In this episode of Hardware to Save a Planet, Dylan is joined by Ben Chehebar, VP of Hardware at RoadRunner Recycling, and Jason Gates, Senior VP of Strategy at RoadRunner Recycling, for a free-flowing discussion on Ben and Jason’s solution of making waste disposal more sustainable and efficient. The concept was ideated at their start-up Compology, where they installed sensors in dumpsters to measure how full the dumpster was so that the disposal agencies would have to make fewer trips. This was a positive step towards making waste disposal more efficient and reducing the carbon footprint of the trucks clearing the dumpsters. RoadRunner Recycling has recently acquired Compology.
Dylan met Ben and Jason at Synapse, where they were developing the first iteration of their camera for the dumpsters. Ben has a Bachelor’s degree in Mechanical Engineering from the University of Michigan, while Jason has a Bachelor’s degree in Civil and Environment Engineering from the University of Maryland.
To discover how smart dumpsters are lowering waste disposal costs while adding to sustainability, check the key takeaways of this episode or the transcript below.
- 06:51 – 08:44 – The seed of an idea – Ben and Jason explain how the hero of the idea was the fullness sensor that could track how full the dumpster was. The idea was for disposable companies to make fewer trips, thereby cutting the disposal cost. It would also save their clients thirty to forty percent of their waste disposal costs. The initial idea was to scale the technology to benefit a neighborhood, but there were bigger and better things to come. Listen to the podcast for the full story.
- 10:37 – 12:04 – Pushback from waste companies – The average dumpster is cleared when it’s only half full because waste collection is based on estimates. The more trips the waste companies make, the more money they make. Understanding how full the dumpster is would make waste collection more efficient and cut costs—however, this opposes the business model of waste collection companies.
- 13:28 – 15:33 – Arrival of the sustainability tailwinds – Jason explains how around 2018, there was a macro-level shift in how the waste generators and government thought about efficiency and sustainability. People started focusing on sustainability and wanted more than just cost savings, which was an inflection point for the business.
- 15:54 – 17:13 – Adding a camera is a game-changer – Ben mentions how, very early on, they decided to add a camera because they found that cameras did the best job. Cameras plus machine learning did the best job of accounting for all the variability. It also allowed them to provide feedback to the waste generators on what should be in their trash and what shouldn’t. They knew the idea had taken root when the clients (waste generators) started contacting them to check how they were doing with their disposal.
- 26:12 – 28:42 –ESG compliance while saving costs, a double win – Ben mentions how the common refrain at most companies is, why are we investing in ESG if it doesn’t help us make more money? The beauty of the waste metering solution is that it allows companies to make more money while moving the compliance needle on ESG.
Dylan: Hardware to Save A Planet explores the technical innovations that are giving us hope in the fight against climate change. Each episode focuses on a specific climate challenge and explores an emerging physical technology solution with the person bringing it into reality. I’m your host, Dylan Garrett.
Hello, and welcome to Hardware to save a planet. I’m excited to have Jason Gates and Ben Chehebar here today. They are the founders of Compology, a waste metering company that was acquired by Roadrunner Recycling last fall. We’ve talked before on the show about the importance of waste management and recycling and its ability to divert materials from landfills and reduce our dependence on producing new virgin materials. The stat that always stands out to me is that 75% of our waste in the US. Is recyclable, but only 30% of it actually gets recycled. Compology has a really important solution for this, which leverages smart connected cameras to monitor the contents of dumpsters. The data they collect helps to improve the efficiency of waste hauling operations and also the waste production and sorting behavior of their customers. I first met Ben and Jason back in 2014 when Synapse was lucky enough to help them develop an early iteration of their smart connected cameras. At the time, I think Ben and Jason were only a couple of years into their first foray into entrepreneurship, but I was really impressed by their maturity and dedication and it has been awesome to see them grow their team and the business through to the recent acquisition. To this day, they’re among my favorite clients I’ve ever worked with at Synapse, so I’m really excited to hear more of their story. Ben and Jason, thanks for being on the show.
Ben: Thanks for having us.
Jason: Thanks for having us.
Dylan: Yeah, I’m excited to do this. So like I just said, I picked up your story after you’d already started the company, but I’d love to know what happened before that. So maybe just to pick one of you. Ben, when did you two first meet and how did you end up starting a company together?
Ben: Yeah, that’s a great place to start. I think we have a fun founding story. So Jason and I met in middle school, so we were both in Jason, I think it was Mr. Rivera’s, like advanced math class or something, where
Jason: Sounds right.
Ben: Dylan’s kind of funny. We would actually go to the high school for our math classes and we always got there too early and we’d have to sit outside the class and then we get heckled by the high schoolers because we were like the young 8th graders or something hanging out in the hallway. But I think Jason and I met in that class and bonded pretty quickly and have really been best friends ever since. And through high school, we always collaborated on random stuff, random projects, parties were one of them. And I think that kind of seamlessly transitioned into wanting to start a business together.
