Exploring Warehouse Automation

January 25, 2024

Jon Schechter (VP of Fulfillment Solutions, AutoStore) joins Karl to discuss robotics and warehouse automation.

Details #

Jon Schechter (VP of Fulfillment Solutions, AutoStore) joins Karl to discuss robotics and warehouse automation. Listen in to learn about the acceleration of robotics in the past decade, automation challenges and how great people and processes accelerate ROI. Then hear from Jordan Lawrence (Flexe Director of Logistics Strategy) and Ben Dean (Flexe Sr. Director of Network Strategy & Optimization) as they discuss the rapid growth of warehousing and automation solutions, automation’s ROI and future innovations in the space.

Some of the topics explored:

  • The rapid acceleration of robotics over the past decade

  • The benefits of automation, from maximizing space to meeting customer SLAs

  • How investments in automation intersect with 2024 labor trends

  • How great people and processes accelerate robotics ROI

Logistics Leadership Podcast legal disclaimer

Episode Transcript #

Karl Siebrecht 0:00

You know, robots are sexy. They're cool. Let's go get some robots.

Jordan Lawrence 0:04

I'm Jordan Lawrence.

Karl Siebrecht 0:05

I'm Karl Siebrecht.

Ben Dean 0:06

And I'm Ben Dean. This is the Logistics Leadership Podcast.

Karl Siebrecht 0:19

Welcome back to the Logistics Leadership Podcast. Guys, good to see you again.

Jordan Lawrence 0:24

Karl, always great to be with you. Ben, good to see you again, as usual.

Ben Dean 0:28

Likewise, Jordan.

Karl Siebrecht 0:29

So in our last episode, we talked about onshoring and nearshoring, which is a cool conversation. And you know, one of the dynamics that came up was that relative scarcity of labor in the U.S. and the high cost of labor in the U.S. is acting as a drag on adoption in general, compared to other markets, markets where manufacturing currently exists, it's just more expensive. And one of the potential solutions to bring that barrier down is the adoption of warehouse automation. So in this episode, we're going to explore automation. And I'm going to speak with Jon Schechter, who is the VP of Fulfillment Solutions at AutoStore. He has many, many years of experience in the space, has kind of seen this thing evolve from its early days. So, very excited to speak with Jon, learn about the industry from a macro perspective, and also understand some of the dynamics in terms of adoption, barriers, etc. So are you guys ready for this?

Jordan Lawrence 1:29

It's gonna be a great conversation. I'm excited.

Ben Dean 1:32

Let's do it.

Karl Siebrecht 1:32

All right. Well, I'll talk to Jon, and we'll see you back here shortly. I am really fired up for our guest in this episode, Jon Schechter. And we're going to talk all things warehouse automation. So let's kick this off. Jon, if you wouldn't mind, give us a sense of your career and your background and kind of what you've been up to, to get you to this point today.

Jon Schechter 2:01

Sure, Karl, great to be with you. I'll work backwards. I'm presently at AutoStore, one of the leading automated storage and retrieval systems. I'm leading new projects in North America across our business development team. Before that, I spent a couple of years piece picking robotics, which is kind of cutting edge manipulation of items out of something like an AutoStore, or within a warehouse for a different job function for a subset of inventory items. I took a couple of years out of the industry to do business school. And before that, I spent four years at Amazon really helping them scale out their acquisition of Kiva, and get them from zero to the first 500,000 robots that were in the network. I really was lucky to start my career at Kiva Systems, graduated from MIT and went there for four years before that Amazon acquisition, and learned a lot about solving customer problems and how robots can help in a warehouse and the kinds of technologies that are available to end users today and their supply chains.

Karl Siebrecht 3:02

That's fantastic. What, 15 years of professional experience in warehouse robotics? You've got, obviously, tons more expertise than I'll ever have. But that seems kind of like almost the full lifecycle of production level warehouse robotics. Is that a fair statement?

Jon Schechter 3:23

Yeah. I was really fortunate to pick this industry or this career path. I think there's been a lot of technology and automation over a long period. But in the last 15 to 20 years, we've really seen a birth of new kinds of robots, not the six axis robots, and not this kind of slow moving large, you know, automated guided vehicles that were using different navigation or the very fixed crane systems, but now a whole diaspora of multi robot fleet systems driven by software, connected over WiFi or other wireless technologies, and really taking on a lot of new tasks within supply chains. So that's been amazing to see and enabled by a lot of the other mega trends like wireless connectivity, increases in battery technology, scale down of compute, and just good hardware production. So there's a huge number of new tools for supply chains.

