Achieving profitable online grocery order fulfillment

The explosive growth in North American online grocery presents a vexing challenge for retailers. While the need to serve this channel is obvious, achieving profitability continues to be a challenge. In the meantime, consumers have gotten used to the convenience of delivery—the faster, the better. Grocers are under rising pressure to meet these ever-increasing expectations and mitigate online grocery’s dramatic impact on order economics.

To serve online demand without significantly cutting into margins, executives must focus their investments on the two major drivers of online fulfillment cost: handling and delivery. Grocers can manage these drivers by selecting from a range of fulfillment and delivery models. The right solution varies depending on a region’s demographics and population density and is also shaped by the target customer value proposition. Organizations that navigate these options wisely will be better positioned to grow profitably.

Online is here to stay, and consumers are more demanding than ever

Online grocery penetration has grown steadily over the past several years but was greatly accelerated by COVID-19. According to a recent McKinsey consumer survey, some trends that took hold during the pandemic will likely endure and even deepen:1

The online channel continues to show strong growth. Online and delivery orders rose by about 50 percent during the pandemic and are expected to increase two to five percentage points in 2022, depending on the delivery option.

Consumers increasingly want delivery rather than click and collect. For online grocery shoppers, click-and-collect offerings have been supplanted by home delivery. According to our survey, 63 percent of online grocery shoppers in December 2021 preferred home delivery to click and collect, up from 48 percent a year earlier. Consumers are particularly gravitating toward same-day and instant delivery. The latter increased 41 percent over the course of the pandemic and is expected to rise further in 2022, with a net two percentage points of survey respondents expecting to increase their use of instant delivery.

Personal contact continues to decrease in importance. Before the pandemic, 46 percent of consumers preferred personal contact in stores. The past two years have muted its importance: now just 31 percent value this type of engagement.

Convenience trumps all. Consumers were drawn to online shopping for its convenience and relative safety during the pandemic, and our survey suggests shoppers have embraced these benefits and aren’t interested in returning to the prepandemic normal.

As these insights reinforce, even grocers that rapidly adapted their operations early in the pandemic have no time to rest; they must continue to keep pace with evolving consumer preferences.

Online’s economics problem

For grocers looking to design an effective online offering to complement their traditional brick-and-mortar channels, the challenge boils down to unit economics.

For grocers looking to design an effective online offering to complement their traditional brick-and-mortar channels, the challenge boils down to unit economics. A typical North American grocer achieves a positive net profit-and-loss (P&L) margin of about $4 on a traditional $100 grocery basket when the customer is walking the aisles in the store (Exhibit 1). By contrast, when the grocer has to manually pick from the store and deliver to the customer, net P&L margin becomes approximately –$13 for a $100 online grocery basket order (assuming no additional fees are collected from the customer). Today, a combination of higher pricing for online orders and fees collected from customers for the convenience of online ordering and fulfillment enables positive unit economics in some cases. However, we expect to see customer willingness to pay this “e-commerce tax” decrease as online offerings come to be seen as table stakes.

1
Costs for home delivery can have a significant adverse effect on margins.

Since the ability of grocers to use fees or higher prices to drive incremental gross margin is limited, controlling fulfillment costs is paramount. The negative impact of online on unit economics is largely caused by two incremental cost drivers: picking labor and last-mile delivery costs. On a $100 online grocery basket, manual-picking labor costs and associated overheads would typically amount to about $8, with the cost of last-mile delivery to the customer amounting to another $8. Clearly, the incremental costs of fulfilling online grocery orders far outweigh the razor-thin margins on a traditional grocery basket, causing online fulfillment to fare poorly on unit economics. The economics get worse when customers require near-immediate fulfillment; grocers will struggle to optimize the use of labor and delivery fleets to meet this faster fulfillment timeline.

Achieving profitable fulfillment

Grocers have been experimenting over the years with various strategies to reduce picking and delivery costs. They must now double down on investments that will help offset the higher costs of rising online penetration and be strategic about where they make these investments across their networks. Retailers should not keep waiting for a better solution or attempt a one-size-fits-all approach. Instead, they must be prepared to adapt based on market dynamics.

Grocers that may be inclined to wait for further technology evolution to save them will be disappointed. Over the past few years, several retailers have been experimenting with different models that can meet rising demand. Many grocers have yet to fully commit to one model, instead waiting to see what others do or expecting a new technology to emerge that will reduce costs. Those options are unlikely to materialize anytime soon. Still, profitable online order fulfillment is achievable. Grocers need to start making strategic long-term investments that are aligned with their online value propositions or partner with third-party logistics providers that have made such investments.

Some retailers are already gaining a first-mover advantage. For instance, Walmart has begun to invest heavily in store microfulfillment centers (MFCs) across its network. This approach involves setting aside roughly 15,000 square feet in the back room of the store for automation. With the MFC solution, retailers can pick at more than five times the typical manual picking speed of approximately 60 to 70 units per hour (UPH) assuming no augmentation by technology.2 The MFC also halves the typical pick-and-pack cost per order relative to baseline manual picking while enabling faster picking and delivery (for example, a two-hour pickup), levels that are difficult to achieve at scale using manual picking.

