The E-grocery Automation Outlook for 2021: Strategic Refinements and Emerging Technologies
2020 will prove to be a pivotal year in e-grocery automation. Coming into the year, e-grocery sales were showing signs of accelerated growth after several modest years, and many grocers were developing plans to automate manual fulfillment processes.
Then the pandemic hit, and e-grocery demand spiked to levels no one expected. Suddenly grocers who had a two-year horizon for automation were compressing their timelines to bring in automation as quickly as possible. At Swisslog we saw demand for e-grocery automation surge to levels we hadn’t expected to see until 2023.
In short, the industry packed several years of experience into nine months and the lessons learned during 2020 will prove invaluable as e-grocery fulfillment continues to scale and mature in 2021 and beyond.
Building the Foundation
E-grocery fulfillment has benefitted enormously from technologies originally proven and refined in other e-commerce applications. However, e-grocery is also different than other e-commerce businesses in terms of the distribution model; variety of product shapes, sizes and storage/temperature requirements; and order complexity.
Much has already been written about the impact of automation on e-grocery distribution strategies. We won’t spend time on that here other than to say that we have seen the hub-and-spoke, micro-fulfillment center and hybrid architectures all be effective when supported by flexible automation technology and automation control software tailored to the unique requirements of e-grocery fulfillment. Each of these strategies will likely continue to be used based on a grocer’s retail footprint and store configuration.
Another challenge that hasn’t gotten the same level of attention is the process for making decisions around what products will be fulfilled through the automation system and what products will be fulfilled manually.
There is a natural tendency for some grocers to apply a traditional warehouse mindset to this decision—products that are segregated in the warehouse, such as milk, alcoholic beverages and cleaning products, should remain segmented and picked manually. However, we know that this approach unnecessarily limits the number of products that can be fulfilled through automation, diminishing its value.
The approach that has proven more effective is to evaluate each SKU based on automation-based inclusion criteria. These criteria include the size of the product, its temperature requirements and its cubic volume velocity.
Some products simply won’t fit in the automation storage bins and these can be excluded quickly. Others, such as frozen foods, may require additional investment to automate and the business case for that investment can be evaluated based on the percent of basket of these products. The decision on the remaining products then becomes about their cubic volume velocity—based on their size and sales velocity, is it more efficient to automate fulfillment of these products or pick them manually?
Here’s an example of how that can work. Rather than deciding that all fluid milk should be picked manually, a grocer and their automation partner should evaluate each SKU within the category based on cubic volume velocity. What they’ll likely find is that a large number of specialty milks can be most efficiently fulfilled through automation while the top sellers, such as gallons of whole and 2% milk, will be most efficiently picked manually.
This process change can have a big impact on the success of a project, regardless of the distribution architecture or automation technology being employed. Using broad product categorizations to determine what products are “appropriate” for automation, a grocer may find that half of all SKUs will be automated and half picked manually. When SKUs are individually analyzed based on the automation-specific criteria described previously, the ratio of products being picked through automation versus manually can shift from 50:50 to 80:20.
Considering that automation systems typically have pick rates 5-7 times higher than manual processes, the addition of more SKUs to the system can dramatically impact productivity and order fulfillment times—and create a better foundation for future growth.
Is Robotic Item Picking Next?
With many grocers still in the process of implementing their first automation projects, it may seem a little premature to be thinking about emerging technologies such as robotic item picking. But this technology is already being piloted in other industries and as it gains traction in 2021, grocers may rightly wonder whether it should be considered as a second phase of their automation plans.
The good news on this front is that robotic vision and gripping technologies have advanced considerably in recent years and today’s systems are capable of picking a high percentage of the products in a typical e-grocery order.
The challenge today is more related to the complexity of e-grocery orders. Business rules have to be developed that allow robots to manage this complexity, including everything from packing orders with the heaviest items on the bottom to picking items sold by weight. As is already being done in other industries, grocers will need to work with their automation partners on pilot programs designed to identify exceptions robots may encounter and define business rules for handling them.
There is also the opportunity to use robotic item pickers to supplement human pickers. They could, for example, be deployed to work overnight to pull all of the packaged goods for the day’s orders. Human pickers could then focus on topping off orders with produce and other harder to pick items.
While it’s something to keep an eye on, it’s unlikely we’ll see anything more than limited adoption of robotic item picking in the grocery sector in 2021. What we will see is more grocers adopting automation and those that have already automated continuing to refine their processes to get more out of their investments.
Derek Sorensen is a senior consultant with Swisslog Logistics Automation, Americas