Dylan: I saw that you went to different colleges. Right. I think, Ben, you were at Michigan, and Jason, you were in Maryland, so you must have kind of kept the friendship going through college, even though you weren’t going to school together. Then. Did you end up living in the same area after college again?
Ben: Yeah. So Jason went to the University of Maryland. I went to the University of Michigan. So both Big Ten schools now, I guess they weren’t when we went for both UMs and after college. Yeah, we both were. Jason, you moved to Philly for your first job, and then I moved to Portland, Oregon, for my first job. So we were. I don’t want to use this terminology. We were long distance for about a year, but yeah, we kept in touch, obviously. And I think even when we had our first jobs, we were continuing to collaborate remotely to think about what we’d want to do together as a business.
Dylan: Okay. Oh, interesting. So you knew you wanted to start a business together.
Dylan: And you were looking for an idea. Jason, did you have something?
Jason: Yeah, I think it started when we were undergrads, and there were a couple of different projects that we were collaborating on where we thought this could turn into a business and that we found to be really fun, the process of ideation and then fleshing out what the business model could be, and that continued. And I think we were both not really happy with where we were in our first jobs. They were incomes after you graduated, but then that was really a catalyst for us saying, let’s leave our full time jobs and start something. But we didn’t actually have the original idea for Compology until after we had moved on from our full time roles.
Dylan: Oh, nice. Okay. You made the leap without knowing exactly what you were going to do.
Jason: We had a high level concept.
Dylan: Okay, before we get into Compology, I’m curious what other business ideas were kind of on the cutting room floor? What else did you guys explore back in the day?
Ben: Well, Jason, so one of the first businesses we applied to, Y Combinator Techstars, is to remember that? It was like Dylan we were going to find a way to assess the mood in a bar or restaurant using audio and video. That’s kind of like the missing piece when you go to a restaurant. Like, is it cozy? Is it the right music? Like, things like that. We were attempting to bottle that into a software product where people could find all that out before they actually visited. So that was one of the early ideas. Well, yeah, cutting room floor, for sure.
Jason: The one I always remember is the triangular chapstick so it doesn’t roll off the table, man.
Dylan: I’m sold already. Yeah. Oh, nice. Brilliant. Okay, good. Wow. So among all that Compology sifted to the top, somehow what was kind of the insight that led to compology. How did that start out?
Jason: I think we were always really interested in the environment and sustainability and there were some courses that I was taking at Maryland where we were learning about anaerobic digestion and how you could actually use food waste to generate electricity. And there were some projects that Ben and I had started to read about that were very much the benchtop scale but businesses that were trying to make at home anaerobic digesters. So taking technology that was being used at utility level scale and package it so that it can be put right next to your air conditioner at home, allow you to process your food waste at home, generate electricity. And in theory it sounded really neat. And so we kind of set out with the original goal of pursuing something around applying anaerobic digestion for the single family residential market.
Dylan: Wow, talk about a pivot. You always hear how startups go through many kinds of iterations. That’s interesting. Yeah. So you ended up how that evolved into what complexity is doing now?
Jason: So that evolved into scaling that technology up to a level where it didn’t make sense to do at a single family property, but maybe to be able to serve a community or a neighborhood within a city and being able to offer composting services without having all of the transportation in and out of urban areas. So think like a shipping container sized anaerobic digester, which we then learned that there’s a reason why that technology works on a utility scale and not on a residential scale. And there’s cost considerations, but also just the science behind it that turned into a smart dumpster for food waste specifically, where instead of having to collect the material frequently, we would add some type of fermentation solution to the food. It was kind of like spraying over the food and effectively pickling it so that you could leave it on site for a longer period of time without having to collect it. And as part of that system, there was a sensor to tell us how full that smart dumpster was. And that smart dumpster for food waste is actually what we ended up raising our first outside investment to build. But it was only after we raised venture capital and really got to work on scaling up those really early prototypes into something that was a little bit more ready to go to market. We learned that more of the value was in the fullness sensor itself. And through going out and talking to customers or folks in the industry, they said, hey, well, the smart dumpster is interesting, but could you take that sensor out and put it in the dumpster that I already have and tell me how full the dumpster I already have is? That was kind of like the AHA moment.
Dylan: Got you. Did you have prototypes of your sensor at that point?
Ben: We did actually, we got a funny story. After we won our first business competition for this smart dumpster for food waste, we were driving back from Oberlin, Ohio, and we got pulled over by a police car. And I think it was because they saw the dumpster in the back of our car and they said, what is that? And they made us explain what it was. We were hot off our pitch competition. Gave him the whole pitch. You’re like.
Dylan: Hey, you can fund this thing.
Ben: We did not get a ticket. I think he was just annoyed by us, and we went on our way. But we certainly made at least one prototype of it, just get a feel for what it would look like and all that.
Dylan: Nice. I think you did a mechanical engineering degree at Michigan, right?
Ben: I did.
Dylan: Right. Ben and Jason, you also were in environmental engineering.
Jason: Civil and environmental.
Dylan: Civil and environmental now. Kind of. Jason, you’ve been more focused on the business side. Ben, you’re focused on the product and the technology. Was that how it worked from the very beginning? Was that always kind of a clear breakdown?
