Retailers Turn to AI to Assist with Forecasting Challenges

Share this:

When the world has seemingly turned upside down in the midst of a global pandemic, predicting how many units of merchandise you’re going to sell next month can seem trivial. But to retailers and CPG brands faced with navigating an extraordinarily challenging year, being able to accurately predict and plan for demand has been paramount.

Retail forecasting has always been challenging, especially given the rise of e-commerce and other outside competitors disrupting the traditional market. However, those challenges have been compounded in 2021, with widespread uncertainty around profitability and customer service. Many retailers are struggling to maintain inventory and meet customer demand. Finding ways to accurately, and efficiently, handle forecasting and demand planning has become a top-line priority.

“Covid-19 was a key trigger for the retail and CPG industries to switch to AI when it comes to forecasting business,” says Amitabh Bose, chief practice officer at Fractal Analytics, a firm that provides AI solutions for the retail industry. “Retailers and CPG brands didn’t just embrace AI to develop better forecasts for their existing channels like store sales but also for their accelerated channels of digital commerce – where demand boomed as a result of the pandemic.”

With machine learning and AI, retailers have been able to navigate the continued imbalance between supply and demand. This is especially true for digital-first retailers and on-demand businesses. The number of online grocery shoppers increased by 35 million during the pandemic. That opened the door to new opportunities, but it also opened the door to certain logistical challenges that grocers never experienced previously. For example, online grocery shoppers expect the items in their mobile apps to be in-stock and available for delivery immediately, which is different from a shopper who casually browses store aisles to see what’s available at a brick-and-mortar location.

Next-Gen Forecasting

Next-generation forecasting platforms learn from recorded data using AI, patterns, and predictive insights. Bose says the retooling in forecasting strategies that occurred during the pandemic actually happened twice for the grocery industry. In the summer of 2020, when forecasts were turned upside down due to Covid-19, forecasts had to be retooled by experimenting with AI algorithms that were providing better forecasts than the conventional time series models that grocery retailers had typically used.

Shoppers eventually became more accustomed to the “new normal” as 2020 progressed, but then grocers had to retool their forecasting strategies once again in the fall. During this second retooling, they focused on different parameters, like recent sales, new product, and end of life product strategies. The approach was meant to better align grocers with the conditions of the industry that existed at that specific time.

“These new AI models were built around experimentation with various parameters such as mobility data, hot zones, unemployment, and competitive pricing indices and feature engineering,” Bose says.

Given the significant movement to online buying within the pre-millennial demographic — who have transitioned directly from in-store to mobile app purchasing in a “native” way — Bose says more retailers are now looking to use AI to build tools and strategies that can retain these types of buyers for the long- erm.

“We are seeing companies in the industry embracing AI to not just track demand and plan for it, but to also shape demand itself,” he says. “This includes tying AI insights in with retail marketing platforms to provide shoppers with a seamless online-to-offline marketplace to consumers that is shaped to them.”

Although we’re now in a place with the pandemic where vaccine distribution is growing and case counts appear to be dropping in many places, retailers aren’t in the clear just yet. While AI is helping to alleviate large-scale forecasting and demand planning challenges, it would be a mistake to assume that this technology alone is a cure for all that ails the industry.

“I believe we are at a pivotal moment in the industry as AI will become the mainstream forecasting engine — albeit with oversight from human demand planners,” Bose says.

Stephanie Miles is a senior editor at Street Fight.

Stephanie Miles is a journalist who covers personal finance, technology, and real estate. As Street Fight’s senior editor, she is particularly interested in how local merchants and national brands are utilizing hyperlocal technology to reach consumers. She has written for FHM, the Daily News, Working World, Gawker, Cityfile, and Recessionwire.