Is our outlet store in San Francisco hurting foot traffic and sales at our full-price store two miles away? Or is it doing the opposite—attracting new customers and making them more likely to visit both stores? How are our five Manhattan stores affecting our e-commerce revenue? Are they making consumers more likely to shop on our website or to search for our products on Amazon? If we open a new mall store in the Dallas metro area, what impact will it have on sales at our existing stores, at our department-store partners, and online?
The answers to these kinds of questions are increasingly crucial to a retailer’s success, as more and more consumers become omnichannel shoppers. Guessing wrong can lead to lost sales and expensive real-estate-investment mistakes. Yet most retailers don’t give adequate thought to the cross-channel impact of their stores. They rely on gut feel or on high-level analysis of aggregated sales data to gauge how their offline and online channels interact with each other, and they assume that cross-channel dynamics are the same in every market—when, in fact, every single customer touchpoint affects the rest of the retail network in its own unique way, depending on a vast range of factors.
The good news is, there’s a way for retailers (and other omnichannel businesses) to quantify cross-channel effects, thus taking the guesswork out of network optimization. Through advanced geospatial analytics and machine learning, a retailer can now generate a detailed quantitative picture of how each of its customer touchpoints—including owned stores and websites, wholesale doors, and partner e-commerce sites—affects sales at all its other touchpoints within a micromarket. In other words, using geospatial analytics allows a retailer to see its retail network as a complex system, rather than just individual locations or independent channels coexisting in a market.
This broader view helps a retailer make better decisions about precisely where and how to reshape its network to maximize value—whether it’s by opening new stores in underpenetrated markets, shifting its channel strategy in oversaturated markets, or making store-level refinements in underperforming markets. Done right, the result of data-driven network optimization can be double-digit revenue growth. Some retailers have identified opportunities to increase their sales by as much as 20 percent.
The omnichannel consumer journey
US retail sales are on an upward trajectory. In 2018, American consumers spent approximately $3.68 trillion on retail purchases, up 4.6 percent from 2017—and, despite the growth of e-commerce, the vast majority of these purchases still happened in brick-and-mortar stores. Even brands that started as pure-play online retailers—eyeglass retailer Warby Parker, mattress company Casper, and even Amazon, to name a few—have expanded or have announced plans to expand into the brick-and-mortar world. So why have US retailers closed thousands of stores in the past year, with thousands more closures to come?
Clearly, one big reason is that the consumer journey is changing and has been for some time. Consumers aren’t just transacting in different channels, shifting more of their spending from physical stores to e-commerce sites; they’re also engaging across multiple channels, often simultaneously rather than sequentially. It’s therefore critical for omnichannel retailers to have a detailed understanding of the interplay between online and offline touchpoints, and between owned and partner networks.
In our previous article, we explained how the use of geospatial analytics enables retailers to understand the sales drivers in each store and zip code in their network. But there are several other powerful applications of geospatial analytics for retailers—including, for instance, shedding light on foot-traffic patterns and consumer demographics in a retail network, or on nascent trends in cross-shopping behaviors. In this article, we focus on one of the more cutting-edge applications of geospatial analytics for an omnichannel retailer: sales attribution. In other words, geospatial analytics can help a retailer accurately quantify the effects of offline and online sales channels on each other, thereby illuminating opportunities to capture the market’s full sales potential.
Quantifying cross-channel effects
With any geospatial-analytics initiative, the starting point is data. A retailer seeking to optimize its omnichannel network must assemble data from a wide range of internal and external sources (see sidebar, “It all starts with data”). Inputs into a geospatial model would ideally include not just transaction and customer data but also store-specific details such as store size and product mix; site-specific information such as foot traffic and retail intensity; environmental data, including local-area demographics; and anonymized mobile-phone location data. Using machine-learning algorithms, a retailer can learn which factors most influence sales in every zip code, then calculate actual and potential sales for each store and each local market.
A simulation model can then quantify the sales effect of each of the retailer’s customer touchpoints on its other channels within a local market. The model must be sophisticated enough to simulate the upward or downward revenue impact of adding or removing a particular touchpoint.
