Automation opportunities in North American grocery

Recent trends and technological advancements have made automation viable across the value chain.

Cost pressure in grocery is an evergreen topic. Despite unprecedented sales growth caused by the pandemic, grocers can still pursue many opportunities to become more profitable. As sales levels stabilize, both public and private companies will again face cost pressures. Using excess cash accumulated during the pandemic to invest today can help to manage cost pressures tomorrow.

Automation is an often underexplored cost lever in grocery, mainly because it historically required high capital investments and yielded lower or unpredictable ROIs. As other industries have shown, automation can help streamline labor-intensive practices, reduce rework, and create opportunities to foster more creativity within a company. We’ll explore automation across different parts of the grocery value chain: stores, supply chains, last mile, and corporate centers.

Stores

For traditional formats, labor in grocery stores has long been the second highest profit-and-loss item (after the cost of goods sold). More grocers are now beginning to explore how automation can enable the reallocation of their workforce to areas that can be a differentiator in the marketplace.

Theft-detection software

Until recently, camera technology wasn’t sharp enough to catch incorrect scans at self-checkout lines. By one grocer’s estimate, at least $30 per self-checkout machine is scanned incorrectly or deliberately stolen each day. Grocers are beginning to explore solutions that use camera technology and AI to pause and flag questionable transactions. Theft detection will be critical as grocers migrate to more economical checkout options.

Smart service counters

A common pain point among most North American grocers involves labor in deli departments. Staffing the service counter is unpredictable and frustrating for both customers (due to wait times) and employees (due to constantly changing demand patterns). To remedy this problem, grocers are using order-ahead technology through an app so that employees know what to produce. Before a customer enters the store, the deli counter can create the order during lower-traffic periods, cutting customer shopping time.

Electronic shelf labels

The ROI for electronic shelf labels historically hasn’t been attractive for most grocers. However, one recent development is changing this equation: the more agile deployment of dynamic-pricing software, which allows grocers to reset prices (normally once the store has closed) according to local demand patterns. Dynamic pricing instantly alleviates the lag time in implementation with once-weekly price changes. To experiment with this technology, some retailers are placing the tags only in select areas (for example, planograms with high promotional density) or categories that take longer to change (such as in frozen sections). The combination of rising labor costs and dynamic pricing has cut the previous investment payback of four to five years to less than two years.

Cameras and robotics

In-store robotic equipment isn’t new to the grocery format, which has used floor-cleaning robots for several years. However, the advent of surveillance cameras in stores has allowed for better tracking of inventory, planogram compliance, and in-stock measurement. These alerts can be pushed to store associates as tasks to fulfill in a prioritized fashion. New robotics can assist with replenishment while also doing more traditional tasks (such as cleaning) more effectively.

The best way to assess the business value of these solutions (and other labor-saving initiatives) is to build smaller-scale tests—for example, across a subset of stores or a banner. Historically, such tests have proven difficult, even for the most agile of retailers. The rapid response of grocers to the COVID-19 pandemic, however, has demonstrated that they are able to move at a remarkably fast pace and that customers will share feedback on what works and what doesn’t. In addition, creating measurement plans is critical to support investment discussions once tests are completed.

Supply chains

Within supply chains, significant automation is already underway. Microfulfillment-center (MFC) providers such as Dematic, Fabric, and Takeoff are partnering with grocers of all sizes and operating models—from physical retailers (such as Walmart and Albertsons) to pure-play online retailers (for example, FreshDirect). At sufficient densities, MFCs can handle 4,000 to 5,000 orders each week, with 100 to 150 units picked per labor hour. These fully automated platforms also buoy the fulfillment economics, bringing home-delivery basket margin to five times the level of centralized fulfillment centers such as Ocado. The average cost per location of an MFC is now approximately $3 million and takes anywhere from four to 12 months to launch, providing a 70 to 80 percent internal rate of return when scaled appropriately.

Grocers are making additional investments in supply-chain automation, mainly in retrofitting existing distribution centers with picking tools that reduce cost to fill. According to McKinsey estimates, these supply-chain tools will generate more than $20 billion in value over the next three to five years. 1 Last, while more closely tied to AI than automation, improvements in inventory algorithms are helping offset rising capital costs by more nimbly predicting where to deposit goods in the value chain, given the more decentralized fulfillment ecosystem needed to support omnichannel.

