for teams to see their data visualized and quickly receive merchandise demand insights based on predictive analysis is game-changing.

Moving forward.

Predictive analytics and the ability to make data-informed decisions from planning, buying, allocation and operation teams will be table stakes for modern retailers wishing to remain viable in an ever-increasing competitive landscape.  FIND is well-positioned to capitalize on this opportunity with proprietary AI technology and a team with experience in developing technology that integrates into an enterprise-grade retailer’s workflow and most importantly accelerates turns, creates high-margins and drives significant increases to top-line revenue.

 

Reference:  IHL Group Report.  Out of Stock. Out of Luck.