Geospatial analysis reveals that the consistency and magnitude of cross-channel effects vary significantly across channel types and markets. Exhibit 1 shows that, in one market, a retailer’s full-price stores consistently boost online sales. Its wholesale channels, on the other hand, have a mixed record: some of its department-store partners have a positive impact on its online sales, but the rest are detrimental to the retailer’s e-commerce performance.
While nationwide channel trends sometimes emerge, we’ve found that sales-channel behavior is highly market specific. Retailers should therefore make market-level channel decisions rather than sweeping, networkwide mandates to arrive at their optimal footprint.
Furthermore, geospatial analysis typically reveals that two stores, even if they’re located near each other, can have very different effects on the overall network. For example, Exhibit 2 shows two of a retailer’s stores in the same town: a full-price store that contributes more than just the in-store revenue it generates, and an outlet store that cannibalizes other stores and online sales, thus reducing its net value to the network.
Three types of market opportunities
Acting on insights derived from geospatial analytics, retailers have been able to optimize their networks and unlock growth in three ways: by expanding in underpenetrated markets, by rebalancing the network in oversaturated markets, and by fine-tuning customer touchpoints in markets performing below their potential (Exhibit 3).
We strive to provide individuals with disabilities equal access to our website. If you would like information about this content we will be happy to work with you. Please email us at: McKinsey_Website_Accessibility@mckinsey.com
In ‘white space’ markets, open new stores
Some markets are underpenetrated—that is, the retailer’s sales in the area are much lower than potential sales, the retailer isn’t fully capturing cross-channel halo effects, and there’s enough latent demand to support more retail doors. These markets represent attractive expansion opportunities.
Case example: Through geospatial analytics, a global specialty retailer identified a number of markets in which there was a large gap between actual and potential sales, and in which the company had a wholesale footprint and strong online sales but no owned stores. In each of these markets, the retailer opened one or more full-price stores and subsequently saw sales increases of 4 to 10 percent.
In oversaturated markets, rebalance the network
Other markets might be oversaturated, with less-profitable channels or retail doors cannibalizing the more profitable ones. In these markets, retailers can rebalance their network by converting one or more stores to a different format (for instance, turning a full-price store into an outlet or vice versa).
Case example: A specialty retailer found that in one US market, its outlet stores cannibalized not only each other but also its full-price stores and its website. Furthermore, it discovered that Amazon was essentially functioning as an outlet store in that market: local customers who purchased the retailer’s products on Amazon opted for heavily discounted items, largely ignoring the brand’s full-price offerings.
Further analysis showed that there was enough demand in that market to sustain more full-price stores. The retailer decided to convert two outlets into full-price locations and a third outlet into a digital showroom, with limited on-site inventory, more space dedicated to product displays, and interactive screens for customers to browse the website and place online orders. In addition, the retailer developed new strategies to win on Amazon, such as focusing on “power SKUs” (high-volume items with limited style variance that could be profitable for both Amazon and the retailer), developing “only on Amazon” items to discourage price comparison, and marketing its products more aggressively on Amazon via sponsored listings and keyword buys.
In underexploited markets, fine-tune each touchpoint
A retailer may find that it already has the optimal density and variety of sales channels in a market. But with some fine-tuning of specific touchpoints—for example, moving a store to a higher-traffic location within a mall or partnering more closely with wholesalers to better tailor the assortment to the local market—the retailer could maximize sales and take full advantage of cross-channel halo effects.
Case example: For one specialty retailer, a valuable insight was the high tourist traffic in one of its markets. The retailer found that shoppers who live outside the metropolitan area accounted for 90 percent of sales. The retailer was able to increase sales in that market by 1 to 2 percent by making its stores even more tourist friendly—for instance, offering free shipping from the stores to shoppers’ homes and investing more in localized marketing and signage to direct tourists to the stores.
No analytical model can predict the future. That said, the power of geospatial analytics in retail-network optimization is undeniable. With advanced capabilities in geospatial analytics, retailers can now view their network through an omnichannel lens and clearly see (and, to an extent, foresee) channel interactions that previously were practically invisible. Armed with these insights, retailers become better equipped to make bold decisions about their sales channels—decisions that translate directly into significant top-line growth.