The rise of omnichannel during the pandemic—fueled, in large part, by the meteoric rise of e-commerce—is no passing fad. In the United States, for example, online-grocery sales penetration as of June 2021 was three times above prepandemic levels, rising from low or middle single digits to low double digits.1 In addition, consumers have elevated expectations: convenience and value were often cited as top reasons for shopping at a new retailer throughout the COVID-19 pandemic.2
While omnichannel continues to offer added convenience for consumers, profitability remains a challenge for grocery retailers. Sales, particularly during the lockdowns of the pandemic, went through the roof but were accompanied by spiraling costs. In particular, delivery costs and costs for store picking and other operations can range from 10 to 13 percent and 11 to 14 percent of sales, respectively.3 With omnichannel here to stay, retailers have no choice but to find a path to profitability across all channels.
Operations hold the key. Focusing on three areas—network, fulfillment, and last-mile transport—offers a range of opportunities to cut costs. In this article, we explore each in detail, as well as the factors grocery executives must consider to develop more cost-effective strategies.
Balancing a differentiated offer with costs
Across the food and grocery landscape, retailers are achieving differentiation with their commercial offering and delivery capabilities. Many are simultaneously emphasizing speed while navigating complexity across expanded assortment and convenience (such as substitutions and scheduled delivery windows). They are also gauging the impact on profitability of density of the consumer base, delivery-service radius, use of third-party partners, and consumer incentives (for example, free-shipping-order minimums).
How should grocers balance these trade-offs? We see several decisions that will lead them to the right answer.
Delivery offering, precision, and speed
Grocers have many choices for consumer delivery, including pickup, curbside, locker, scheduled, and same-day and instant delivery. While these are convenient for consumers, retailers must consider the implications of each on operations. For example, grocers must determine the precision and duration of the delivery window. A narrow window creates little room for error or to flex shipments for the most efficient routes. However, too broad of a window could lead consumers to choose another retailer. Indeed, research on US consumer willingness to pay indicates stronger preferences toward precise windows versus faster delivery.4
In addition to precision, many players now offer same-day service for large orders. For example, Asda promises delivery within an hour for a basket of up to 70 items from full assortment.5 As speed increases, the density of drops declines, typically creating higher delivery costs. The long-term viability of full assortment and instant delivery for all consumers can be challenging. A segmented approach to service offerings and consumers is critical to support a full-service model. While consumers residing outside service boundaries may miss out on instant options (though same-day delivery could still be offered), the benefits of providing more consistent and reliable service to core consumers can outweigh those risks.
Fee structure is a close neighbor to delivery offering, speed, and precision—and it can play a critical role in subsidizing the additional operations required to fulfill an e-commerce order. Delivery fees and associated service levels require trade-offs relative to local competitors and localized consumer preferences. In Europe, for example, “no frills” scheduled-delivery models (such as Picnic) do not include any delivery fees. By contrast, Amazon instituted a standard fee of $9.95 for all Whole Foods grocery deliveries through Amazon Prime in the United States, a bet that consumers would be willing to pay for convenience.6
Range of SKUs offered and alignment to store assortment
In fulfillment operations, the number of SKUs and categories creates additional complexity. The assortment size and special requirements (for example, cooling, packaging, and weighting) have process and inventory implications that retailers must balance against consumer needs and expectations. A broader assortment includes trade-offs on available space to store a longer tail, the accuracy of inventory (and subsequent impact on picking efficiency), and the productivity of operations when fulfilling large, complex orders.
Retailers with physical locations also need to determine the degree to which their offline assortment will mirror their online offering. The more these diverge, the greater the fulfillment complexity. In addition, decoupling online and offline promotions is a key lever to profitability, so companies must be prepared to invest the necessary resources to handle these channels separately.
According to IRI’s latest CPG Supply Index, 11 percent of edible packaged consumer products will be out of stock in store.7 In a pick-from-store model, this means 10 to 15 percent of orders will experience stockouts and potentially require substitution during fulfillment. A variety of factors can drive stockouts, such as inventory accuracy and changes in availability between order placement and fulfillment. For these situations, retailers need to decide how they want to interact with consumers regarding substitutions. One-to-one engagement significantly reduces order-picking productivity but provides a premium service. Alternative strategies include prohibiting substitutions, using advanced analytics to determine real-time substitutions during picking, and providing options during the prepurchase shopping journey to preselect choices in the event exceptions occur.
Operations are critical for profitability, but the range of optimization and efficiency gains for online operations requires clear trade-offs with decisions related to consumer value propositions. Once grocers determine the strategy to meet consumer expectations, the focus then turns to execution and operations.
Imperatives for omnichannel grocery operations
We believe grocers need to focus their efforts on operations to increase the efficiency of their omnichannel operations and reduce costs. We see three critical areas.