Karl Siebrecht 4:12

Got it. Just taking a very broad perspective of the warehouse automation category, can you kind of break that down for us into the different subcategories, kind of this mental map of how to think about warehouse automation in its different segments?

Jon Schechter 4:33

Absolutely. So you have equipment like an automated storage and retrieval device that can take in inventory from new receipts stored on the shelves or into a special storage area, and then bring that out to operators as it's needed. This has existed for a long time for pallets and cases in the beverage industry and other large formats. And in the new realm for serving eCommerce and each level distribution, there's a lot of new tools that can handle a bin or a split case, less than a case of inventory, as well. Beyond that, there's different types of mobile robots that are doing, sort of, conveyance of inventory. There's piece picking robots, like I mentioned, that do manipulation, maybe at an ASRS at a picking station, maybe inducting to an auto bagging, packaging machine. And then you have all kinds of sorters, conveyors, pouch, or pocket sortation systems that can really convey or move goods, whether it's new receipts, finished products, or customer orders, over to a dock door, over to a different processing area. So really mapped all the touches with different equipment connecting them. And there's still people, a tremendous amount of labor content in these fulfillment centers and distribution centers. And so people are really at the heart of getting this work done. And these automation tools are there to serve their work and make their job function a little easier.

Karl Siebrecht 5:58

One of the things I've heard a lot about in the context of warehouse automation is, some of the lessons learned and mistakes that companies had made in the earlier days and, I don't know, sort of like getting excited about automation technologies and working to deploy them without yet having a stable process. And you're sort of looking for, I don't want to call it a quick fix, but you're like, it's a shiny object, let's get some robots, and this thing's going to rock and roll. But if you don't have the stable process where you understand the operation and how automation would drive an economic needle, you're kind of cart before the horse. Is that a fair assessment?

Jon Schechter 6:39

That is so spot on, Karl. I see a lot of operations that have had failed innovation projects where someone got excited with a little extra budget, went out and bought one robot or point solution. And they brought that in and it didn't work so it got shelved and, you know, collects dust today. I think of it like Maslow's hierarchy. Of course, you need excellent people and processes first. You can do amazing things with just a good leader with a good team underneath that's motivated, that knows how to get the job done. And that's the level one. I think on top of that is then mapping all of that expertise to software flows. So you have things like enterprise resource planning or ERP, warehouse management software, WMS, that are essential to be able to match those physical flows with virtual flows in software so that there's visibility to what's happening and you can track down issues, you can really measure and monitor processes better. Then you've sort of entered a realm where you can start to augment with automation to find where are the areas for improvement? Where are the places where we're spending inordinate amounts of labor, or we don't have enough resources, so we need to augment with systems and equipment like we're talking about today. And you bring those things in, in the right way, in a sort of incremental approach. Rather than trying to take a warehouse with no software and just throw robots at the problem.

Karl Siebrecht 8:03

Right. I've had the great pleasure of getting to work with a number of folks from Amazon, and then living in Seattle, I know, many, many, many more and have gotten to see that journey a bit. And that's one of the things that has popped out very consistently. Almost at its fundamental DNA level, Amazon, at least on the logistics side, they're operators, operators, operators, and then look to apply technology to the operation, rather than, you know, of course, they are a great technology company, but it's the DNA in the first thrust of that business is operator operator, and then figure out how to apply the tech, which seems like a great lesson to learn.

Jon Schechter 8:41

Yeah, absolutely. The best companies at adopting automation have already been great operators. Amazon, one of the best, they spent, you know, their first 15 plus years with almost no automation. It really wasn't until, say, the 2000s, when they started using a lot of sortation. And then in 2012, when they acquired Kiva, when they started to roll out automated storage and retrieval through the Kiva systems, to then augment and take away some of the picking walking time and put away walking time. And then that's allowed them to cut their productivity by, say, the total spend by 20 to 30%. They follow that trajectory I talked about before in terms of people and process, warehouse software that mirrors that, and then layering in different kinds of automation to execute well. And now, of course, as they're scaling further, that's a huge part of what they do going forward is more and more kinds of automation. But they earned the right to be there, for sure.

