a16z Podcast - a16z Podcast: Construction Under Tech -- Info Flows

Episode Date: June 13, 2018

with Greg Lynn (@greglynnform), Gina Neff (@ginasue), Tracy Young (@Tracy_Young), and Hanne Tidnam (@omnivorousread) Construction has been one of the industries most resistant to innovation and change over the last decades -- productivity has actually decreased there while it has risen in other industries around it. So how are new technologies (finally!) beginning to transform the most brick-and-mortar of all the (literally!) brick-and-mortar industries? This episode of the a16z Podcast -- with Tracy Young, co-founder and CEO of PlanGrid; Greg Lynn, architect, professor at UCLA, and co-founder of Piaggio Fast Forward; and Gina Neff, sociologist at Oxford University (in conversation with Hanne Tidnam) -- considers the problems, and progress, in the construction industry. Information flows in particular are one area where tech is already making meaningful inroads into the construction process... will coordination follow?

Hi, and welcome to the A16Z podcast. I'm Hannah, and this episode is part of our series on construction. This conversation includes Tracy Young, co-founder and CEO of Plan Grid. Greg Lynn, architect, professor at UCLA, and co-founder at Piaggio Fast Forward, and Gina Neff, sociologist at Oxford University, who works on how new technology helps architects, engineers, and construction professionals collaborate. Given that this is one of these industries that are literally brick and mortar and has tended to be less obviously impacted right away by these new technologies, we want to wanted to get into what is happening, how tech is beginning to make an impact, how we think about the challenges in architecture and construction as an industry, the way information flows and

coordination happens, and how technology is beginning to improve that flow. I want to start by asking you guys about the basic tools. Are they still the same? So in 2007, as a rookie engineer working on my first construction project, I was issued one of these tablet PCs. There were like ruggedized laptops. And so the innovation at the time was running Microsoft Excel through these like seven pound tablet PCs in the field just to track issues. To track issues that would come up as construction was going on. Yes. And a seven pound laptop is not that easy to carry in the field for over half an hour. I remember my superintendent buying me the strap that goes around your head and attaches to the laptop so that you can just like crouch over to carry it better. So you know,

It's 10 years later, and obviously the hardware has changed, and because of that, companies can be building software specific for our project manager trying to access the information to build these projects in the field. There's been so much rapid technology change in how buildings are visualized and how these visualizations get shared. The idea behind building information modeling was to create a shared information space that architects, engineers, and builders could rapidly exchange files, and they would be, interoperable. I like to make the analogy of working together on a writing project in Microsoft Word or in Google Doc, as opposed to sharing locked PDFs by email. And sadly, in the construction industry, that kind of shared information space isn't developing as fast as we might see. The technology is certainly there. It's the legal, regulatory, and history of the industry that makes companies is very slow to change. One of the things we noticed around 2008, 2009, is that the technology

was bringing the people who plan construction and design projects closer together, but organizationally, they were still very far apart. You mean like behind the scenes, you know, planning in other offices? Absolutely. So the planning for the detailed construction, especially of the mechanical electrical plumbing systems, happens in construction. So, construction trailers near or close to the site. And so there would be the ways in which these systems could come into closer coordination for planning. And general contractors are using building information modeling to coordinate the work that people do. And yet, what was happening in a day-to-day process was often those construction plans weren't being able to be shared

quickly with people in the field. They couldn't quickly adapt them in those settings. They were still having to go through many layers of bureaucracy through many different companies in order to get changes made. And what was the difficulty in sharing that information stemming from? If you look at a construction site with a time-lapse camera, it's the most insane organization you could ever imagine to try to turn into a vertical streamlined process. You mean to actually watch a building be made over time from a pile of dirt to... between the logistics of machines coming to the site, goods arriving just in time, everything ready to be installed. It's an incredible logistical problem, and it's not happening in a factory

