Demand Forecasting

How To Ensure More Accurate Demand Forecasting

Demand Forecasting

How To Ensure More Accurate Demand Forecasting

As an industry, online retailers have no need to worry about demand for the upcoming holiday season (supply may be a different matter). After all, Cyber Monday 2020 posted the biggest ecommerce day in history, with sales of $10.8 billion. At an individual store level, however, there are plenty of concerns to address. 

The current approach to demand forecasting isn’t working. It fails to deliver the tools and insight that retailers need to capitalize fully on the holiday season, or the retail year as a whole. That could leave retailers who rely heavily on holiday sales (typically around 20% of annual revenue is from Black Friday/Cyber Monday) under- or overstocked at critical moments. 

Here’s how online retailers can flourish with enriched product data backed by AI-powered customer intent data for more granular demand forecasting.

The Legacy Data Retailers Rely On Today

Too many retailers are still relying on historical sales data to plan for the seasons ahead, leaving them slow to react to evolving trends. As McKinsey notes, “Retailers are already wrestling with omnichannel’s demands on their supply chains.” At the root of the problem is a lack of actionable insight and granular product & customer data.

We’re no longer in a landscape where spring and fall fashion shows set the schedule for the year ahead. In today’s digital marketplace, trends can break out and gather momentum much faster than the traditional sales cycle allows. A single Netflix series, for example, can trigger a spike in demand for a look, culture or subculture that was entirely off the radar just months before. 

Similarly, the tools that retailers have at their disposal restrict them to broad segmentation (by age, gender, location, etc.) to forecast demand and establish revenue targets. As a result, the interactions that customers have with products on-site give little insight. Sales figures alone indicate only which units sold well, not why they resonated or whether they will perform strongly in the future. 

Retailers could look to real-time data for clues, such as customer reviews, yet most lack the machine-learning and artificial-intelligence capabilities to aggregate and analyze contextual data. Instead, they are stuck with objective sales figures only.

What Retailers Are Missing

As a result, retailers have little visibility into the context behind sales figures, and no insight into demand. That leaves them exposed to the following, particularly around peak holiday sales season: 

  • Discounting stock that could have been saved for full-price sell through 
  • Missing out on upselling opportunities to increase average order volume
  • Witnessing even higher cart-abandonment rates, bearing in mind that an 80% abandonment rate is already common in fashion 
  • Processing high returns, which are already breaking all-time records in retail at around 25% 
  • Losing out on the chance to turn holiday shoppers into return customers

The Data That Ecommerce Retailers Need To Forecast Demand

With enriched product data at their fingertips, retailers can see the full picture at last. Granular product attributes can articulate why customers want an item, what they might also like right now and what they will be looking for next. Crucially, retailers can appeal to individual shoppers with hyper-personalized product recommendations based on psychographic profile, rather than generic suggestions informed by demographic. 

As always, the clues are in the context. Conventional data, for example, might reveal that knits sold well last year but not understand why. A deeper look at the product attributes would reveal that it was an overarching trend for nautical themes that was triggering demand, opening up the potential to drive demand for similar items in the future.

How Lily AI Enables Smarter Planning, Buying and Allocating

We have reached the point where traditional allocation planning tools are too crude for the modern digital retail landscape. They don't provide enough granular product attributes to work with, forcing retailers instead to take a shot in the dark with new launches. 

Lily AI allows retailers to go deeper into demand, with the power to mine more than 15,000 product attributes and 1 billion data points for insight. And it’s not just about the numbers. The platform can even identify products that are visually similar to items that perform well, to support improved trend forecasting and planning. Retailers can integrate that enhanced data gathering into their own ecommerce platform to power up forecasting across the entire stack. A true customer intent platform such as Lily AI's pushes accurate and actionable data into every system of record used by modern retailers. 

Lack of demand forecasting leaves money on the table. Worse, it can reach a critical point where cracks appear throughout the whole supply chain, threatening a store’s survival overall. Contact us to find out how to lock in better forecasting as a strength in your store with the industry’s most powerful customer intent platform.


FinancesOnline - 78 Black Friday Statistics You Must Read: 2020/2021 Market Share & Data Analysis

The Balance - What Is Black Friday?

McKinsey - Future of Retail Operations: Winning in a Digital Era

SaleCycle - 10 Cart Abandonment Statistics You Need to Know – 2021

McKinsey - Returning to Order: Improving Returns Management for Apparel Companies