1. Omnichannel distribution network: A diverse set of operational models based on different constraints across the country
What it is: As grocers rapidly look to capture new markets with different attributes (for example, order density, assortment size, and delivery speed and precision), they will need an ecosystem of order preparation and last-mile operations that takes into account costs, service levels, and product quality.
The optimized ecosystem should be determined based on different options for order preparation, including traditional stores, dark stores, microfulfillment centers (MFCs), and highly automated central fulfillment centers (CFCs). These options need to be evaluated in combination with different last-mile models (for example, in-house fleet, outsourced fleet, and crowdsourced services). In addition, the online-grocery network isn’t static; it needs to be reevaluated over time as value propositions and demand evolve. Retailers should also devise a robust plan for how to test, learn, pilot, and adapt as new operational models and automation are deployed. Grocers will also need to look beyond a specific end-state model, as a mix of assets and flexibility will continue to be as important as ever.
Examples in practice: A network of dedicated order-preparation assets has been a part of the portfolio of many retailers for some time. Tesco, for example, has used a mix of traditional and dark stores with different automation levels to fulfill grocery orders since opening its original dark store in 2006.8 However, technological advancements present new options, such as automated MFCs. Several retailers are adapting their models with the launch of automation solutions: Tesco recently announced the launch of ten MFCs per year for the next several years,9 Ocado has committed to an additional 56 CFCs,10 and Walmart announced plans to scale MFCs across three technology providers.11 In many cases, these MFC models play a specific role in high-density areas to augment pick-from-store volume.
What grocers should consider going forward: Grocers will need to assess the full set of order-preparation options (Exhibit 1) together with last-mile options by market. These decisions should be considered against specific market scenarios (for example, the share of orders with same-day delivery versus the share of orders picked up) to guide the range of long-term network options.
2. Omnichannel fulfillment—node operations: Best-in-class order-preparation execution
What it is: As grocers look to adapt their order-preparation network, they have alternatives to increase the efficiency of their in-store and warehouse picking operations. These opportunities span changes to the operating system, such as wearable scanners allowing both hands to be free for picking; the management system—for example, labor scheduling to match picking capacity to demand; and people systems, such as performance structures to encourage sustained employee unit-per-hour performance (Exhibit 2).
Examples in practice: During the pandemic, retailers rapidly added capacity by using their stores to achieve incremental efficiency gains and support additional e-commerce orders. Many are expanding their fulfillment capabilities. For example, Kroger plans to implement a picking and packing software in partnership with Ocado to improve efficiencies within stores.12 Similarly, Morrisons introduced capped shelving to improve product availability and replenishment times, adapted back-of-house space for online operations, and tested digital shelf labels to improve its online-picking process.13 Tesco adapted its operating model to increase the flow of orders, including starting picking earlier, extending click-and-collect hours, and adapting picking processes.14
What grocers should consider going forward: To establish best-in-class operations, grocers need to look at all aspects of the operating model. Some retailers are already taking the following actions.
- Operating system:
- Multiorder carts. Picking carts can optimize both the number of orders and containerization for each order. Pickers can focus on discrete orders or zones of orders (such as produce) to batch their picking efforts and substantially decrease the average travel distance for each order. This type of multiorder picking does require a certain level of order density and enough lead time to ensure orders can be combined.
- Pick to light with electronic shelf tags. In an in-store picking environment, the pick path and search efficiency are often only as good as a picker’s experience and the clarity of images provided on their handhelds. With the introduction of electronic shelf tags, their illumination can improve search and picking efficiency, much like the pick to light used in warehouse settings. StrongPoint has recently implemented this technology in ICA stores in Sweden, a move that has reduced both picking time by six to nine seconds per pick and errors by 30 percent.
- Advanced analytics for substitution logic. For in-store picking processes, differences in available inventory and consumer-requested inventory can create significant process inefficiencies since pickers either have to make direct one-to-one contact with the consumer to select a substitute or pick a substitute on an additional picking run. Accurate inventory can minimize those issues, but most retailers will struggle to have better than 95 percent unit availability with online orders during store picking. Grocers can use new advanced-analytics tools to provide consumers with up-front, like-for-like substitutions while ordering, eliminating the need for pickers to interact with consumers during runs while maintaining satisfaction levels.
- Management system:
- Dynamic labor scheduling. Enabled by adaptable work systems and scheduling, dynamic labor scheduling ensures consistent coverage and creates flexibility for store associates to organically accommodate unmet consumer needs and order windows. In many cases, retailers will cap orders for certain delivery windows to distribute volume across the day or week to achieve greater utilization of both labor and any automation asset involved in the picking.
