Online Growth Demands Real World Change
Even with sales shifting online due to COVID-19, demand-based decisions about the physical store retain significant financial impact, and this new demand can be used as a tool to reduce the need for markdowns.
Introduction
Retailers have been impressively nimble at developing new operational processes to support a dramatic increase in online shopping, but they must now demonstrate equal agility in optimizing the profitability of these new processes. Winners will do so by deploying new technology to address fundamental changes in how they need to operate their stores.
Increases of 200–400% in online purchases, whether directly shipped to customers or picked up in store, have significant implications for the role of the physical store. While processes to service this surge in online demand have impressively been put in place often in a matter of weeks (rather than the typical months or years of piloting), most retailers have not yet optimized the profitability of operating in this new reality.
In particular, many are still making critical decisions about staffing levels and order fulfillment using legacy systems and frameworks, leaving millions of dollars on the table. And in an environment where sales is shifting from 90% physical / 10% online to more like 70% physical / 30% online, these areas will only become more leveraged.
Relearning In-store Labor
For store labor, setting a profit-optimized staffing level requires two things:
- Understanding the new optimal trade-off for labor levels for a given level of customer demand. How exactly has the revenue-labor relationship changed? What is the right level of staffing needed to support online-sourced activity?
- Predicting likely customer demand far enough in advance to set schedules, and knowing not just the most likely level, but also the range of potential likely outcomes.
For fulfillment decisions, 70/30 retailers can make significant profit improvements by considering demand, not just shipping cost.
The key question when a customer order is received and can be shipped from multiple locations is:
‘What is the likelihood that this item will sell or be marked down in its current store?’
And further,
‘How does this probability change as a selling season progresses?’
This knowledge must then be incorporated into real-time fulfillment decisions.
The margin differences alone from a right or wrong decision can be multiples of the difference in shipping costs. Aggregated over the thousands of individual decisions per day, this can represent millions of dollars of profit opportunity for a large retailer.
Capitalizing on Online Demand to Reduce In-store Markdowns
Leading companies are discovering they can make better decisions about staffing levels and order fulfillment by utilizing some of the very same AI techniques originally developed by the largest and most analytically sophisticated 100%-online retailers. More specifically, the same prediction and machine learning technologies developed to optimize their profitability can be repurposed and redeployed for the particular challenges of retailers with a physical presence. Incorporating new types of external data, and making better use of existing internal data, means that forecast accuracy can be improved, and economic trade-offs more rigorously calculated in ways that directly increase profits.
An additional challenge presented by sudden shifts in customer demand, such as the one we are now experiencing, is that past history becomes less informative. However, this is also an area where cutting-edge techniques can help. In particular, modern AI tools allow predictions for a particular store or SKU to ‘learn sideways’ from subtle trends also evident in similar stores or products across the rest of the network. This dramatically accelerates the speed of learning and adapting to the new patterns of demand and can be a source of advantage for physical retailers.
Maximizing Profits from the Online Boom
We have seen retail executives willing to embrace these new technologies demonstrate that material improvement to staffing and fulfillment decisions is possible quickly, and that simply integrating a smarter ‘brain’ into existing operational systems and processes provides a fast and low-disruption deployment path. At Predion we are pleased to be leading the development of software addressing these urgent needs and are happy to engage in dialogue about how this, our organization more broadly, and the potential for collaboration.
About Predion.ai
Predion.ai is a cloud-based software platform that combines a state-of-the-art forecasting engine with a sophisticated economic analysis engine to generate profit-optimized decisions for labor, inventory, fulfillment, pricing and production.