Ben: Not really. Yeah. I think at the beginning, we as founders, you end up doing everything together, which is great, and you learn how to work together, but obviously very inefficient. And at some point we had to have a conversation about how we have to divide and conquer here. We still can’t do everything together. What are the things that you’d like to peel off and focus more on? And Jason, I think it ended up kind of being a natural conversation where it made sense for me to handle lean more on the product side. Jason, more on the financing and sales side.
Dylan: Cool. I’ve always been impressed at how well you both collaborate and kind of lead together. So that’s cool to hear that.
Jason: Yeah. It’s not always easy, though. Yeah, I’m glad it comes off that way, though.
Dylan: Hey, you’re still here. You’re still doing it together. That’s right. Okay, so people were interested in the sensor itself, and what was the problem at the time they were looking to solve with that?
Jason: Well, early on, we learned that the average dumpster in the United States is collected when it’s 51% full, and that is because the collection schedules are largely set based on estimates. And we’re talking about dumpsters that you might see behind a restaurant or an office park and typically between two and eight cubic yards. Right. And the customers that we were talking to were the folks who are actually generating the waste. So for them, understanding how full the dumpster is would allow us to provide better collection schedules that are more efficient and reduce the amount of truck trips to their location and ultimately reduce the cost of waste collection. One of the really interesting learnings is on the flip side of that is you have the waste companies themselves and the reason why a waste company might sell you services so that your container is 51% full on average, is because it’s very, very profitable for them. The more air that’s in the container when they pick it up, the more profitable it is. So you have this misalignment of incentives in the industry, and Ben and I thought that by using technology, we might be able to realign those incentives and provide value to the end waste generator so that they could have fewer truck trips and less spend. We thought the hypothesis was that we would also be able to provide value to the waste companies themselves. To have more efficient routing and less truck trips.
Jason: What we ultimately ended up learning yeah.
Dylan: It sounds like maybe that hypothesis wasn’t proven to be true.
Dylan: Okay, tell me about that.
Ben: I’d say the first, almost like, year of the business, we were pounding the pavement, talking to as many people as possible. We really thought haulers, the companies that pick up the trash, were going to be our core market. And we soon learned that they wouldn’t be. We actually deployed, Jason and I think we deployed some cameras and some pilots with haulers to demonstrate at least that the savings were there. But as Jason was mentioning, just the business models didn’t work for them. They realized that if this technology did scale, it would drastically reduce their revenue or just change their operations. A point they didn’t want to do.
Dylan: And at the end of the day, that’s because the dumpsters would be more full and there’d be fewer pickups for them. And that’s how they make money. Interesting
Ben: For a hauler, the most profitable pickup is an empty dumpster. Because they go pick it up and they have no cost of disposal. Right. And so, yeah, we’re making things more efficient for the system, but less profitable for them.
Dylan: Wow. Just as an aside, have you seen any movement in changing those incentives to be better aligned? Because that just seems like a fundamental issue.
Jason: Short answer is no. But I think in 2017, 2018, there really started to be a macro level shift in the way that waste generators and governments thought about efficiency and sustainability. And that was really an inflection point in our business. So Ben mentioned that we sold some waste haulers. They were primarily specialty recyclers. So they had some other angle where there was a benefit to them to use the technology. But really where our business started to take off was when we had those tailwinds of people starting to really focus on sustainability. And wanting more than just cost savings. They wanted to quantify, how much waste am I actually producing at these buildings? And I want accurate measurements at my distributed locations that might be all across the country. I want consistent, reliable, accurate data to report in a public disclosure. And those types of changes were what really helped fuel Compology’s growth.
Dylan: Okay, so actually, you said when you first started looking at company ideas, you were interested in sustainability. And then it almost sounds like kind of as you got customer feedback and stuff, you went into this down this route that wasn’t necessarily driven by sustainability. It was more about kind of operational efficiency. But then as the trends started getting more and more aligned with sustainability, it became more of a sustainability play again.
Ben: Yeah, it was always in the back of our heads when we were building the company that we knew there was a sustainability angle to this, but it just didn’t land. Early on, we would try to bring it in as, like, an additional value and kind of be told to shut up, focus on cost savings. And so I think we were excited as the market shifted maybe a couple of years ago, where sustainability became a very key component of what we were building and what people cared about.
Dylan: And we’ve talked about how it’s about measuring the fill level, but as I understand, you’re also measuring contents and contamination rates. And can you talk a little about that? And was that always the plan, or did you start to add in more sophistication about understanding what exactly is in the dumpster in the first place, as you understood customer needs better and that kind of thing?
