The Hidden Costs of Inaction in Inbound Logistics

Welcome, and thanks for joining us today as we discuss the hidden costs of inaction in your inbound. I'm Shirley Goh, and I'm joined today by my colleagues, Tamara and Scott. Welcome. So when considering making the switch to automation, there might be some hesitation. There's that high price tag up front. There might be concerns about disruptions when you install and integrate. But your most expensive strategy might be maintaining the status quo and doing nothing. Before we dive in further, I want to introduce our speakers. I'm Shirley Goh, and I work in Product Marketing here at Boston Dynamics. Prior to working here, I worked in the research spaces at MIT and Boston College. And before that, I was a staff writer and editor atTheBostonGlobe.Tamara, can you introduce yourself? My name is Tamara Belkina. I'm the director of Product for Warehouse Automation here at Boston Dynamics. I've spent many years in robotics, shaping, scaling various products at Amazon Robotics across the fulfillment and sortation networks there, as well as spending some time at online grocery robotics startup. Outside of robotics, I've also directed strategy for technology and innovation across the public transit and entertainment industries. I'm really excited to be here with you today. Hi, I'm Scott Quigley, senior director of the Customer Solutions and Integration team. So my team is responsible for delivering all of the robots that Boston Dynamics produces to the customer sites. I have over 20 years of experience with different types of robotics, first starting in the government and DOD side, all the way back to bomb disposal robots, and then moving on to commercial warehouse applications, so essentially sortation and distribution systems. So I'm excited to be here as well. Before we talk about the tech and the cost, let's talk about some of the reasons that people want to automate in numbers. Some analysts calculate that there's an average 49% annual turnover rate in warehousing. And a 2026 MHI survey found that 58% of industry leaders named talent acquisition and workforce challenges as the top supply chain challenge, and that's reinforced by the number of warehouse jobs that were unfilled in 2025. That was 370,000. Then there's the issue of safety. In 2024, there were over 252,000 work injuries in transportation and warehousing. So some facility managers look at that rate of turnover, and they think of it as an unavoidable cost of doing business. But it's a really expensive mistake to maintain that mindset. Scott, can you tell us about some of these silent costs that people don't think of when they fail to make that switch to automation? Sure. So I think that high turnover rate is certainly one to be considered. So you have the cost of offboarding any of those people that either quit or decide they want to move forward. You have to go and find all of those new personnel that you're going to bring in. There's typically background checks that are associated with those. There's a training to come on board and actually get people up and running. And then there may be any union considerations, other benefits. There are those costs that go into it. And then there's also just the loss in throughput that comes along with this. Especially as you have people that maybe would just walk off mid-shift, you immediately have a system that goes a line down, whereas having something like Stretch in your facility allows for a consistent throughput. One of the challenges when you don't have an automated facility or you're using and relying on manual unloading is that you are slower to unload the materials from the trailers. And ultimately, you start to incur fees as those trailers spend more time either in the shipyard or at your dock door. So really, the ability to quickly unload those materials, get those trailers off your yard is going to reduce the overall costs that are not necessarily tied to just the product or the throughput but to those additional fees that start to stack up as you have slower throughput in your facility. I think one of the other hidden costs is that a lot of these warehouses are located very closely to their competitors. And if there is a system where-- or a company that has something like Stretch or automation, where it's already an easier job for the users, and they know they're not going to have those repeated heavy lifting injuries or that stress, you're going to lose talent just for people that are trying to go over and work in those facilities as well. So I think there's a number of ways to look at that. Yeah, I think having a predictable source of throughput in your inbound dock really allows you to protect any automation you may have already invested in further in the warehouse so that you're not creating bottlenecks there. And whether it's manual operations deeper inside the warehouse or you may have an ASRS or other type of automation, having a steady stream of product flowing through really makes sure your warehouse is operating efficiently. Yeah, I think that's a good point because you reduce a lot of the bottlenecks or the inconsistency in flow that goes through as