setting. And what happened in architecture is a lot of tools were invented by architects. Tools from the aerospace industry, in particular Revit and BIM, were adopted to allow people to collaborate on a joint three-dimensional file. So every screw, every bolt can be described, not just as a form, or a shape, but also what its weight is, what its cost is, where it's coming from, who's responsible to deliver it, who's responsible to install it. So what architects and engineers and everybody will do is you'll write the equivalent of a performance spec for another industry. So you'll convert everything from three-dimensional files, which are very robust and very sophisticated now, into two-dimensional drawings or into two-dimensional documents. And is that what

the blueprint is essentially. It's a stock and trade, but it has nothing to do with reality. I mean, you never build off of a plan or section. Really, you use it as a performance spec to measure dimension. It's simplifying it. 3D model's hard to consume. It's too much information, and the software just isn't in place to easily consume it in the field. For a sheet model person, all they need to know is that their duck work is a 10-inch round duck work, and it's 8 feet off finish floor, and they're going to hang their duck work. They don't need to see a model to do their work. Although the modeling in the coordination has been very successful, right?

You know, you can make a really rich visualization or a really rich model. And still, people are lacking the contextual information of what went into these decisions. So the sites that we observed, teams were continually going back and looking at decisions that had been made in months beforehand. But because there weren't people there in the room, who could tell them why that decision was there. And so when you're looking at a drawing or a specification, you know, I think this goes almost in any industry that there's only so much that we can expect

from large-scale information models and databases. What I think is really interesting in the construction industry is, you know, we bring up every bolt can be modeled in 3-D modeling. But, you know, the killer app for building information modeling really has been coordinating your plumbing work and your heating. and air conditioning work so they don't collide within the building system. Matching it all up, this huge amount of data, but without context or history. What you start to see is a lot of people now modeling process, construction process and

logistical processes, but when that steel beam arrives and when that sheet metal duct has to go through a webbing, and when the lighting comes in and then when the sheet rock and closes the ceiling, that whole process is captured occasionally in a specification. The performance spec. So the flow of labor and everything coordinating. It's not even in the spec. It's in the superintendent's schedule that changes from week to week because everything is always delayed. So it's basically in somebody's brain.

In several people's brains. Coming into the industry in 2007, where BIM wasn't as adopted in the U.S., we did something called light table coordination, which is this giant light box with a piece of glass where it shines through and you put paper over it. Then you layer on the mechanical drawings and the plumbing drawings and the structural drawings on top of it, physically. Physically layering them. In a room with like 12 other people to see like, okay, we're going to play. We are going to play Tetris because we're going to make all these systems fit. And so it's an awful, awful process.

Did that work? Thousands of hours of coordination of just trying to make these systems fit together because there's so little space over our head. And so what BIM does is eliminate that entire. light table process and make it fit in the model. The infuriating part with light tables is someone would always bump into those drawings, at which point everything would scatter. And it's like, oh, God, here Bob goes again. And then you have to realign it on one of the projects I was on. The owner paid something like $14 million for us to essentially play Tetris inside the model. We went into the field and the inspectors came out and said, hey, look at my stamp, look at my

signature, it's on the 2D set of drawings. It is not in this model. So how are the tools in your industry being used now in the construction process? The state of the field right now is that by developing some tools and borrowing some from aerospace, the collaboration between engineers, architects, building owners, and all the constituents is fairly integrated. The ability to cost and plan and logistically scope. Construction is pretty good. It's the minute you get all the entities that go into a building, which is a very flat structure from medieval people digging holes in the ground

to sophisticated robotic factories making curtain walls. And try to bring that all into a site in a timely way. It is less efficient now than it was in 1960. Because on the front end, we're very sophisticated and drawing in very detailed ways three-dimensional models. So there's a real disconnect. Down to the screw. And on the job site, you've got this mix of people, some of which are using phones, some are still using drawings, some are, you know, working on 3D files on trailers. And honestly, the waste is increased and the schedules have increased. Yeah, the gaps in timing and the workflow must be. Buildings are delivered

later now than they were in 1960 and more over budget than they were in 1960 because of this problem of getting those two things together. One of the things that we saw were teams. of coordinators who are helping to design and plan the systems that are going into buildings. And the real ingenuity and innovation that they would have, they were really problem solvers. And the frustration they would face, when they came up with a time-saving idea or a cost-saving idea, getting it through the levels of approval was so frustrating for them. Because even if they got approval from the general contractor, it would still have to go through multi-layers of engineering and architectural approval in order to get them information.

