In recent years, retail has begun an unprecedented transformation. Called everything from renaissance to Darwinian survival of the fittest, retail has changed dramatically as consumer habits shift toward digital. This has proved out in our own data. For instance, in surveying 15,800 global consumers, only 45% were interested in grocery or pharmacy delivery, with North American consumers ranking among the lowest at only 34%, preferring grocery delivery. Europe was a laggard in delivery adoption at only 36%. And, globally, anyone over the age of 55-plus was only mildly interested in the idea (27%). Fast forward through the health crisis to today, and there’s a massive shift in buying behaviors. U.S. e-commerce sales surged in several categories in April, with e-grocery sales growing 110% between March and April, and overall e-commerce sales rose 49% in the same comparison.
Many retailers had believed that their online equivalents could not replace habits like in-person grocery shopping, but this assumption was quickly flipped on its head with the health crisis, forcing retailers to respond to this unpredictable event. Online grocery sales grew, supply chains were tested and the face of grocery shopping was perhaps forever changed. Will this exceptional online grocery momentum continue? Only time will tell, however it is interesting to observe the movements of the large, dominant players in the online space. How are they evolving and adapting to this surging retail environment, and what methods are they using?
This trifecta of descriptive, prescriptive and predictive analytics can create a competitive advantage across the retail enterprise. The dominant strategy for significant e-commerce players is to double down on data using the insights derived to act swiftly and cost-effectively to create an unparalleled customer experience. Using this data intelligence as fuel, retailers can continually evolve and shapeshift to fit customer demand in a way that allows them to outperform their rivals.
The new challenge
What does this shift to a super-charged level of customer service and efficiency look like? It includes next-day home delivery and competitive fees. It demands the cultivation of extreme levels of customer loyalty. It provides “endless aisles” of product and an infinite assortment.
While this level of success may be intimidating, it’s not the work of a magic potion or secret spell. Rather, the serious and dedicated attention to data patterns comes down to noticing the unnoticeable and the slightest variational ripples. Retailers are extracting maximum intelligence from customer data emissions and using this to solve backend supply chain issues, and this paves the way for next-level predictive analytics that drive true customer satisfaction.
Cut through complexity
Consider an average medium-sized grocery chain offering a standard omnichannel experience. Imagine it has three distribution centers, 200 providers, 200 physical locations, 15-20 million orders, 10,000 SKUs and around 1 million customers a year. Let’s also imagine that 5-10% of all of the chain’s orders will be placed online, either for home delivery or curbside pickup, and there are regular daily deliveries to each store.
Even the average retailer interested in understanding demand at the consumer level would have no fewer than 2.73 trillion data intersections influencing the calculation of demand. This number doesn't even consider the never-ending increase in assortment offerings and a tireless trend for products with shorter lifecycles. This scale makes inventory management one of the biggest challenges of modern retail.
With this extreme fragmentation of demand, retailers must rethink the methods that facilitate their predictions so that they can accommodate an individualized approach beyond what is possible with an antiquated time series. They need to change up to something that allows for the continuous provision of new, personalized items with a shorter lifecycle.
Embrace the analytics trifecta
Giving the customer what they want is always a good rule of thumb, and while it is possible to offer customers a dynamic, ever-adjusting range and variety of products without an internet titan, this makes forecasting demand even more challenging. To tackle this, retailers are turning to tools like machine learning, which can now give accurate forecasts—even with a highly variable and fragmented product selection, and all while dealing with the complex supply chain that comes with it.
With increasing demand fragmentation and the exponential growth of data intersections in retail today, it’s not realistic to expect actionable information from applying traditional methods at the item level using regular sales history. Similarly, analyzing demand at an aggregated level—such as a category or a department—while tempting, often masks more granular details about how specific items are performing within a broader item group. Thankfully, through artificial intelligence and machine learning, retailers can rely on a more dependable approach built on an array of product and consumer “attributes.”
These attributes represent a less erratic pool of data, which allows items and customers alike to be grouped in different ways based on relationships and similarities. By applying descriptive, prescriptive and predictive analytics, retailers can analyze and map these forensically determined characteristics, in turn yielding accurate predictions, and ultimately, breaking down the complexity of the omnichannel supply chain into a more stable and predictable matter.
Pivot to the customer
By placing customer preferences at the center of supply chain decision-making, this strategy lets machine learning assess and adapt to consumer data on an ongoing basis. Using fourth dimension analytics, intelligent systems can burrow into intersections of products and consumer data. From there, these analytics deliver prescriptive insights that produce a next-generation understanding of the consumer, transforming the old adage of “right product at the right place at the right time” into a new question of “what is right for my customer?”
To make this question actionable, a genuinely modern supply chain must be able to predict not only when an item could be sold, but also when a consumer is more likely to make a given purchase. Retail-specific systems leveraging machine learning allow retailers to understand attributes for both items and consumers, enabling the most effective utilization of inventory and the highest level of consumer satisfaction, both of which offer retailers of all sizes a competitive advantage.
In an environment where innovation is both highly necessary and challenging, having this kind of agility allows retailers to deal with rising consumer expectations, retail volatility and demand fluctuation. In the past this kind of responsiveness has been the sole domain of digital superpowers, but it is no longer out of reach for the retail world at large. Now more than ever, retailers need the trifecta of analytics to sustain and prepare for supply chain shocks. With more disruption on the horizon ahead, now is the time to put your data to work and plot the smoothest possible course forward.