well. So there are times where you can see going from an automated process to a manual process, one is faster or slower, and you start to have this elastic buildup within the facility that can slow things down as well. So, Tamara, can you speak to what kinds of warehouse tasks are being automated and which ones are ideal for automation? I think in general, warehouse labor spend has more than doubled in the last 10 years, and it's still growing. The most fundamental reason for automating is just finding the people to show up and do the jobs you need them to do. The types of jobs that are hardest for people to do are the ones that involve heavy lifting, being in environments that aren't really designed and built for humans. So when you think about opening a truck that's been in transit perhaps during a summer season and a heat wave, you open it up, and it's very, very hot in there. The lifting is heavy. People really aren't designed to do continuous heavy lifting at a high rate for the whole shift. So when a person shows up for this job, we've heard some of our customers say, I have someone not show up for the shift at all. Some people show up for the shift and leave after break. So it becomes really unpredictable for facility managers to manage. They don't know how much throughput they can count on during the day because they don't know how much staff they may or may not have. So I think any task that involves continuous lifting, strenuous things, such as lifting boxes out of trucks, that's a really great job for automation because it's something that-- it's a dirty, hard job that people really shouldn't be doing. So tell us about our automated unloading robot Stretch. Yeah, so Stretch was built for unloading trucks. Stretch is a autonomous robot. You drive it into the trailer after you open the trailer. And it's as easy as driving it in, stepping out, pressing Play on a tablet, and then it just unloads it. We work with telescoping conveyors. So Stretch will unload boxes continuously. As it moves further in the truck, the conveyor moves with it. And it will unload the whole truck and then back out when it's done. If any boxes got dropped along the way, it will pick them up as it backs out. But it really removes the really heavy lifting from people. Yeah, yeah, that eight hours a day or more of just unloading boxes non-stop. Yeah, exactly. And there's no wired power there. Stretch is-- it's very easy. You don't have to deal with any wiring. And the battery life lasts for two shifts. Sounds great. So, Scott, what does the timeline look like for installing modern warehouse technology like that? It used to be that it would take months to a year. And I think that that's also a source of hesitation for a lot of people considering making this switch. Absolutely. I think when a lot of sites and customers think about automation, they think about large ASRS systems. They think of all the conveyance that goes into the facility. Those can typically be anywhere from 3 to 12-month installations, and then they have some time on the back end for that to actually bring those systems up and get them flowing in a way that is optimal for the facility. Stretch is built to be a very fast installation. So when we come into the facility, we are there for two weeks total. The first week of that is the installation of the system. So that's getting it onto the dock, mounting the safety fencing and some of our off-robot systems. And then week 2 is where we go live. So we're already picking live freight at the beginning of the second week. We use that time to train the staff that's going to be operating with Stretch. Ultimately, we train those trainers. And then by the end of that week, we are complete in operation and exiting the site. So it's meant to be a very, very quick time on site for you to get operational use out of the system. Yeah, so complete reversal of what it used to be. Absolutely. Yeah. What's the biggest point of friction that customers face when they make this transition from manual to automated? And how do you and your team help them overcome that? So a lot of it starts with education of the users. So there is a change mentality that comes into play here. A lot of it is having to do with the fact that, first of all, you now have an automated system. So it's not something that's going to have to be babysat all day long. You can have multiple systems that a single operator is working with. So it's getting in that mindset of having an operator that's roaming back and forth, as opposed to dedicated to a single dock or a single system. The overall workflow of Stretch in your facility. So how are we going to inbound? What does that do upstream to your different systems or some of the folks that you have along the way? And I think that the primary one is that Stretch has a level of eligible freight that it's able to pick and inbound. So making sure that you're bringing that type of freight to those dock doors so that you're always feeding Stretch with the SKUs that it's able to pick and move into the facility so that you can then orchestrate anything that is maybe not eligible for Stretch to-- either manual doors or some other location. And