Karl Siebrecht 9:38

Let's talk a little bit more about the benefits or selling points of automation. I'm an operator or I'm a 3PL or I'm a CPG company or whoever, and I'm considering doing this. What do I think about the metrics that will guide my decision making, should I automate this or not?

Jon Schechter 10:01

You know, most warehouses hardly use most of their space and labor potential. So the first thing to look at is, when you're at these peak times and you're sticking inventory everywhere, maybe there's an automation solution that allows you to fit a lot more inventory and use more of the vertical height and the 3d space available to you. That's number one. Most warehouses that say they're quote unquote full are not nearly full, they're actually using maybe five to 10% of their cube potential. I think the second area, and it comes down to the demands being placed on these warehouses is, are you meeting customer SLAs? In the traditional fulfillment model, you know, one to two to three days cycle times was acceptable when you received an order. And that's just not the case anymore. That needs to be on the customer's doorstep by then, which means you have mere hours to get those units packed and out the door to the end customers. The third, of course, is how are you leveraging your people? Are you asking people to walk 10 miles a day? Because that's not a valuable use of their time, it's not a good way to retain talent. And it's certainly not a sustainable way to use employees. And so goods to person is one of the solutions for reducing that walking and really empowering them to get a lot more of their work done, which is really filling orders, not wandering a warehouse. And the last thing, of course, if you're using your space and your labor most efficiently, it's just going to be much more sustainable from an ESG standpoint. You're going to use less energy, you're going to be, you know, instead of building a new, larger facility, you're going to fit more in the warehouse you have, which is a literal, smaller footprint. And that's important to almost every supply chain I talk to, not just the transportation emissions and carbon footprint, but the footprint of the facilities that they have in their network.

We were talking about labor productivity and just a more effective use of the human skill set, as opposed to, sort of, human acting like a machine, have a human do the value added human bits. It's getting to the point where when you're competing for labor, if there's a job down the street where they could work alongside automation, that might be preferable. It sort of starts to create a requirement for more and more companies to adopt automation, just to compete for labor, almost.

Yeah. So it's really important that when employees that you hire show up to work, they're given at least the level of tools that they expect from their personal lives. Everyone you're hiring has smartphones, has computers, has automation in other forms that is at their beck and call. So they're expecting that in the workplace as well. If you've got them running around with clipboards and pushing a cart and their friend is down the street at another warehouse with an automated system that's doing some of that work for them and they're just working through a beautiful UI workflow that's helping them get their job done in an ergonomic workspace, they're gonna go work with their friend, 10 times out of 10.

Karl Siebrecht 12:45

Are there any stats that you're aware of? What's the state of adoption?

Jon Schechter 12:49

The belief right now is that about 15% of warehouses globally, in, say, the top 50 economic countries, have any level of automation. There's a huge amount of manual warehouses that are still moving things in a manual way. So there's a huge amount of opportunity to do better in our supply chains.

Karl Siebrecht 13:08

Got it. Is that total any type of automation as you described about the category earlier, kind of automation and all its different flavors?

Jon Schechter 13:16

Yeah, that is our estimate for warehouses that have, say, beyond conveyors and forklifts. So any kind of mobile robots, automated storage and retrieval systems, automatic sortation and packaging equipment, that haven't adopted those things yet. So most of the brown boxes showing up on your doorstep, the boxes were built by hand, somebody walked around the warehouse to shop for those items to get them in there and they were sorted out with a paper pick slip, believe it or not, and put in your box by luck. So if the quality of your order wasn't achieved, then it's possible that it came from one of those lesser warehouses, so this becomes now a competitive edge, and that your customers feel in a very direct way when you make those mistakes. And the root cause will actually be because it was from a manual facility that was impossible to get that job done correctly because there's tens of thousands of items and hundreds of thousands of orders a day. You just can't do that without good software, good automation, of course, on a base of good people and process.

Karl Siebrecht 14:15

It's early on. Okay, got it. Hard to get to a specific number, but order of magnitude 15%, call it. Are implementations happening more in greenfield warehouses or retrofit scenarios with existing or is there any way to think about the split? What do you see most of?