Ben: Yeah, I would say it was sort of in the plan and sort of not. So, I think one of the critical decisions we made on the technology side very early on was deciding to use a camera in the first place. And the reason we actually did that is we experimented with a whole number of different sensors. There’s ultrasound, there’s some time of flight sensors you can use. And we just found that cameras did the best job. Cameras plus machine learning did the best job of accounting for all the variability you’ll see in a dumpster, dumpster sizes, how the trash recycling is actually loaded into the dumpster, all these different variables. And I think at the time, we said, all right, great. We have cameras. You probably see other things going on. And that was the high level thought. But it wasn’t, again, until we kept getting more customer feedback, more around, how am I doing with my recycling, my haulers actually finding me for having the wrong material in the wrong dumpster. That’s when we said, all right, well, now we have these cameras, we have all these images. Let’s build a feature around this where we’re actually identifying material that should be recycled that’s maybe in the trash, but also trash that’s in the recycling that should be taken out.
Dylan: Okay, so you’re doing both. You’re in trash dumpsters identifying. You’ve got a bunch of cardboard boxes in your trash, and you’re in recycling dumpsters saying you have a bunch of food waste or something, and you’re recycling.
Ben: Yeah, exactly.
Dylan: I want to understand the customers a little bit more. So Jason, who is the typical customer. It sounds like the waste producers. Can you describe a little bit more about who they are?
Jason: Yeah, our primary customers are Fortune 500 brands with distributed real estate portfolios. And the rule of thumb is the more brick and mortar locations, like unique addresses that a business has, the more valuable our solution is. So some of our marquee customers are concentrated in the retail, quick service restaurant, multifamily housing, and office space. Those four property types are primary, although we have customers in 17 different property types. The value is recognized regardless of what business you actually run. But it’s brands like McDonald’s, right. And Starbucks and Google, and it’s brands that consumers see every day who have a real drive to go above and beyond when it comes to their own sustainability goals.
Dylan: And just quickly, it seems I’ve read I think I was reading about a pilot you did with Miami that showed a 40% reduction in costs and 50% increase in recycling rates. So there’s clearly an environmental impact, but these people must be looking at the financial benefit as well, right? It sounds like that’s pretty significant.
Jason: Yeah. So on average, our customers will reduce the amount they spend by between 30 and 40% after implementing Compology’s waste metering solution, which for a McDonald’s can end up being several thousand dollars a year in savings per location. And that’s where it becomes. The more locations that you meter, the benefit really adds up.
Dylan: And what are the kinds of changes they make just to help us understand the impacts of this? What are the kind of changes these companies are making once they have access to this data and visibility to it?
Ben: So the primary thing that we do for a generator is we recommend new collection schedules so their trash dumpster maybe before compology is picked up five days a week with our data, we may recommend that it gets picked up two or three times a week and so they get savings for that reduction in pickup. That’s probably the biggest thing. On the contamination side, we give them a breakdown of the types of contaminants inside their container and we provide real time alerts for them to actually remove that material before it gets picked up. That just helps get it out of the stream.
Ben: And also I mentioned some haulers actually issue fines when your trash recycling is contaminated, and that avoids those fines as well.
Dylan: Oh, wow. Okay. So like, a McDonald’s franchise manager would get an alert saying, hey, you’ve got a bunch of food waste in your recycling dumpster. You might want to go deal with that before the pickup comes.
Ben: Yeah. Typically, like, whoever’s on site who helps manage the waste, we would alert. And it’s funny actually, most of the alerts that we see are the more obvious things. Like someone threw a sofa in a McDonald’s dumpster and that definitely needs to get removed, right. Or a tire or a pallet, things like that. It was just very obvious.
Dylan: Yeah, I just heard somebody making the point that cardboard boxes, if they’re not broken down, just the amount of volume they take up in a recycling dumpster is really detrimental to kind of the whole ability to be a dumpster. Is that the kind of thing you can look out for with your system?
Ben: Yeah, totally. So we actually have a contamination category that’s just called uncollapsed cardboard boxes.
Jason: Yeah, we call that artificial fullness. So the statistic that we share is that about 15% of an average container is artificial fullness from just not breaking down the cardboard boxes.
Dylan: Jason, what do you know now about running a business like this that you wish you had known when you started this company?
Jason: Well, we touched on it a little bit earlier in the conversation around alignment of incentives and I think Ben and I were really eager to have identified a clear mismatch in incentives in the industry and we thought that technology was going to be what was going to help get people aligned. But I think what we overlooked is that there’s a reason why waste companies set up their business model the way they did and there’s a reason why they might be resistant to changing that. And that resistance is often much harder to change than you might think. As young guys starting our first business, we would read all the tech crunch articles and we would hear the rhetoric around disruption and world changing and of course we believed that we could do the same for the waste industry. So now when we look at problems and even within Compology as we started to really have these learnings is really thinking through who are the different stakeholders that are involved. And that can go well beyond the obvious, that can go to the regulators who are overseeing an industry, that can extend to the unions who are representing drivers who may be asked to use a specific type of technology. And once you understand those different stakeholders what their incentives are and what they might be not incentivized to do at all, then you can really build products that people want to use. And when you have an entire industry who’s collectively incentivized to see a product succeed, that’s when you can achieve really breakout growth. And so I think that’s probably the biggest thing that I’ve learned.
Dylan: Yeah, I love that because it is counter to a lot of the, I guess talk about starting a company and coming in and just disrupting everybody or disrupting an industry or an incumbent. So how do the waste haulers view your solution now?