Retailers can apply the following lenses to the different fulfillment options in order to optimize picking and last-mile delivery costs:

Picking costs

Picking costs typically represent the greatest aggregate cost for online fulfillment. The average grocery basket contains about 30 items. Assuming a pick speed of 60 UPH, the 30 minutes spent picking a single order translates to about $8 to $10 an hour.3 This total includes the labor costs associated with picking as well as order consolidation, staging, customer contact, and handover.

Grocers can take advantage of a range of picking options with different degrees of centralization, capacity, and automation. Grocers need to build a portfolio of fulfillment options by geography, with the mix varying by the density of current and projected demand and their fundamental online value proposition (for example, cost or speed of delivery). Even within a geography, grocers need to build the capabilities to support multispeed delivery, since customers increasingly prefer delivery (particularly instant) over click and collect but have different fulfillment speed preferences (especially if they do not pay for delivery). Each fulfillment option offers different features (Exhibit 2):

2
Features and required investment levels vary from one type of facility to the next.

In-store picking. With little investment, grocers can stand up a manual in-store picking operation. This option, which can handle 800 to 1,000 orders per week, comes with higher operating costs (due to the highly manual nature of the activity) and is best suited for low-volume areas.

Warehouse in store. Retailers can create a dedicated area within or attached to the store where pickers can aggregate popular items for delivery. This low-capacity option requires a medium level of operating costs and can handle 1,000 to 1,300 orders per week at a UPH of 80 to 100. This option works well for moderate-density suburban areas.

Robotic MFC. These MFCs require up-front investments in technology to replace manual pickers with robots, which lowers operating costs and increases capacity (up to 4,000 orders per week with a UPH of 400 to 500). Typically located close to consumers, robotic MFCs enable delivery in two hours or less. This option is best suited for high-volume, high-density urban areas.

Dark store. These stand-alone facilities offer an optimized layout to support manual picking for e-commerce orders. After an up-front investment, dark stores can handle 2,400 to 3,000 orders per week. They are ideal for moderate-volume areas.

Traditional warehouse. This larger stand-alone facility features a wider assortment of products, as well as the potential to incorporate automation to increase capacity. Traditional warehouses, which require a medium up-front investment, are often located far from the end consumer and hinder same-day delivery.

Highly automated warehouse. By deploying automation across several process steps, this option significantly increases capacity to nearly 30,000 orders per week thanks to a UPH of 600 to 800. The product assortment is more generic, making highly automated warehouses suitable to a broader customer base. The substantial up-front investment leads to lower operating costs.

To understand how these options could be applied, consider a California-based grocer with a statewide presence. This grocer might choose to invest in two MFCs each in Los Angeles and San Francisco to handle a combined 16,000 orders a week. Alternatively, it could set up one central fulfillment center (CFC) outside each of these cities to serve customers within a 100-mile radius. Fresno would not have the order density to justify an MFC, so the grocer could choose between a CFC to serve that city and the surrounding areas or a traditional grocery store with manual in-store picking combined with a dark store.

Delivery costs

As more customers shift their preference from click and collect to delivery, the last mile accounts for a rising share of costs. Sources include labor and fleet costs (for retailers that have built their own delivery capabilities) or a share of revenue paid to a third-party provider (such as Instacart) for picking and delivery. The total cost of delivery can vary significantly by region, population density, and order perishability.

A variety of models have emerged in last-mile delivery, enabled by growth in the gig economy (Exhibit 3). Grocers should assess each transport model based on its value proposition. The fundamental trade-off is between speed and cost of delivery. For example, to enable delivery in two hours or less, grocers can tap into the gig economy or use a third-party provider such as Instacart. Both of these options are typically more expensive. A less-expensive model could be to develop an owned fleet. However, the economics work only with an order density that can support high fleet utilization. In most geographies, grocers must wait until enough orders come in to limit the cost per order, which can significantly increase delivery times.

3
A variety of models have emerged in last-mile delivery, enabled by growth in the gig economy.

Grocers must have a clear idea of what their customers value. If they care about quick delivery, are they willing to pay for it? If so, grocers should use the gig economy and third-party providers. If customers care more about costs and wouldn’t switch grocers for faster delivery (for example, same day versus two hours), grocers should consider building their own fleet—but only in regions with the requisite order density to make this option cheaper than tapping into the gig economy or third-party providers.


Grocers can choose from a number of paths to achieve better online order economics across large parts of their network. The rapid evolution of available technology over the past few years has helped to clarify the best options to support picking and delivery. To make the right bets, grocers should take the following steps:

  • Evaluate their customer promise and value proposition, and consider what customers really care about.
  • Assess how well their fulfillment strategy fits with the customer value proposition.
  • Look at the big picture to consider whether they have the right technology in the right places.
  • Make the aggressive investments required or partner with established third-party logistics providers.
  • Set EBIT and margin targets and monitor these metrics to track improvement.

Grocers that design a portfolio of options tailored to each geography’s order density and customer preferences can chart their own path to online profitability.

Download the full report here.

Explore a career with us