Jon Schechter 14:34

About two thirds of the projects we do are actually retrofits or brownfield situations. We just published an article about that, because it's such a huge part of what we do, and a lot of people are asking the question, is my warehouse compatible with new automation? And the answer 99% of the time is yes. There's some building preparation to make the floor flat enough and to maybe augment the fire suppression system. But beyond those things, most facilities can handle automation. Now some of the newest sites, somebody's rejiggering their network design and they need a new facility, a new place. And so it makes sense to build from scratch. But for many warehouse projects, the ROI is driven just on the fact that they can leverage the asset that exists today, and not have to build a whole new facility.

Karl Siebrecht 15:19

Presumably, for these brownfield implementations, you know, these are operating facilities, like they've got to ship the stuff tomorrow, next week, the next week, and then you want to layer on top of that a project, which is, while the train's moving down the tracks, so to speak, we're going to layer in some new automation. And that's part of the deal. And presumably, that must be very doable if two thirds of the implementations are in these facilities.

Jon Schechter 15:45

Yeah, what's cool about the newer technologies is they're highly modular. And so you can do a phased rollout. Of course, you need a little bit of wiggle room within the facility, I think of it like the nine piece puzzle where you've got one slot. But if you can build a chunk of an automation system, move a lot of inventory into it, now you're denser. And so you've moved maybe two puzzle pieces worth of inventory, and all of a sudden, you've opened up all the additional space needed to finish the project. So we'll see a lot of customers and brownfields do that in, say, two steps. And then they're only building ahead to that next peak. They can augment and add to the systems every year as they need to meet their growth projections, which are really uncertain. And so they don't need to make those decisions years in advance. They only need to make them, say, six months in advance of that upcoming peak period.

Karl Siebrecht 16:29

Yep. So let's talk about what you've seen as some of the big challenges. So once I've done an ROI case, and I want to do this, when I start thinking about implementing, what are the big challenges and the watch outs?

Jon Schechter 16:43

Yeah, I would say the main thing is assembling the right team within your organization. So it can't just be the operations team that asked for this, it can't just be an industrial engineering group that's pushing for it. And of course, the CFO and the financial team needs to be on board with funding this project and converting OpEx to CapEx. The maintenance team needs to be ready to handle this equipment like any other equipment. The HR team with how you use this to recruit and retain talent. And the software team, the product management teams that are thinking about how this integrates the warehouse management software, how it talks to other subsystems, how it complies with information security. All those stakeholders need to have an opportunity to review and understand these projects and be part of them from day one. Those projects end up being the most successful, when all those people are aligned that this is a good decision for the organization. It can be really, really frustrating if you don't have those things aligned. That's why we still see 85% of the warehouses out there manual, because they can't get that kind of stakeholder alignment that unlocks this type of potential.

Karl Siebrecht 17:48

When do you come across conditions where you're like, you know what, it doesn't make sense for you in this facility or at this time?

Jon Schechter 17:54

Yeah, I would characterize it typically in two major camps. First is complexity of what you have to get done. And the second is scale. And so on terms of complexity, if you have a thousand SKUs and a thousand orders a day, that's enough order complexity that your team might need to think about automating. If you have less than that, it's pretty manageable. Someone can memorize where those locations are, and can get the volume out maybe by themselves. So we see really small sites that can be incredibly efficient. And automation doesn't add a ton of value there. In terms of the scale, if you have less than 10 people doing your pick, pack and ship process, it's probably not enough scale to justify the types of investment. Because of some of the fixed costs, like the software integration, and the equipment itself, you're just not going to have the payback there, unless that service level agreement that you're trying to reach is really essential and you can make a business case that says, it's essential that I ship with a cutoff of midnight. We've seen a lot of customers change from a two day cutoff to like, if you order today, I will get it to you tomorrow. So if you can make that business case on the total customer experience, then you might be able to even go below that 10 picker threshold.

Karl Siebrecht 19:03

That's great. A thousand and a thousand. And that works from an ROI perspective, too, because presumably, you get x dollars per unit economic benefit, and across a thousand units a day that will pencil out an ROI that makes sense.