Jason: Typically not in favor. Yeah, so the way that our business has grown is by selling directly to the waste generators. Our contract is with the McDonald’s of the world and they’re the ones who put the pressure on their waste hauling company to say, if you want the service contract for my locations, you will allow me to install this metering technology or you’ll allow Compology to install it on my behalf. And it’s only through that pressure that we’ve been able to overcome the hurdle of waste companies resisting the installation of the metering.
Dylan: And just thinking about incentives, you talked about how companies are more and more incentivized by environmental impact. I guess I don’t fully understand why that is. Is that driven by regulation, by kind of consumer demand? What do you think has caused that shift?
Ben: Well, don’t actually think it is those two things you just mentioned. It probably stems most from consumer sentiment, especially with the customers we work with who are brands, right, who care a lot about their customers, who are consumers. So when there’s a shift in what they expect out of a brand, someone who cares about the environment, who’s thoughtful about the way they waste or recycle and things like that, it’s going to cause a shift in those companies to make sure they’re addressing their customers properly. And I think relatedly regulators, if they don’t see the movement they’re expecting out of companies as part of a broader goal to net zero or whatever it is, will put in regulations to do that. So I think those are probably the two biggest drivers of the sustainability sentiment that we’re seeing.
Jason: I’d say maybe the third. I totally agree with Ben that shareholders are starting to take a much more active position. So for publicly traded companies, the idea that they’re not doing anything related to sustainability is not acceptable anymore to most public markets investors.
Dylan: So is that related to ESG reporting and the company?
Jason: Yeah, exactly. So the hot topic right now is being able to quantify carbon footprint and waste fits in as a large contributor to carbon emissions.
Dylan: So ESG is environmental social governance. When a company is reporting on their ESG activities or their ESG impact as a company, part of that reporting includes what they are doing with their waste?
Jason: Right? So it’s how much electricity they use, how much water they use, how much waste are they producing. And then other things might be sustainable farming techniques they’re employing or how they’re also working with their employees for a more equitable or diverse workforce. So it kind of captures all of those things.
Dylan: So you’re kind of supporting that whole process with visibility into the waste. And once a company is able to measure their waste and figure out where they can then figure out where to improve on it. Makes sense.
Jason: Yeah. And historically, there’s been great technology development in metering for water, electricity, and natural gas consumption. Those have been around for decades. So you have a great data quality and consistent data amongst all your locations for those utilities. But for waste, it’s always been estimated, and the method for aggregating that data is to go to each location and say, how much waste did you produce? Or ask a waste hauler, how much waste did that location produce? And when you have a fragmented network, you get a lot of different answers and different methodologies and makes it very difficult to report on waste the same way that you might for your other utilities. So with Compology’s waste metering solution, we’re able to give operators the confidence that the data is accurate, that it’s consistent across all the locations where they have it installed, and that they can report on it and talk about using the data to inform their business decisions the same way they would data from water meter or a gas meter.
Ben: Dylan, I think in the conversation about ESG, I think there’s like a conflict right now where shareholders or stakeholders more broadly are saying, well, why are you investing in ESG if it’s not actually adding to profit? Right. So why are you investing in carbon-capture projects or reforestation when it’s not making any more money? And what we’ve always leaned on, it’s really where the company started was, yeah, we have all these ESG benefits and we help you track what you’re wasting so you can reduce it. But there is also a fundamental cost reduction by implants and technology as well. And we found that really having both of those things makes it a very sensible technology to implement. Right. It does both things. Right.
Dylan: It’s not a green premium. They’re not paying more to meet their ESG goals. They’re actually saving money and meeting their ESG goals.
Ben: Exactly. Jason, what’s the tagline for that? Doing good
Jason: Doing good Oh, doing well.
Ben: By doing good.
Jason: While doing good. Doing good while doing well.
Dylan: Doing good while doing well. Yeah.
Jason: The ESP solution, that makes financial sense.
Dylan: So getting that quality data is obviously a key part of all this. And then you talked about how you kind of made the decision to go with a camera over other sensors. I want to dig into the hardware a little bit more. Can you maybe just start by describing what it is? What does it look like, how does it work?
Ben: Yeah, fundamentally, it is a wireless camera. So there’s a camera in there. We’ve got a flash to illuminate the dumpster when the lid is closed or it’s dark out. We have an accelerometer inside as well, which can detect when a container is being dumped. So we know we can confirm service, but also know when to take the next picture that shows that it’s empty. We’ve got a big battery. And so our cameras last about five years in the field when they take three images a day, and it’s all in a very rugged enclosure, which, Dylan, I know we worked with your team quite a bit on that.
Ben: In the early days, I think the shotgun test, Darren shot a shotgun at it. I’m not sure what data that gave us.
Dylan: And I remember a video of a dump truck driving over one. Yeah. And it’s surviving. What is that, like a 20 ton truck or something driving over one of these things?
Dylan: That was pretty impressive. Yeah.
Ben: That was probably one of the areas that we spent a lot of effort on, and we continue to spend a lot of effort on, is just making sure that our cameras can actually survive in these gross, super rugged environments.