can you speak to that eligibility just so people understand what would work for a solution like Stretch and what might be outside of that scope? Sure. So Stretch is able to lift up to 50-pound boxes. So anything obviously exceeding that weight limit is something that wouldn't go in there. There is also a minimum and maximum volume or dimensions that we can have. So essentially, anything that fits within that spec that we provide to the users ahead of time, that's the key piece. And then as you start to look at freight, some of the requirements are it's floor-loaded freight. So we are not picking palletized or shrink-wrapped freight. And then as far as the way it is actually loaded into the trailer, there are some configurations and conditions that are not optimal for Stretch to pick at. And maybe, Tamara, you can talk a little more about that one. Yeah, so Stretch won't be able to pick canoes out of a trailer. So any extra-large, oversized items, maybe someday, but not today. And things like poly bags and totes is something on our roadmap. But yeah, it's really we're sticking to cardboard boxes. Yeah, and that's still a huge share of what people have. There's always going to be bags of oranges, but there are so many cardboard boxes. So, Tamara, one more question before we take a short break for Q&A because we have a lot of questions coming in. Some companies want to wait for technology like automated unloaders to mature a little more before they invest in that. Maybe they want to wait two or three more years to see how it shakes out. What's the cost of waiting for that? What's the cost of not jumping in right now? Yeah, that's a great question. I think a lot of times, the mindset is, well, automation can be expensive. So by delaying purchasing it, I'm not spending money. But that's actually not true because labor wages continuously go up. And so by not automating some processes, you're actually incurring compounding annual labor costs. And so the longer you wait, the more costs you incur, as opposed to if you invest in automation, then you're solidifying your costs up front that you can rely on, as well as predictable throughput. So I think that's one way to think about it. Another one, there's also different kinds of inaction and waiting involved. You can decide to delay purchasing at all. Another thing that we see some companies do is they treat it as an experiment in one corner of their warehouse, where they will purchase one robot in large multi-dock operations. They wait a year or two to see how it performs before really rolling it out further. The customers that we see that get the biggest ROI and see the largest success, they take a completely different playbook, where they have very targeted, systematic approach for, OK, here are the throughput rates and the type of performance I need to see in order to roll this out network-wide. They do a very targeted test where they, at a smaller scale, test to make sure they're getting that throughput out of the system for a few months. And then once they've validated the performance, they go really all out in a system-wide installation across their entire network. Because I think there's also a hidden cost to fragmenting your operations where most of your staff still is unloading manually, but then we have this robot that we're playing with, but not everyone is fully trained on how to use it if it's just in one corner of the warehouse, as opposed to if you make it a consistent type of training, then you get the consistency of throughput that the rest of the warehouse sees. Yeah, and I think some of that, too, is as you get further down into the different operational groups, you can get a siloed mentality where it's, my piece works fine, so everything's OK. So we don't necessarily need to automate. But if you can take that top-down view, you can see where maybe some of the other groups are over-indexing on solving a problem without automation or without working together where you could be more holistic across the entire site. So with an automated solution like Stretch deployed across North America and Europe, how much freight is that moving? Good question. I can give you an answer now, but also know that as Scott's team is out there deploying more robots, the number keeps changing. So right now, it's at about 700,000 cases per week. By the end of the year, I expect will be at least a million cases a week, if not more. Yeah, and I think one example of that is in a recent deployment. Over a pretty short period of time, we moved just under 7 million pounds of freight. So again, that's 7 million pounds that associates are not moving along. And because we'll use anything but the metric system, we equate that to 22 blue whales when we're talking about this internally. So it was really good. More than 5,000 tiger sharks if you like those. Easier to envision. Yeah. Absolutely. It's a lot of weight. And just imagining people not having to do that job is pretty powerful. Yeah, yeah. All right, so we're going to take a short break. We're going to get to some of your questions. I'm told that we're getting questions about Stretch's working speed and also about the possibility of Atlas