Jon Schechter 19:18

We're talking about 20 to 50% labor reduction for fulfillment. So that can be, you know, 20 cents all the way up to $1 depending on your current fulfillment costs and what kinds of items you're shipping. That can start to justify these systems just on itself. And then you layer in the other economic benefits of space savings, the energy reduction, and meeting that service level, which, you know, is a little bit harder to quantify. It's outside of the operations budget, but it's really about what does the marketing team need to compete online and get demand? What are the kinds of SKUs you're gonna offer and being able to add SKUs, it's really easy to get to a thousand SKUs, especially if you're in the fashion business. And so SKU complexity is always growing if you're in a successful business, so it's really easy to meet those ROI thresholds.

Karl Siebrecht 20:04

That's the other key is robots themselves are modular, you can just add more robots as you need, as your scale and/or complexity increases.

Jon Schechter 20:12

You can certainly scale up very easily. You can reallocate the equipment to different facilities. For certain equipment it's easier to do that than others. But you can move robots from one grid to another if you have a network of facilities, like DHL just announced, you're talking about a three year ROI for most projects, even down to a two year ROI. So most supply chains have that level of visibility. And the key thing is that you don't need to predict 10 years out, that's impossible to do. No one can do that today. That's what the old type of automation systems might have been asking supply chains to do is make a bet on their 10 year plan. But now you don't have to do that anymore. You only need to plan out two or three years. And then you can start to look at an annual cycle where you say, okay, what is my operations requirement for next year and that capacity plan can be met on a regular basis rather than on a, you know, a forward looking 10 year plan.

Karl Siebrecht 20:59

There's, I would imagine, as I think over the life cycle or expected life of the robots, you know, if I have more certainty around my multi-year forecast, my spreadsheet that does the ROI calculation can come out with a better answer on here's my expected productivity improvements, because my forecast is reasonably certain. If it's less certain, that may create some more risk in the equation. Is that fair to think of? Although, as I say that out loud, I also think, well, you could add robots incrementally as you go, you can scale up, but you probably can't scale down unless you're leasing them.

Jon Schechter 21:39

Absolutely. This equipment is part of a network strategy. And so there might be buildings that are at the very edges of the nodes that are just too small that don't need automation, say, they're 7-Eleven sized. But most of the time, you can scale it down and still put it in a micro fulfillment center in the urban locations and in your huge central fulfillment center. So it kind of works in both places. And you'll just scale the amount of equipment you use.

Karl Siebrecht 22:01

As you work with bigger customers, bigger meaning businesses that have multiple warehouse locations in their network, is it typical that they would look at this broader network and say, okay, I want to automate in these locations, but then not automate in these other locations, so that I have kind of a network of different capabilities to handle my different lines of business or the different sort of dynamics in my business. Is that typical?

Jon Schechter 22:32

I think we've seen a mix of every possibility in our, you know, we've got 1,250 locations. In the early days I'd say it was more common for the end shipper to buy the automation and then hire someone to run it, one of the 3PLs or run it themselves. We're seeing more and more that the 3PL is now offering that in their contract. They're buying the equipment and they're using that as a way to get a longer contract, to get a better commitment from their end shipper. But just to make a better fulfillment center for them, frankly. Even going forward, we're seeing a lot more 3PLs especially thinking about multi tenancy, and so now are accumulating multiple businesses together. And they're then opening up to more of the automation potential across many businesses. And so they can swap out the boxes, but use all the same bins of inventory for different customers and fulfill that and offer that service and then use it as a competitive edge to win more business as a 3PL.

Karl Siebrecht 23:28

Right. Yeah, that really resonates. We have a lot of perspective on that question because a lot of the customers we work with, they have, sometimes we talk about them as master nodes, and then satellite nodes. Big facilities that tend to, more likely, to be automated, kind of in the center of the country, that will handle the entire catalog. But then they also want to forward deploy, call it the fastest moving SKUs or certain segments of their SKU mix of their catalog, out to the edge so to speak, put inventory closer to where the people live so you can be more responsive, saving freight. And when you look at that, it's kind of like, hey, the master nodes more likely you want to automate, the satellite nodes, maybe less so until they get to a point of more certainty in terms of volume and stability.

Jon Schechter 24:16

It turns out more than 60%, no matter what kind of product you're selling, fits in your hand. Even if you make parts for a tractor, there's tires and fenders, but for every one of those, there's two or three widgets that are used to attach that. So it's super important that you can fit lots of the items that you have into an AutoStore or a similar sized bin and still get a bulk of the volume out.