Ben: Dumpsters are tough places to be.
Jason: That was one of the biggest areas that we got pushback on early on in the business from prospective customers and investors, was, can you even create something that’s durable enough to live in a dumpster? So we put a lot of effort into that early on, and that’s been mentioned, continued to evolve, and we started to incorporate things like nanocodings on the camera lens to resist schmutz that gets catch up, whatever is in the dumpster, and so continuing to iterate on improving that.
Dylan: Interesting. Yeah. What are some of the things you have to worry about in a dumpster? I’m trying to remember the requirements we had back in the day, but I remember, like, they fill with water. Right. So you have to be able to be fully submerged.
Ben: Yeah. If you’re in a compost dumpster, it may be submerged in, like, a liquidy type material. There’s obviously objects that may hit it. So, like, if you’re just throwing your trash in there, a sofa you’re throwing in there needs to withstand that type of impact. All different types of weather. So we’ve got cameras in uptilling negative 40 degrees in Northern Canada that we need to survive. And then, really hot down in Mexico, we’ve got cameras that need to survive. Yeah. And then I think one of the biggest impacts you see with a dumpster is every time they lift and then smash the container back down on the ground, that’s a huge impact to the camera. And so that if you do it three times a week for five to ten years, it ends up adding up to five years. Exactly
Jason: I think the biggest thing that surprised me was the amount of tampering that would go up, like how you needed to protect from somebody on the street just helping themselves to the camera.
Dylan: Right. It looks valuable. Somebody might try to take it. What do you do about that?
Ben: We do as much as we can. The counter to that is you’ll need to make sure it’s pretty easy to install and can be swapped out if it fails. So we use security screws and things like that. And at the end of the day, there’s really nothing someone can use the device for if they do steal it. So I think that’s the biggest thing. They can’t access any of the data. You can’t use it as a cell phone. There’s nothing there that’s usable.
Jason: Yeah, I think one of the funniest memories I have from the early days is I think it was one of the prototypes that Synapse helped us build, but it was in San Francisco and it got stolen. And I think we had just paid 500, $600 to produce this prototype, but we could track the GPS location of it. And Ben and I tracked it down to where it was. And then we walked over there and we were trying to confront this guy who had stolen the camera. And when we reported back to some of our investors, we thought we’d done a good thing and were good stewards of their capital. And they were just like, you guys are idiots. What are you doing? Are you going to get yourself killed? Like going down an alley. For $600? Just buy another one. Okay, yeah, that makes sense. We didn’t get it back. He refused.
Dylan: How do you maintain connectivity? Is it cellular?
Ben: Yeah, I should mention that. So it’s like a flavor of the LTE network called Cat M, which is designed specifically for low power distributed devices. And it’s funny, when we started, Cat M was really just starting out itself. It was a little bit like the new technology on the block. And I think we were maybe one of the first or second customers of Rise. And I was using Cat M in an actual product in the field, and a whole lot of learnings about using bleeding edge technology like that when it’s not really available, because they were Lightning up new cities with this technology, basically, as we were deploying in these new cities. And so sometimes we would get there too early and the network wasn’t set up yet. So that’s a big challenge. But now we still use that same technology, but it’s been around for five, six years now and very widely deployed and really reliable.
Dylan: That’s a good call out. One thing that always stands out to me and tell me if I’m remembering this right, but for your early customers, you decided, rather than waiting on having a machine learning model built up to be able to do AI image recognition and stuff in dumpsters, you just used a Mechanical Turk model. You just had people manually looking at images to identify contamination and to make estimates of fullness level and stuff like that. I always thought of that as a really smart decision because I think it would be really easy to say, hey, we’re an AI company. We’re a machine learning company. We’re going to go develop those algorithms and make that part of it work before actually going out there and getting data with customers and showing value to customers instead? At the scale you were at, it probably made sense to have people manually doing this. And that stood out to me as a really smart product and technology decision early on.
Ben: And the Mechanical Turks were us for a while, too. That was a good learning experience. But I think the way we thought about it was for any machine learning model, you need training data, right. And to develop train data, you actually have to have labeled sets of images, in our case, and you’re going to have to have people doing it anyway. So if you could overlap those two functions where you’re actually providing live data to your customer by labeling these images and eventually be able to use that as training data, you kind of kill two birds with 1 stone. And so, yeah, I think for the first, probably couple of years of the business, first, it was Jason and I. Then we had an internal team in the US that was actually doing it live. Then we eventually outsourced it to India. Now we have a very, very accurate machine learning model for fullness and contamination, but we still have teams in India that just do sort of like the backup system in case the automated system doesn’t know what to do with an image, for example, which happens. So we’ve built this whole system to deliver that data.
Dylan: How many pictures of trash do you think you’ve looked at in the last ten years?
Ben: I don’t know personally, but as a company, we’re at about 100 million images.
Ben: That we processed yeah. Over ten years.
Dylan: That’s amazing.