So that's partly why we saw the technology more tightly coupling the work teams together, but making them organizationally really quite separate from the rest of the project. And that sounds like what you were talking about, Tracy, with in terms of the inspector coming in, right? Like then you come up against this archaic, well, is it archaic the way that cities inspect? How does that tack on to this? So by and large, the whole world has figured. out how to increase labor productivity, except for construction, as Greg touched upon.

We are less productive today than we were 20 years ago than we were 60 years ago. And it's completely unbelievable. And strangely enough, it's like because of the innovation in a weird way. It's complicated. So the U.S. Bureau of Labor Statistics has written like an 84-page study on this. And basically the summary is it's complicated. The data's bad. Regulations are part of it, the people are part of it, because the construction involves just so many different entities. We're talking about tens of thousands of different material suppliers are part of this process. And then another 50 construction entities from the general contractors to electricians, to the plumbers, to the drywall folks. And then on top of that, they're counterparts on the

engineering and design side. And when so many different parties are involved, and communication needs to happen. And meanwhile, regulations changing. It's just this incredibly broken process, which requires so much managers just making sure that coordination is happening, making sure that the schedule is going to work, making sure that things are getting delivered on time, on schedule. And there hasn't been software to help with this process until very recently. There's certainly been project management software that's existed for the the last 1630 years, but the hardware didn't exist to support the folks in the field. So any software that was used in the space until 2010, when the iPad was first introduced

to the market, it was on straps around their next laptops. Exactly. It was either on straps, which is rare, or they're in the office where most of the people are not. It used to be 150 drawing set for a $100 million building was kind of a lot. And now you would expect a thousand drawings. So it's like these poor builders are just drinking out of a fire hose because they're used to maybe not so much information, but the right information?

Very different information flow and content. And now there's just so much information that goes into a document that what to use is really difficult. It's even more than that. We're the largest digital blueprint repository in the world. On a $150 million project in the Bay Area, we have seen 15,000 sheets of blueprints. Can you imagine trying to take 15,000 sheets of blueprints into the field?

It's impossible. Is it useful, meaningful information? As builders, your job is to translate these documents and these specs into physical structures. So yes. So is it all just an information management problem right now? Is that the crux of what's going on? Or are there ways in which we're starting to see that information get better managed or other tools that are coming up? I think a couple of things that are very interesting solutions.

One, which is mature robotics in order to go direct to fabrication with machines. Obviously, part of that might be because of the coordination and the data processing aspect of it, yes. But those that robots can understand that level of information that we're not used to drinking from the fire hose from. Certain parts of the building process, there aren't so many options or manufacturers. and so you'll see a company that's building curtain walls and they'll have some stock extrusions. Everything's always custom, but it's all coming from a easily quantified process. Right. So if you introduce robots into that fabrication and a factory setting, and you get the architects

and engineers to talk to those robots, then you get easy to plan, easy to cost solution. And can you just describe what the robots are physically doing and where they fit into that workflow? Well, the construction industry has been an industry for a while. So there have been industrial processes where machines for welding, cutting, extruding would have numerical control associated with them. They would tend to do the same thing over and over, but with better quality computation, you can start to get more and more fineness into that process. But it goes from steel industry, glass, wood, almost every trade. They're working now. with masonry and concrete, 60% of projects built in Switzerland use robotics at some significant