Karl Siebrecht 24:41

This has been super fun. I want to take us home here a little bit with my last question for you. What has been the most exciting new tech in automation and robotics that you've seen in the last few years?

Jon Schechter 24:56

The stuff enabled by computer vision is really cool. So whether that's some of the right size packaging equipment that can sort of see the contents and really cut the box down to the minimum size required. Or, of course, piece picking that's using computer vision to manipulate the items, or for some of the mobile robots now that are really interacting even closer to people than they were before, that's really enabling like a new direct connection and flexibility of technologies that they weren't able to do before that were a little bit more brute force, a little bit less flexible. So all of that is coming to bear where you have these inexpensive cameras from Intel and others, really good software packages that are pretty generalizable. Then a lot of companies can build products on top of those trends.

Karl Siebrecht 25:42

Yes, that's very cool. Well, this has been super energizing, Jon. Talking to you, it gets me all fired up. And it occurs to me that this is a great spot in the evolution of this industry, because it's super real, right? These systems have been operating for a decade plus, at scale, and they continue to evolve, and there's so much more that will come but there's enough certainty that it's gonna come. It's just kind of exactly what flavor and how quickly, right? And to be operating in this time, in this category, must be pretty thrilling. Jon, thank you so much for joining me here on the Logistics Leadership Podcast. I look forward to keeping in touch with you and look forward to our audience learning tons from you as they give this a listen.

Jon Schechter 26:31

Thank you, Karl, it's been a pleasure to be with you. I love this space and I look forward to talking more about it with you and others.

Karl Siebrecht 26:43

Okay, well, that was quite a conversation with Jon, lots to unpack there. Clearly, he's got tons of experience. Jordan, why don't we start with you? Could you help share some of what you've learned from a macro market perspective in terms of automation?

Jordan Lawrence 26:57

Yeah, Karl, a great conversation. When you mentioned you were going to talk about this topic, automation, I went and did a little research to see how big is this market and what's going on with this market today. What I found was really interesting. I was actually a little bit to the downside on the size of the market. Globally, warehouse automation and robotics is about a $16 billion market in 22. But it looks like 23 is going to be around 22 billion. So I think what's impressive is that it looks like 15, 16, 17% compound annual growth rate. Most analysts say they expect that to be maintained for the next five to 10 years. So tremendous growth in this space, which is really interesting. Just for some relative context on that number, say 22 billion for 23, the warehouse services in the United States alone, are estimated to be around a $170 billion market. So I think that's an important context to put it together with how big this is and I think it speaks to where the growth could actually go over the coming years. Ben, you had some similar experience with Kiva Systems, specifically. I was curious to get your take on the conversation and specifically with your experience in robotics.

Ben Dean 28:18

Yeah, Jordan, Jon had mentioned that he was at Amazon during their process of bringing in Kiva as their preferred ASRS robotic system. I was actually at a 3PL that was running Kiva at that same time. So different sides of the same coin. At that time, I mean, this is really bleeding edge automation within the warehouse. That whole experience of the warehouse worker, for a lot of folks in automation was this, there's going to be a boogeyman effect. People don't want to work with automation, they think they're getting pushed out of jobs. But in fact, it increased the value of working at our company versus somewhere where Jon mentioned, you had to walk 10 miles. Instead, the picks are coming to you, you get to play with interfaces that are tech savvy. And to his point, you know, people are used to dealing with screens all day now instead of paper pick sheets. So we found that it was a differentiator in the labor positioning. Karl led into this, talking about the onshoring and what impact that's going to have on labor. I think the only thing we know is that labor is not going to get cheaper in the United States. That's only going to have folks look at more and more new technologies as having a positive ROI. But I think that ROI piece is really the struggle today. What's that payout against your client contract length?

Karl Siebrecht 29:39

Given that robots are generally high CapEx, so the payback happens over an extended period of time, I'm just very curious on the dynamic of what is the return and what do the economics look like to the end customer versus what do they look like to in the case where a 3PL is making that investment and it sounds like more of an evolution to a services based or subscription based model. Some of that tension may be coming out of play. You had some experience with that as well, didn't you, Ben?