in the warehouse. So stay tuned. We'll be back in just one minute. [MUSIC PLAYING] Here at Arvato, we work closely with our customers to fulfill and solve a lot of the challenges that they have with their logistics and supply chain operations. We had a vision for completely automating a carton from loose loaded in a container all the way to the outbound process. We already had the scanning and systematic pallet building and sortation. Stretch was that last puzzle piece that we were missing to have a true automated loose load/unload solution. Stretch can unload everything that we bring in in a loose loaded form. It allows us to improve the back end, the sortation, with just a little bit of our logic change. And we're able to really achieve higher throughputs than we ever imagined. [MUSIC PLAYING] Welcome back. We're now going to answer some of the questions that you both pre-submitted and that you've been asking while we've been talking. So, Scott, this first one is for you. We've spoken about how turnover on the dock is notoriously high. And we've heard stories about customers who walk off the job at a moment's notice and don't come back. Walk us through the math of what it really costs a facility to constantly recruit, onboard, and train that staff just to keep the dock moving. Sure. So I think one of the best or most extreme examples I've had with a client is that their facility saw a 70% turnover month over month for the associates that were actually moving material. And early on, they didn't focus on automation. They actually focused on their onboarding process so that they were able to get people in the door faster and get them operating within the first day. So really, the money went into offboarding people that had either walked off the site or quit or onboarding new personnel and working through the staff. So again, we talked about sometimes the outside processes are spending more money to eliminate some of the challenges internally. Putting something like Stretch in place could really kill two birds with one stone. So I think that a lot of the challenges and a lot of the hidden costs comes in through just that, finding the personnel, bringing them on board. There's background checks associated with any new staff. There's the time to bring them in. In a lot of locations, we're talking about multiple different languages. So training materials have to be translated to all of those languages. In the case of union or worker councils, you have to have associates that can also speak those languages and interact with those personnel. And then there's really just the unknown of when they're actually going to continue to work or maybe not show up that day or leave right in the middle. So again, you don't know how often you're going to have a breakdown in that throughput through the facility. Yeah. And I've heard the cost to-- some companies, track the actual cost that it takes to onboard a new associate, and it's thousands of dollars. So it's not not consequential, especially at those high turnover rates. And if you're talking about doing that at a injury or health and safety point of view, workers' comp costs are even more. Now you're talking tens of thousands and a significant amount of those happening throughout the year, especially in the unloading side of the house. Yeah. And a solution like Stretch requires no background checks. We've thoroughly vetted it. We've gotten some questions also about, how do you handle containers that are messy when you open them up, and there's boxes all over the place, and they're not all stacked neatly? How do you, as a product designer, plan for that? Yeah, we get that question a lot. I like to think of unloading a trailer full of boxes like a high-stakes Jenga game, except you actually don't know how heavy each box is. And the dimensions are all different, so it's not equal-sized bricks. You also don't have the luxury of looking around corners and seeing what's actually interlocked and interconnected and what could fall down if you pull something out. It's also high stakes because Stretch is still expected to meet those high throughput rates. We have seen a ton of advancement in AI technology to help us with that. So as Stretch enters a trailer, we use AI to determine the dimension of the box, as well as predict which box to pick next to prevent an avalanche or a facade of boxes from falling down. That's important for us so that we can keep the pace going and hit those consistent throughputs that we've talked about earlier. I think there's also the hardware component of it. These trailers, they're in transit. Sometimes they're loaded really neatly, but then in transit, things shift. When a person is inside the container, people are very good at instinctively figuring out which box to take next, maybe get some quick feedback. Oh, this one feels stuck. Let me try another one. However, they do still make mistakes because they don't see what's behind that first facade. And so the impact of a box falling on a person-- and some boxes can be 20 pounds, 30 pounds, 50 pounds-- that's a really high-impact event. There are injuries and risks associated with that. The way we designed Stretch