Ben: We actually applied for the Guinness Book of World Records for the most number of dumpster images in the world. And we got a very nice letter back from them signed by whatever the director that says. I think this is a little too niche for the goods of the world records. You got this lady squirting milk out of her eye. Isn’t that niche?
Dylan: Yeah, that’s kind of an admission that you’re also the world record holder, though. Totally sounds like, to me yeah, one of one, yeah. That’s amazing. Is that data set and now you have all this trained data? Is that in and of itself, something that’s valuable and can be used in other ways?
Jason: Well, I think the way that we’ve approached it is, how can we take the same architecture and the models that we’ve developed through dumpsters and apply it to other business use cases? And so that prompted us to launch a line of business in over the road trucking where we use it’s technically a different camera, but it’s all of the same components that Ben described in a different form factor that is installed in 53 foot drive in transportation trailers. And we’ve used our machine learning models to do exactly the same types of analysis. So how full is the trailer? And then over time, we’ve now expanded it to how much floor space is available versus airspace? Or is the loading all the way at the nose of the trailer? All the way at the tail of the trailer? But very specific attributes that a shipping company might want to know about their assets. So the data from dumpsters has been very valuable in being kind of the seed for us to be able to launch into some of these tangential lines of business and do it much faster and more accurately.
Dylan: I was going to ask you if the data about a company’s trash or waste and recycling has any, do they use it in any way to kind of feed back into their own purchasing and consumption habits? Oh, I see. We’re using a lot more of this product than I thought. We’re buying a lot of things that have non recyclable packaging or something like that. Does it feed that far upstream into a company’s operations ever?
Jason: Not in an automated way. And I think that’s a great future vision for where the data could be used. We see it happening organically with just the use from customers where we provide them access to view the images. So along with our automated classifications and service schedule recommendations, you can see the pictures from the inside of the dumpsters. And we have restaurant operators, for example, saying, jeez, I didn’t realize that so many trays were getting thrown out. I really need to talk to my team about why they’re throwing those. We should just be washing them and reusing them, not throwing them away.
Jason: So things like that.
Dylan: Or stop throwing our sofas into the dumpster.
Ben: Good start.
Dylan: So, Ben, similar question for you. As I asked Jason, what do you know now about building a hardware product or a technology product of this kind that you wish you had known when you started?
Ben: I think one of the biggest learnings for me is that everything is going to take longer than you think. Technology development is very difficult. Prototyping is very time consuming. Testing and validating your product is very resource and time intensive. And yeah, your perfect schedule that you built, that Gantt Chart you’re probably not going to stick to because you’re going to have a whole bunch of things that pop up in that development process they need to deal with and get past. That’s definitely a big thing. And I think similar probably on the sales side as well, because you’re always trying to match product development with customer needs and sales take longer than you think, and customers committing to paying for a new feature that you just built is going to take longer to commit. So just keep that in mind when you look at your bank account and how you’re building your team and all of that, that you need to build a little bit of a safety factor into everything because everything’s taking longer than you think.
Dylan: Yeah. Thinking about the future a little bit, how do you see the product or technology evolving over the next five to ten years? Are the things kind of on the Roadmap or you hope to put on the roadmap at some point that you’re excited about?
Ben: Yeah, I think what’s been very fun over our entire journey is just being a part of so many new developments with technology and being able to incorporate them into our product. New battery technology we’ve been able to use which can extend the life of our camera. From when we started, it was maybe six months and now it’s five plus years. Cameras are getting cheaper and cheaper every day, especially because they’re being used in so many different vision applications now cellular modems are getting lower power and more reliable. So I think just continuing to refine our product to where it’s more and more reliable, it can last longer than ever in the field and also just, I think almost to become more affordable too, which helps us scale but also helps customers be able to adopt it more easily.
Dylan: And Jason, we haven’t talked about this yet specifically, but you were acquired by Roadrunner last fall. What’s the future of Compology now that you’re inside Roadrunner? What does that look like?
Jason: So Roadrunner is an outsourced waste management service where we’re providing collection on behalf of our customers. So from a customer base perspective, our customers overlap quite well. It’s the same kind of target customer base. And the idea is that all of the benefits that we’ve been selling directly to brands for cost savings in ESG that we would be able to incorporate directly into Roadrunner’s suite of services. And the primary benefit is that it will help drive profitability of Roadrunner’s core business. So instead of the end customer getting all 30% to 40% of those cost savings, roadrunner would be able to realize a lot of that. But on the sustainability and data side, being able to package that as a newer product to offer and combine it with some of the data that Roadrunner is getting and kind of enhance that sustainability offering even more. And so we think the combination of those two things is really compelling. And it’s been fun because we’re started deploying the technology through the Roadrunner portfolio of customers and we’re just now starting to see the cohorts come, the first cohorts of installations come back and validate those assumptions. So all of the theses that we built with the Roadrunner leadership team building the business case for why an acquisition made sense, we’re starting to see come to fruition and I think everyone’s really pleased about that.
Dylan: Nice. I said at the beginning that what is it? Less than half of all recyclable waste in the US is actually recycled. From everything you’ve seen and kind of your view of the future of this industry, do you think we’ll make good progress on that.