Ben Dean 29:39

Yeah. When Amazon bought Kiva and soft servicing the bots, we had to come up with a solution pretty quick because our entire distribution center was working around the Kiva floor. And Jon had talked about having very mature operations and sliding technology into that. I've seen much of the opposite in many cases, you build the technology and then you build the operation around that. That's not the way to do it. But when we had to replace Kiva at the time, Locus Robotics, they're a big name now, they partner with DHL. But it was a very new concept. So we piloted with them in probably their first year or so of having production level bots and they went to market immediately on robotics as a service, in this case RaaS model, that changed the way we looked at this entirely, because there's literally zero upfront cost to it. And you're only paying for the bots that were in your building the moment they came into your building. That allowed us to flex up, flex down seasonally. And, frankly, swarm pick areas when we had heavy volume for that. So I think he's right on about that point is that you've got to change those economics, because if you think about a traditional 3PL, they're getting their clients on one to three year terms, and the ROI on an upfront capital expense, forget bleeding edge automation, just conveyance and pick mezzanines is five to 10 years. So you're making a long term bet against short term revenues.

Karl Siebrecht 31:47

Yeah, I think that's exactly right. Look, you got a toolbox, got a handful of different tools in it. One of them is automation, generally speaking. Inside that category, you've got some elements, or some types of automation that are really fixed, and others that are a little bit more flexible. And I think those different tools, and by the way, there's another tool, which is very little automation, right? Those different tools map on to different elements of your business or different points in time across your operation, generally speaking, where you've got more certainty, more stability, in an operation, you know, the odds of getting a favorable payback are better. You know, I found it really interesting that he talked about how, I think, two thirds of the implementations that he sees are in existing facilities and one third are brand new greenfield facilities. And I think that goes in conjunction with his framework of, you know, start with people and he had Maslow's hierarchy of needs, start with people, and then get your process dialed in and get your process stable, and then look to apply automation to that. That all makes sense. And I think with more and more of these automation solutions becoming more like a SaaS type of service, that just allows the penetration and the adoption to speed up. I'm guessing here, but that's got to be one of the drivers behind that CAGR that you are seeing, Jordan, in that macro market data.

Jordan Lawrence 33:17

Yeah. I think kind of building on that, Karl, another trend that he kind of identified or at least spoke to that I found was really interesting is what is the minimum complexity threshold needed before you pursue automation and robotics? And the numbers he presented I thought were very low. And I think what that speaks to whether, you know, that number is right for an AutoStore product versus something else, another ASRS system, whatever, I think what it speaks to is the trend is robotics are becoming cost effective and useful at lower and lower complexity thresholds, meaning the addressable market is actually growing very rapidly. And that's feeding that annual growth rate. So Jon mentioned 1,000 SKUs, 1,000 orders a day, and 10 people dedicated to pick, pack and ship. Those are pretty small numbers in the grand scheme of things and probably addresses a very large percentage of the warehouse market today.

Karl Siebrecht 34:11

Yeah, I had the exact same reaction. Ben?

Ben Dean 34:14

Yeah, and I think of that tied to that modular pay as you go approach is the only way you can have success in a 10 person operation at automating anything is if you can bring it along piece by piece, as opposed to making a big bet around it. And the interesting thing here, changing gears just a little bit, is it sounds like we're talking about just fulfillment centers, not distribution, not bulk goods. But there's a lot of less sexy stuff happening in that space in terms of layer picking, palletization optimization. I don't think we'll have much time for that here. But I did want to give some love to those folks who are moving things around on pallets, that there's a lot going along with autonomous lift trucks, for example, a lot better than they were five to 10 years ago. And so that's contributing a lot to that CAGR is like, yes, fulfillments, the least efficient warehouse operation, you've got a lot you can take out there with automation and fulfillments growing, but retail distribution is still the biggest portion of what's happening in warehouses today. And there's a lot that can be automated there.

Jordan Lawrence 35:25

On the other end of the spectrum, the bleeding edge, I did find the computer vision comment, he immediately jumped to computer vision as being one of the most interesting new technologies. I think when you talk about the ease at which computer vision can be deployed and what that means for automation and each picking and very complex things that only humans could do at one point, you know, really unlocks what could be a lot of growth in this industry over the next 10 years.