from the hardware standpoint, it weighs a ton, literally. It weighs about 1,300 kilos. And so if a box falls down, Stretch is able to recover. It continues working. It picks up the box that fell down and moves on without getting hurt. So I think it's both the software component and the hardware, if it makes sense, that help us react to things that shift. Yeah, yeah. And then, Scott, there's questions about the experience of people who operate the robot. What's your team hearing from operators after Stretch is deployed? Do they welcome the robot, or are they hesitant? And what does orientation or training day look like for the staff? So I think it varies based on how much automation might already be in a facility. So somewhere we're coming that maybe this is the first thing that's coming in beyond just the inbound conveyor, there's a natural hesitancy to use this. And is it going to be as good as the people we have? And there's always maybe the initial want to dismiss this as it's just a new thing that's coming in. Very quickly, as the users begin to operate Stretch, get to know it, and realize how much of a force multiplier it is for them in their facility, they really become advocates for it. I think one of the operators commonly referred to Stretch as her baby. And if anyone else in the facility would speak ill of it, she would begin to defend it like it was a child. You begin to fall in love with it as soon as you understand how much of a benefit it is to you and your coworkers. As far as the training, so like I said earlier on, week 2 of an integration is really we're going live, and we're starting to pick freight and bringing it to the facility. So that's the time that we use to train the operators that are going to be working alongside Stretch. We operate with a train-the-trainer model. So the team that is going through that orientation and training, they're getting certified by Boston Dynamics as trainers for the Stretch system. So as you have new employees come on or someone gets moved over to that side of the building, they're able to train their own personnel and certify them as users. And then that way, they can expand their own workforce without having to re-engage or come back to us in any way for that. Another interesting aspect of advancements in AI is that for those folks who look at trailers and think, oh, my trailers look different, they're a little messier, it's actually super helpful to get started earlier because with AI, we get real data once we're deployed at your site. And so we're actually looking at the unique problems that your site sees in their trailers. And then we're training our models based on the data in your real trailers and make Stretch adapt quickly to being able to handle that freight. So it's also an evolving story. It's moving very quickly. The pace of progress with AI is moving very quickly. There's also an added benefit there of moving faster with deployments because that lets us tackle your problem faster. And I think a part of that, too, is the physical infrastructure or physical installation doesn't change. This is a software update to improve the perception and the features and the thinking. So waiting doesn't necessarily get you a benefit. You could have the benefit of the system today that gets obviously more impactful over time. So, Tamara, we also have a number of questions about how fast Stretch works. And does it work as fast as a person? Yeah, that is a question I get asked a lot, too. It varies. We see anywhere from 500 to 1,000 cases per hour, and it really depends on what it is. For example, if we have a customer that has boxes of heavy jeans, then Stretch adapts, and it will be in the lower end of that range so that we avoid any spillage or dropped boxes because-- And they're really heavy. They're very heavy. 50-pounds box full of jeans is a very heavy box. A lot of our other customers have cases that are smaller, and Stretch proactively multi-picks, I've seen, three and four boxes at a time. We can also singulate them and drop them one at a time if the facility needs that downstream. And so it really depends on what's in the trailer. But we're always making software improvements to keep adapting based on what our real customers are seeing and to make performance even higher for the real trailers that they have in their facilities. Mm-hmm. And how does that compare to the way a person unloads? Yeah, so a person could start out with faster rates. However, there's an element of fatigue. So those trailers are big. They take hours to unload. And so as a person gets further in their shift, the rates tend to dip lower because people get tired. It's a very hard and physically demanding job to do. Trailers can get really cold in the winter and really hot in the summer. So that also adds to speeds dropping on and off throughout the shift. Stretch does not get tired. It does not get overheated. It was designed for these environments. So it really gives you a consistent throughput and consistent speed throughout both shifts that it can work on a single battery charge. I think some of it too is not just the speed of the person unloading. It's the speed of the facility and