Jason: I do. I think it is going as Ben mentioned, things take more time than you expect. The numbers have not been reflected, I think, at the macro level or the nationwide level, but on the micro level, when we’re working with individual businesses and individual locations within that business, we are starting to see a real shift in behavior and the way people recycle and those recycling rates. And we can watch that path to zero waste, to landfill. We’re seeing companies really start to achieve that by diligent management of the back of house processes and they’re doing it largely by leveraging technology and making it an important part of their operations. So yeah, I’m optimistic that we’ll get there. I think it’s going to still take some time. And Ben mentioned the economics of implementing the technology that will help continue to make a change and regulations, stricter regulations will continue to make a change, but I think we’ll get there.
Dylan: Awesome. Similar question about your level of optimism, Ben, and I ask this of everybody on the show. How optimistic or pessimistic are you about the future of our planet and why?
Ben: I’m definitely more of a pessimist myself, just generally, but no, I am actually very optimistic. I think the right attention is being placed on climate change and just all the contributing factors to that. And equally, I guess, what climate tech needs to be built or used to solve some of those challenges. So I think I have the utmost confidence in society to build technology to solve problems when we’re focused on the right things. And I think right now we’re the big focus on climate tech as there should be, and I think we’ll do it.
Dylan: Awesome. Jason, who is one other company or person doing something to address climate change that’s inspiring you?
Jason: I’ve been really inspired by the team at Climate Tech VC, Sophie Purdom and her team. They’ve done an amazing job of filling one of the information gaps related to climate Tech, I think awareness and making information about Climate Tech accessible and reliable, trustworthy. They’ve done a really good job of hitting that product need. I’m a twice weekly reader of their newsletter and a big supporter.
Dylan: Awesome. That’s a great call out, actually. I use their info all the time too.
Jason: Yeah, it’s not necessarily hardware specific, but they certainly report on hardware companies and super cool technology like SciFi, mind-bending stuff, which is fun to read about.
Dylan: Yeah, totally. Ben, is there anyone you want to call out?
Ben: Don, other than you.
Dylan: Oh, and that would be a first, I’ll take it.
Ben: No, I love this podcast. So that is part of it. Ice is in a slightly different way. I think people are excited to get into the climate space and I think it can make a huge impact. I’m a huge fan of Dyson as a company, Dyson Vacuums and it sounds like a weird one, but they have an expertise in motors and batteries and airflow, and I think there’s probably so many places you can apply that to solve climate challenges. Heat pumps being a really key tool that people identified to solve climate change. I hope Dyson comes out with a heat pump.
Dylan: Oh, I like that.
Jayson: There’s the call out Dyson. The challenge has been posed.
Ben: If you listen to this podcast,
Dylan: I think he’s a frequent listener. Yeah. Okay, there it is. There’s your call to action. I love it. Let’s see. Jason, what advice do you have for someone who isn’t working in the climate today but wants to do something to help?
Jason: I think doing anything is a good thing, and it really starts with the first small step. So there are opportunities in all different segments of the climate tech industry, and regardless of expertise or previous experience, there’s opportunities to get involved. And so I think for whether you’re just coming out of undergrad and looking for your first entry level position or you’ve been an operator for your entire career, there’s places and there’s work to be done, and the problems are big and meaningful and exciting, and there’s ample funding behind them. And so I think it’s really just taking a first step and that’ll really help everybody.
Dylan: Yeah. How about you, Ben?
Ben: Well, I totally agree with what Jason said. I think just as all as individuals, we can all look at our own habits and do little things. I mean, every day I have the choice of throwing away something or recycling it. It’s those small decisions that do make an impact.
Dylan: Or breaking down your cardboard boxes.
Jason: Yeah. Which I guess I need to do. But I also think there’s just so many different reasons, there’s so many different types of companies tackling different aspects of climate change, and so I think there is a good opportunity for people to find their passion and apply it in a lot of different areas in the space.
Dylan: I love that. That really resonates with me. I mean, there’s just so much that needs to change. Every aspect of every business, almost every product in the world needs to be reinvented or has the opportunity to be reinvented with sustainability goals. So yeah, so much opportunity out there. Cool. Well, guys, that was really fun. Really. I love getting the true story, the origin story that I’ve never heard before and always love catching up with you. So I really appreciate your time.
Ben: Yeah, thanks so much, John, for having us.
Jason: Yeah, thank you. This is great.
Dylan: Hardware to Save a Planet is brought to you by Synapse. To find out more about us and how we develop hardware solutions for the world’s most ambitious companies, head to synapse.com and then make sure to search for Hardware to Save a Planet in Apple Podcasts, Spotify and Google Podcasts, or anywhere you like to listen. Make sure to click subscribe so you don’t miss any future episodes. On behalf of the team here at Synapse. Thanks for listening.
- Ben Chehebar on LinkedIn
- Jason Gates on LinkedIn
- RoadRunner Recycling Website
- RoadRunner Recycling Career Link
- RoadRunner Recycling Embedded Engineer Role
- Synapse Website