Release the Robots
Collaborative robotic solutions are safer, smarter, and more efficient than ever. They’re also the perfect partners for human workers.
By Seth Winterroth
There is a new reality in the world of robotics: commercial viability. Robots are moving beyond cages into less controlled environments where they are doing varied work safely, reliably, and at a cost basis that makes economic sense. The media devotes a disproportionate amount of attention to “holy grail” applications like autonomous vehicles and flying cars. However, real businesses will be built in the next five years by those who leverage the strengths of robotics to deploy practical, full-stack solutions into more controlled environments where work can be done reliably and economic value created. I believe retail fulfillment is an ideal environment in which to deploy these types of focused autonomous solutions.
E-commerce shifts the supply-chain model
Traditional retailers and third-party logistics companies have designed their supply chains around distributing cases containing large quantities of the same item to brick-and-mortar locations. The bulk deliveries are unpacked and shelved by store associates and customers travel to the store to make their selections. E-commerce shifts this model substantially, putting much of the burden of fulfilling a customer’s purchase on the fulfillment operations.
And what a burden it is. E-commerce fulfillment requires more individual picks and touches than traditional fulfillment in order to ship more individual total parcels to customers from the fulfillment center (FC). Here’s what the typical E-commerce workflow looks like:
● Pallets stacked with cases are delivered to FCs and de-palletized
● Each case is opened and its contents are processed, picked and shelved
● After a customer submits an online order, FC associate #1 picks individual items from shelves and processes
● FC associate #2 sorts and processes
● FC associate #3 packs individual customer’s order and processes
● FC associate #4 further sorts packed parcels and prepares for shipment to the customer
● Customer’s parcel is shipped from the FC to local parcel hub where it is processed, sorted, and sent out for local delivery
The increase in individual item picks results in a corresponding need for more labor to meet the need. For a supply chain that is already breaking at the seams, this shift from a distribution to customer-centric fulfillment model is hindering growth. Robotics solutions can, and must, begin to be applied to many of these steps if E-commerce is to continue to expand.
In 2016, E-commerce retail was a $390 billion market, growing approximately 15% YOY. That just scratches the surface of how big E-commerce will become as it is still only 8.2% of the total $4.8 trillion traditional U.S. retail market. As E-commerce grows, it will demand more of what are already scarce commodities, from optimal FC real estate to the workers that will staff them.
In 2016, Amazon (which accounts for 3.3% of the U.S. retail market) employed 115,000 full-time equivalent (FTE) employees and 100,000 seasonal workers. This year they have publicly announced plans to hire thousands more warehouse associates. If Amazon grows to 15% of total U.S. retail, it would need 538,000 FTEs and 468,000 seasonal workers. But, there is already an E-commerce labor shortage. As competitors across this landscape push to scale fulfillment efforts the competition for labor will only increase. Amazon has sharply raised customer expectations for fast fulfillment; they have been able to achieve this by implementing automation solutions. When Amazon spent $775 million on Kiva Systems, it was a clear signal that robotics was going to be key to their growth strategy. At every turn, Amazon enhances their ability to serve the customer and competitors in this market are struggling to catch up. Unless retailers push forward with an aggressive automation strategy, they will be unable to compete.
A closer look
Robots have been able to do single actions with great efficacy for some time. But performing varied actions with industrial performance requirements has eluded the robotics world. The task profile for today’s labor in fulfillment centers is largely rooted in this problem. E-commerce needs an automation solution that can pick multi-geometric objects with the speed and reliability of the average human laborer. FCs also need low-cost, flexible robots capable of autonomously navigating the environment and moving items from point to point. These types of solutions can be leveraged to collaborate with human workers to increase the overall throughput, or the rate at which items are picked through each stage of the fulfillment workflow. A reliable increase in throughput, at a cost-basis with favorable economics, through automation solutions will allow E-commerce fulfillment to scale.
Kindred Systems, a company in which my firm, Eclipse Ventures, is an investor, is deploying a commercial automation system that directly addresses many of the scaling issues that FCs are facing. The company’s approach to training a machine to detect, classify, plan a path to and then manipulate objects has been the most effective that I have witnessed to date. Current Kindred systems can perform a high percentage of an FCs sorting workflow. They are designed to work side by side with staff to increase productivity, and are adaptable to new and existing FC processes. A flexible, infrastructure-light design results in an easily implemented systems that can be scaled up or down depending on seasonal demand. Kindred’s systems leverage state-of-the-art machine learning enabling each individual robot to become more intelligent over time and require less human oversight.
Another Eclipse investment that bears watching is 6 River Systems, a Massachusetts startup founded by a team that includes two former Kiva Systems executives. Like Kindred, this technology startup has an industry-problem-first approach to building the technology needed to make fulfillment more efficient. The technologies that enable mobile autonomous systems has progressed a great deal since 2003 when Kiva Systems was founded. We’ve seen substantial advances in sensing, edge-compute, computer vision, and cloud integration; moreover, we’ve watched the cost of these capabilities decrease dramatically. As a result, 6 River is able to develop autonomous vehicles that require no on-site infrastructure in order to localize, navigate, and actuate reliably. This makes for a highly flexible, low-cost solution that can be easily implemented in large scale FCs, metropolitan FCs (think Prime Now centers), or even within brick & mortar retail locations to enable omnichannel capabilities. Furthermore, the collaborative nature of the system was designed with the intent to enhance, not replace, the FC associate. With combined backgrounds building, selling, and implementing thousands of robots during their time at Kiva, the 6 River team has a clear understanding of what it takes to work with customers to operate well-designed, reliable automation solutions for fulfillment. They understand that, for customer-centric fulfillment to scale, a new, automated, digital supply chain is required. 6 River Systems is building the end-to-end solutions to make this possible.
Thanks to the promise of a new breed of technologies, the field of robotics has never looked more exciting. There are real, commercially viable automation-use cases being built today. I believe that, by deepening their understanding of the workflows in semi-controlled environments like E-commerce fulfillment and manufacturing, robotics innovators will build successful businesses in the next 5 years. These efforts will advance fundamental technologies, facilitate greater commercial adoption, navigate regulatory hurdles, attract alternative capital sources to the ecosystem, and amplify public awareness of the value that can be derived from well-designed robotic systems. Furthermore, these successes will lay the foundation for how to commercialize solutions to more challenging autonomy problems like mobility (land and air), logistics in uncontrolled environments, and advanced machine intelligence for human assistance.