Reviewing your demand forecasting? Here are the three questions you need to ask


It’s been a tumultuous eighteen months and we all need to take a breath. It’s the perfect moment to look back and consider whether it’s time to overhaul any of our processes or modes of operation. What has worked and what hasn’t? Are there any capabilities that we wish we’d had in place before they became business-critical?

Most retailers know only too well that accurate demand forecasting is essential if they are to maintain a competitive edge, stay profitable and grow sales. They also know that achieving these goals requires a full understanding of the complexity and reach of the modern retail supply chain. Given the changes that have taken place in retail since the onset of the pandemic, it’s no surprise that many organisations are beginning to review their demand forecasting processes.

If you have found yourself wondering whether your demand forecasting processes are working – and fearing that it isn’t – here are three questions you should be asking. The answers will tell you whether you have the right systems in place, help you plan how to get back on track and be ready for the future demands of your customers.


1.    Can you connect the dots between categories across your stores?

Shopping behaviour is fluid. Unless we are on a grocery run to grab a forgotten or last-minute item, most of us shop across categories. Our basket makeup will vary, but we are likely to buy some fresh produce or prepared foods while also making selections from shelf-stable, frozen and non-food categories. It’s important, therefore, for retailers to have a holistic understanding of shopping behaviour, taking into account demand signals from both the perimeter and the centre of the store. A fluctuation of demand in one area will have a ripple effect into others.

Yet cross-category demand planning is complicated by the fact that only 36% of supply chain professionals say they operate on a single supply chain platform. Fresh item management has its own nuances and complexities, so many retailers manage fresh-specific demand forecasts in one system while forecasting for the rest of the store in a disparate view. However, this siloed understanding of demand won’t deliver on the big-picture goals retailers have for sales and revenue. Fresh produce demand does not exist in its own bubble.

Fresh and perimeter items can be fickle to forecast, as they are more sensitive to weather or other external events. In the past year, however, we’ve also witnessed how demand for everyday items can throw in-store demand into disarray. Products including toilet paper, hand soap and flour were highly sought after with pandemic-induced panic buying. How did that reality impact shopper’s purchases in other areas of the store? If dry, packaged pasta was missing from the shelf, for example, did a shopper opt for fresh pasta from the deli, or perhaps a frozen option? Answering questions like these is critically important to understanding all influences on demand.


2.    Are you forecasting demand holistically across all channels? 

In late 2019, prior to any notion that a public health crisis would shake things up, we asked retailers about their supply chain priorities. Seventy-five per cent said that cross-channel fulfilment would drive them to rethink their supply chains in the next five years. This shows an appreciation for the importance of a holistic view of demand across channels, even before COVID-19 drove consumers to drastically change their shopping behaviours.

Just as a demand forecast has to account for factors across categories, it must also incorporate demand data from all retail channels. This is especially important when it comes to order fulfilment as we saw its amplification during the pandemic with the explosion of online grocery shopping and click-and-collect. To be effective across all channels, retailers’ demand forecasts must be fully integrated into inventory allocation and replenishment both online and in-store, for all forms of fulfilment.

This means that retailers need be able to map demand forecasts to each channel in order to effectively prioritise and execute orders, determining the best source from which to pull inventory. That might be a brick-and-mortar store, a dark store or a warehouse, but confidently determining the best source for inventory will ensure that a retailer fulfils the customers’ needs at the lowest execution cost. A synchronised inventory view will also provide accurate pick times and locations for store associates and third-party fulfilment partners.

If you are struggling to understand the complete picture of demand from each fulfilment channel, that’s not just a problem for order picking, on-shelf availability and out of stocks, but also for the long-term effects on shopper loyalty. Consistently missing the mark in fulfilling customer needs will be costly not only because of the logistical headaches, but because your customers may be driven directly to your competitors.


3.    Can you react fast enough to meet changing shopper needs?

We’ve established that having clarity around cross-category and cross-channel demand is critical. However, that knowledge alone won’t translate to business success if a retailer can’t take action on those insights immediately, or if the information is irrelevant and out of date by the time that a retailer decides how to execute.

As a retailer, it’s important that you have the ability to understand shifting dynamics as they happen so that you’re not operating from an unprepared and reactive stance when disruption occurs. With the right technology, you can immediately identify an event as an anomaly, incorporating historical and real-time data to make daily or even intra-daily inventory decisions. Analytics engines can even make recommendations on the next best steps to optimise inventory in response to demand changes.

The ability to react quickly also depends on alignment with other players. Retailers and their suppliers must be more aligned than ever to ensure that any supply-chain disruptions don’t result in forecasts for promotions that cannot be accurately executed. When a demand forecast doesn’t recognise a disruptive event for the unique changes it causes in demand, it has a cascading effect throughout the retail organisation and its partners. This can result in inaccurate plans, poorly executed promotions and strained relationships. To increase agility and improve inventory availability, retailers must share forecast information with suppliers. This collaboration will lead to faster reaction times, better executed promotions, and stronger relationships across the supply chain.


Demand forecasts work best when fueled by AI

Demand forecasting works well when there is a perfect balance of people, processes and technology. This is why a unified AI-based forecasting system with machine learning has become the new standard. AI recognises demand patterns that may go unnoticed by the human eye or by siloed forecasting systems, freeing up supply chain managers to execute more strategic work. Using AI, retailers can have confidence in their forecasts, which allows them to reduce inconsistent inventory buys, overstocks and their resulting markdowns, out-of-stocks, and ultimately, margin erosion.

Ensuring that demand forecasting ‘works’ is incumbent on a complete view of demand. AI allows your technology to act as an automated data scientist, ingesting and analysing the massive data sets from all categories and all channels to inform the forecast. Retailers today must have a holistic and interconnected view of demand in order to fully understand what customers want and when. AI-enabled analytics across the supply chain make that possible.