Ben Dean 35:51

We're talking about so much within the four walls, how can I make the four walls more efficient? But if you're CSCO leading a supply chain, you know that that is one of the lower portions of the overall supply chain spend that you've got there. And transportation is driving most of it. And I'm not talking about autonomous trucks here. But I am stating that things in the warehouse can make transportation more efficient. So Jordan started to touch on the way the software that lies above the automation, making sure that that's optimized and running within the warehouse. But especially with DIM weighting and the parcel carriers really jacking up their costs, one of the last consultancy projects I worked on was auto boxing machines, so trying to cut as much DIM and weight out of your packaging as possible. I think there's gonna be a lot more bang for your buck as a supply chain leader there than inside the box of the warehouse.

Karl Siebrecht 36:44

Yeah, you're spot on, Ben. I mean, when you zoom out and look at supply chain distribution logistics costs in total, you know, what is it, four or five times, or maybe even more than that, is spent on transportation than is spent on warehousing. So whether it's automation inside the box, so to speak, to take air out of the shipping equation, or just the intelligence to help predict better where the inventory should be so that the last mile is as short and efficient as possible and the mid mile is as efficient as possible. Huge bang for buck there, too.

Ben Dean 37:25

Let me give one story to illustrate that. So I mentioned my Kiva operation a couple of times here. We built that at the start of 2010, in Louisville, Kentucky, because that's where UPS was and that's where the warehouses were. We were told you get everywhere in the U.S. within two days, right? That's a dinosaur concept. That facility was highly automated with a huge amount of investment into it. Today in running my supply chain would rather have a dozen non automated facilities, so I can get one day service to the entire United States. And I can guarantee that's cheaper from a parcel transportation perspective and overall cost perspective than running one very sophisticated automated DC.

Karl Siebrecht 38:09

That's exactly right. And that's what we see time and again. As you think about the potential for automation in that context, now as the upfront fixed costs come down, the applicability of robots to less and less complex operations in terms of you can still get a payback, those two things start to merge together. It's just a far better model than the sort of big automated behemoth in the center of the country of you know, circa five, 10 years ago.

Jordan Lawrence 38:38

Well, guys, this has been a great conversation. I really learned a lot. The automation space, robotics, it's really the sexy, exciting thing, people love to talk about it. There were definitely some very interesting tidbits, I think the space is moving forward very quickly. The growth rate, of course, really surprised me. Any final thoughts on your end, Karl?

Karl Siebrecht 38:58

My biggest takeaway from this was, you've just got to start with the process. Once you have a stable operation, now look to apply automation to it. And it's tempting to do it the other way. You know, I think we've all seen it done the other way. Ben, you mentioned this where, you know, robots are sexy, they're cool, let's go get some robots. And, you know, to take that important step of just understanding the operation deeply and then figuring out how to apply automation. That's the real key and look, now, there are several different options for what types of automation to apply. And those continue to evolve pretty rapidly, as well. So no question it's a real space. It's very real, you know, real market penetration and the innovation trajectory that many of these companies are on and the adoption trajectory is pretty exciting. So lots more to learn as this space continues to evolve, but certainly enjoyed having this discussion with you guys, and learned a lot from our guest, Jon. Look forward to continuing this conversation and we'll see you on the next episode.

Jordan Lawrence 39:13

See you next time.

Narrator 40:14

You've been listening to the Logistics Leadership Podcast presented by Flexe. If you'd like to learn more about the podcast or join the Logistics Leadership community, check out this episode's show notes and visit flexe.com/logistics-leadership-podcast. Keep the conversation going: Email us at leadershippodcast@flexe.com. The Logistics Leadership Podcast features original music by Dyaphonic. The show's producers are Robert Haskitt and Adam Kapel. Here's a quick pro tip: Instead of chasing down the next episode, why not just follow the show and have it appear in your feed automatically. Thanks for joining us!


  • Karl Siebrecht 2022 Headshot 2

    Karl Siebrecht

    Co-founder & CEO

  • Jordan lawrence flexe

    Jordan Lawrence

    Director of Logistics Strategy

  • Ben Dean

    Ben Dean

    Director of Network Strategy & Optimization


  • Jon Schechter Headshot 2022 1

    Jon Schechter

    VP of Fulfillment Solutions, AutoStore