Even before the world first heard of COVID-19, retailers had been rightly questioning if their traditional planning process was broken. Consumers’ behavior has become much more fickle, while their expectations from retailers have shot up over the last few years. Simply plugging in the prior year’s numbers, then, was already beginning to feel unreliable and incomplete. But the onset of the pandemic has completely upended planning.
Considering how much the world and consumer behavior has changed over the last 18 months, trends and numbers from 2019 are almost certainly dated and stale, while 2020 was just a fundamentally unusual year. And although 2021 has offered some indications of what our post-pandemic future holds, there are still too many unknowns. In addition, data from this year will be skewed by macro factors such as pent-up consumer demand, the federal stimulus feeding a shopping frenzy, supply chain stresses, among others.
All this means that many retailers are scratching their heads about how to plan for 2022. So, what is the answer? For years now, retailers have been talking about balancing the art and science of retail, but now is the time to really embrace the science part of it. Retailers must take a mathematical approach to planning and build a model that incorporates multiple drivers and variables, is updated much more frequently, and gets smarter over time by constantly learning and applying insights to future analyses (Figure 1).
While there is no recommended single model, predictive analytics can help clearly identify drivers and adjust how each factor is weighted to make an informed and defensible forecast. The main goal of the model is to understand how each variable impacts the output, which can help you stay nimble and adapt quickly. There are three main steps to building a model:
Determine your baseline: This can be achieved by identifying key demand-supply inputs during other instances where stores were closed, there were severe supply dislocations, or situations where business had to be restarted. Examples of such instances could be during prior extreme weather events such as hurricanes or trade or port strikes.
Identify data sources that will predict outcomes: This would not only include your internal numbers, but also factors such as government data to model where new COVID-19 outbreaks may occur and external foot traffic data to estimate the probability of store visits.
Iterate and learn continuously: Be prepared to continuously forecast. As new data becomes available and inputs change, predictive models will update, and a strategic response will be required.
There are both demand and supply considerations when building a model, and each input demands thoughtfully answering several questions around it. Demand considerations include channel shifts, category shifts, traffic fluctuations, and pricing and promotions strategy, while supply considerations include labor, materials, and transportation.
Channel shifts: Determine how your channel mix has shifted and if consumer behavior by channel has changed. Are customers who made the move over likely to stick to digital channels only going forward? Honestly assess if your inventory is truly flexible in an omnichannel world. How are your main competitors investing in channel capabilities and which way is the broader market trending?
Category spending shifts: What categories gained during the pandemic, and are these trends expected to hold or reverse? Assess if customer priority changes impacted demand in certain categories and how any upcoming macro trends, such as back in the office or back to school, may affect sales. What category bets are your competitors making, and has the category competitive landscape changed due to the recent shifts?
Traffic fluctuations: How was your traffic impacted by COVID-19 and are these changes likely to be permanent? This may vary by location and geography and by overall population fluctuations in certain areas. Ask how your customer loyalty mix and share of wallet may have changed. Then determine if your competitors are seeing similar shifts.
Pricing and promotions: Determine if your customers became less or more value-oriented during the pandemic. What impact could changes in discretional spending have on your pricing decisions, and what will the future of stimulus payments -- or lack thereof -- have on purchasing power? Relatedly, are you planning to be more or less promotional than the year before?
Labor supply: Determine how labor availability has been affected for manufacturing units and factories, as well as for various store locations and distribution and fulfilment centers. How are prevailing economic and social conditions affecting labor cost and supply?
Availability of materials: Assess if and how labor constraints may, in turn, affect the supply of raw materials. In addition, account for any shortages in availability of materials, such as chips, that are predicted to last in the short and medium term. Account for suppliers that have either faced or caused issues during the lockdown and beyond, including making down payment demands, showing unwillingness to store goods, and not receiving credit.
Transportation issues: The current customs and shipping backlogs must be accounted for in any planning exercise. In addition, the cost of both shipping and trucking has gone up significantly, which must be factored in.
Inflation and cost increases: Identify changes to raw material cost and any impact of inflation on supply. Are necessary materials too expensive to meet margin targets? Monitor costs to maintain visibility to potential shifts in cost of goods.
The number and variability of these factors suggests that meaningful and accurate planning for the coming year will be among the harder challenges a retailer would have faced. However, this is an opportunity to swiftly make lasting improvements that were making their way to the industry at a much slower pace. The pandemic and its fallout have prompted businesses to take a closer, more critical look at the way things were always done and, in many cases, the resulting change has led to better workflows and increased efficiency. For retail, this forcing mechanism holds the promise of bringing in a much-needed sea change in accuracy and speed of decision-making.