what they can do with that. One of the things that we see as we start to get up to those higher CPH around or near a thousand is we very quickly can jam up the other systems that are inbounding. So some of that is more-- You always look to the person first. Can I be as fast as a person or faster than a person? But if you can achieve that, you can very quickly start to be faster than the building itself. So that goes back to, again, how you look at automation across the whole facility and understanding where your next choke point or friction point comes from as well. Yeah, absolutely. And we're getting into conversations where a lot of our customers have more than one type of truck coming in, some with maybe really easy-looking freight. Some could be more challenging. And they're actually proactively making decisions of, let's put a truck that we know Stretch can get into at or even above 1,000 CPH right next to the truck that has more complex rate to balance out the downstream automation so as not to create bottlenecks. So it's really important to look at holistically. You can't just say, oh, well, if you only did 2,000 CPH, my ROI would be so much better. But in practice, that could be meaningless if your downstream sortation can only handle a thousand across a couple of docks. So that's holistic approach is important. We have some viewers who also want to know whether Atlas can do truck unloading. Yeah, we get that question a lot. We are so, so excited about Atlas. We're building it as a multi-purpose robot that can do anything that a human can do. While we're very hard at work developing that, it's still some time away from being a product you can purchase today. However, Stretch is a product that is available right now. I like to think of it as our customers being able to learn about us as a company earlier, seeing our products that are available today doing real work in their warehouses. Also, it's an opportunity for us to understand things what kind of freight your building gets, as well as what other warehouse automation tasks you may have in that facility that are priorities for you. It may be truck unloading. It might be other things that Atlas may be a better fit for. So it's a way for us to start the collaboration earlier for folks to get started with Spot and Stretch that are available today to solve real problems in the warehouse. And we're absolutely excited to keep talking to anyone about their potential applications for Atlas within a warehouse, logistics, or any other industries. So if someone is considering making that transition to automated unloading, but maybe they need to convince someone on the accounting side, what's maybe the one data point for each of you-- I'd love to hear both your viewpoints-- that you think a viewer can take with them today to convince their team? I would look at it as a fully burdened cost per case, and that includes not just what your labor costs are today, but also the expected growth over the next three, four, five, et cetera years. I would also add in, what are your turnover rates? What are your costs to retrain someone? What are your costs to help people who may get hurt with a work-related injury doing those hard jobs? And so really, doing that whole holistic math and then dividing that by the amount of cases you get gives you a more realistic number of the cost of manual operations. The other thing I'll say is also, it doesn't have to be a big upfront cost. Some of our customers prefer to buy with CapEx, but we also have robotics as a service, RaaS, option as well. So if that's something that your finance team prefers, we are very open to that, and that's available as an option. No, I think Tamara nailed it. But I think that looking at it across the entire organization is the key. Stretch is an unloading robot, but unloading is not the only thing it's going to impact. It's going to impact your staffing, your training. It's going to impact downstream process. And I think really looking at not just the cost of the robot, but also just simply how leveling out the flow through the building and making it more consistent and predictable is going to help your business is really a key piece to look at when you're making these decisions. Yeah, I think another point I hear is that if you own the inventory that's just sitting there in the truck and you're delaying, there's a cost associated with delaying, getting it inside, processing it, and being able to pick real orders from it. And that also can be a factor. But also, I think that some of the product that you're moving, too, has expiration dates. Some of those things can be wholesalers that have-- essentially, a clock starts the minute it hits the yard before they have to discount or liquidate. So really, the ability to get that in the door and back out again is key to making sure that they're maximizing as much as they can with their profit margin. All right, I'm afraid that's all we have time for today, but I want to thank all of you for joining us and also Tamara and Scott. If you want to learn more about Stretch or our other products, be sure to visit our website at bostondynamics.com. Thank you. [MUSIC PLAYING] PLAYING]