Last mile

During the pandemic, e-commerce penetration in North American grocery tripled to 9 to 12 percent. This shift has suddenly made the last mile a busy and potentially profitable part of the value chain. Against this backdrop, several types of automation have become more common in grocery recently, and these developments have garnered attention.

Outsourced delivery labor

Traditional delivery options include company-owned assets, contracted delivery services, and parcel companies. Automation is playing an increasingly central role in the ability of outsourced labor models to execute effectively, giving rise to two new formats: in-house and outsourced gig networks (Exhibit 1). In-house networks give grocers much more control over talent pools and resource allocation. Larger retailers (such as Amazon and Walmart) are using their own labor and gig networks as a way of flexibly distributing resources when needed. The outsourced gig providers—such as DoorDash and Uber Eats—will often contract with retailers as needed, and the gig model allows independent contractors to choose their own hours, providing added flexibility.

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

Autonomous vehicles

In the coming years, grocery retailers will shift from regular delivery to assisted autonomous delivery before moving into droids and fully autonomous delivery (Exhibit 2). Major retailers have already begun experimenting with these methods. The implications of this transition will be massive—both for technology companies that are currently developing solutions and for the grocery industry. One estimate suggests that the advent of fully autonomous vehicles could reduce the cost of delivery orders by an average of 40 percent. 2

The last mile has long been capital intensive. However, the evolution of interpretations of labor and public sentiment (for example, California’s Proposition 22 on app-based transportation and delivery drivers, which was approved by voters in November 2020) has given retailers more flexibility. In high-density environments, company-owned models typically yield better economics than gig networks, mainly due to network effects of batching and routed delivery models. However, in subscale locales, gig economics present more favorable unit costs due to the removal of fixed labor. The embrace of gig economies and start-ups represents an early step on the road to more automation.

Corporate centers

While automation is newer to field functions (for example, stores and distribution centers), it has long been used at corporate and store support centers as a way of improving efficiency and reducing errors. In fact, automation will affect nearly every part of a retailer’s operations (Exhibit 3). The following two functions reinforce the opportunities that automation can unlock:

Automation will affect most elements of the merchant role.
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

Merchandising and buying

In merchandise planning, automated predictive analytics are replacing time-consuming reports. Automated-replenishment models also support the movement of inventory while maximizing margins, saving the time previously spent on scenario planning. Even the more “artistic” elements of a merchant’s role, such as program negotiations and supplier management, are being streamlined through automated playbooks and synchronous data sourcing.

Depending on the organization, pricing sits either within its own center of excellence or a hybrid model. Web scraping (executed even at the zip-code level) helps pricing teams assess gaps in known value items compared with the competition and can dynamically adjust prices to remain competitive. This capability, coupled with electronic price tags, turns a price-change process that previously took four to eight weeks (based on aggregate point-of-sale data comparisons from vendors such as Nielsen or IRI) into a one- or two-day exercise.

Corporate functions

Automation has significant potential within several back-office support functions. In finance, for example, software can streamline accounts payable, accounts receivable, and financial planning and analysis. In HR, automated criteria screening and aptitude testing can ensure that retailers place candidates in the right roles.

Such innovations can take many forms, depending on the end user. For example, HR team members can use expected attrition estimates and projected demand to forecast the number of new hires and place advertisements in external talent apps. Candidates submit resumes through the digital portal, which automatically scans for relevant criteria; an HR colleague then reviews the screened sample. This process can cut hiring times by more than 50 percent. Although integrating automation into these roles tends to take longer, it comes with an impressive ROI as a result of higher margins and improved accuracy.

One challenge is that conducting pilot projects is often more difficult due to the extensive reach of these systems. As a result, companies need to think carefully about which pain points to automate before making investment decisions.


Vaccines will bring back more in-person dining, and grocery spending is expected to return to prepandemic levels. Grocers will need to reexamine their automation ambitions as a way of addressing structural costs and improving productivity across the value chain. Doing so can increase profitability levels and reinvestment potential not typically seen in such a traditional and competitive industry.

Explore a career with us

Related Articles