- Performance-management boards. Once labor is effectively scheduled, visual indicators within picking areas and picking apps can help significantly increase productivity and performance. Physical productivity boards should be used during stand-up huddles, and in-app picking performance should deploy a simple “green, yellow, red light” logic to tell a picker if they are on pace.
- People system:
- Retailers are exploring ‘pushed,’ tailored training for pickers and warehouse employees. These tutorials are delivered directly to workers’ handheld devices based on tenure and job performance across certain areas. Instead of standard time-phased training, this approach provides customized insights to help colleagues continually become more productive. Kroger, for example, has implemented a customized training model that offers employees daily five-minute sessions to support continuous learning.15
3. Omnichannel fulfillment transportation: Last-mile transport as part of the end-to-end supply chain network optimization
What it is: While consumers’ increasing preference for convenience is fueling online growth, last-mile delivery remains one of the key obstacles for online grocery profitability. Supermarkets, e-commerce players, and online grocers are all piloting new delivery models and need to concentrate on decreasing costs while still offering convenience to consumers.
First, as companies decide how delivery speed aligns with their consumer value proposition, they need to understand its impact on operations. Depending on expected shopping-basket size, market density, and target delivery speed, the optimal last-mile solution can differ substantially. Less-prominent alternatives to home delivery—such as drive-throughs and collection points—should also be considered (Exhibit 3). Notably, in less-dense markets, profitable instant delivery is exceptionally difficult to achieve. Last-mile options may need to be differentiated by local market (for example, rural versus urban areas), even within the same brand.
Within the last-mile models, companies should consider a selection of levers to optimize costs. They need to address three key challenges for home delivery.
First, flexibility in delivery time windows allows for more efficient routes by combining several deliveries in a milkman run. However, fixed, narrow time slots lead to suboptimal routing and possible wait times, potentially increasing costs by more than 100 percent. Flexibility in time slots also allows retailers to balance peak demand with nonpeak hours of the day. Since consumers appreciate short windows, features such as text messages to confirm a driver’s arrival can help manage trade-offs. A benefit scheme for consumers who accept flexible delivery, for example, could make a difference.
Next, optimization of real-time routing and drop density to balance multiple constraints (for instance, delivery windows, incoming orders, or available drivers and vehicles) is critical to control last-mile costs. Strong integration with commercial incentives (for example, reduced fees for next-day delivery) can help shape consumer demand in ways that allow efficient order execution. Next-generation routing optimization considers the holistic network, including flexible start and stop locations, to avoid travel back to starting depots.
Finally, insourcing—as opposed to an outsourcing or partnership model—assesses whether to own and operate a dedicated fleet, use an outsourced gig-delivery platform, or outsource delivery (on demand or dedicated). These options need to be carefully assessed along financial, capacity, control, and capability criteria. Cooling requirements might also have an impact on this decision: using passive cooling can make it possible to ship through existing same-day delivery and courier networks. The pooling of last-mile transport across multiple stores, businesses, and sectors may be explored, as it allows retailers to increase drop density and improve the utilization of the chosen distribution model. All adjacencies and category additions aren’t considered equal; for some grocers, the complexity can outweigh delivery efficiencies (for example, freshly prepared food).
Examples in practice: The Dutch no-frills online grocer Picnic provides its consumers with a user-friendly app combined with competitive prices, a low minimum order value, and no delivery fees. The retailer increases efficiency through a targeted assortment and a milkman-run approach, where electric delivery vehicles visit a given area just once or twice a day to increase the density of last-mile deliveries. Next-day delivery is possible if the order is placed before 10 p.m.; consumers can then choose a one-hour delivery time window. For consumer convenience, the window is narrowed to 20 minutes on the morning of the delivery. This allows Picnic to optimize routing plans while limiting service restrictions. With this approach, e-grocers can also provide incentives for consumers to choose delivery windows in which other consumers nearby are already planning to receive a delivery.
What grocers should consider going forward: Depending on advancements in technology, consumer acceptance, and regulatory restrictions, several innovative solutions currently being piloted or in R&D could eventually reduce delivery costs (Exhibit 4). For example, electric or autonomous vehicles, automatic drop-off points, or droids may provide new options to serve markets in a cost-efficient manner. These innovations will help to recover last-mile costs while allowing grocers to enter previously less-attractive, low-density markets.
The path forward for omnichannel grocery retailers
Increased consumer expectations on convenience and product offerings are here to stay. As grocers think about profitable operations for online grocery with these raised stakes, they shouldn’t look for a silver bullet. Instead, an interconnected series of decisions across the value proposition and their impact on operations—including network design, order preparation, last mile, and best-in-class processes—will be crucial to bend the curve toward profitability. With many decision points across the omnichannel journey, a coordinated and consistent strategy will be critical for success.