level in the construction process. And it's for a lot of reasons. Partially, the mortgages in Switzerland tend to be 50 to 70 years, so construction budgets are about double, so you can innovate and save more money there. But there, it's the most robotic construction industry on earth, is in Switzerland per capita. And what have they seen that do to jobs? Well, higher-skilled robotics jobs, but more and more of the buildings are built in factories and delivered as kind of mega components. The real bottleneck is what can you get on a road or on a train or on a boat. There's an Austrian kid that just has been teaching a computer to see bathroom plans and has been training it. The same way we're training our robots to see people and bicycles,

he's been training a robot to see bathroom plans. It's kind of the first job everybody gets out of graduate school is laying out bathrooms. Really? That's hilarious. I did not know that. Because there's not that many options in a bathroom. Is that the idea? So you just draw an 11 foot three by seven foot six rectangle and it'll lay out a bathroom for you. It's the kind of thing you would probably want to train a robot to do unless you really care about the way out of your bathroom. We want to innovate in that space. Or prefabrication, right? Because it takes somewhere between 10 to 20 different subcontractors that touches a bathroom, which is crazy, right? You only think you're like, you know, maybe the plumbers. But there's a concrete subcontractor. There's a mechanical subcontractor. the electricians, the plumbers, the drywall folks, the painters, the fixture installers, etc. Bathrooms typically are this common room.

Like there's one design and then it gets repeated over and over and over in the building, which makes it the perfect thing to manufacture off-site and then bring in. And we're seeing a lot of our contractors and customers working together off-site because they can control the schedule. Interesting. They can fabricate 20 of these things and then drop it into the building and then gain two weeks in the schedule versus trying to get one scope done at a time and then sequencing everyone in. You know, we don't think of them this way, but Disney builds cities. And I was able to visit

their building research facility in Shanghai. And they're thinking about the ways in which how do you understand heat generation on a block scale? How do you understand water? How do you and what kind of systems could you build if you were not thinking of the units of analysis as individual commercial residential homes? Because they're doing it in one fell swoop, right? It's not over two centuries. How is that vision of how we build that city changing? If we really want to take seriously the challenges facing the future of the planet, we're going to need to think both about how we reduce consumption, but also how we'd be really smart and the production of the future of our buildings and our cities.

The construction industry is so large. It is such an enormous part of American GDP that small improvements in energy output, small improvements in efficiency end up having enormous implications. Hospitals alone are 15% of energy, greenhouse gas production in the United States. So what are some of the ways that you're starting to see that happen?

some of the cool tools that we're going to start seeing. This is why we wake up every morning. Construction accounts for 13% of global output. Literally, one out of every 10 people in the world, eat, sleep, and breathe construction every single day. We're focusing on the human factors. Builders, frankly, spend a lot of time not building. They're trying to find information.

They're duplicating work by transcribing their notes that are written on paper over to whatever they're using to track these issues, the cloud exists, mobile devices exist, computing power can be taken out to the field with a battery life that lasts all day on a device that's super lightweight that doesn't waste seven pounds. And so we look at every other industry

and we think about how did the rest of the world find this massive increase in labor productivity and one of the main reasons is they started using computing power to do their jobs better. And construction couldn't benefit from that until very recently. When you think of a robot on a construction site, you think of an autonomous bulldozer or an autonomous backhoe or a crane that's able to drop goods at a specific time, a specific place.

And all those things exist. There's a finer grain with how you move there and you move on a scooter or a bike or a motorcycle rather than a car. And that's just because you don't need two tons of steel to carry a couple of gallons of milk. You could do that on an electric bicycle. So a finer grain approach to movement and mobility is also happening, I think, on construction sites at the scale of wheelbarrows, forklifts. Instead of these giant cats. Exactly.

Like if you're in a factory, what drives a factory are these little things called point-of-use carts. And if you're just screwing things together with an electric screw gun, the screws just show up when you need them on these little carts. Right. ready to go. If we can take that scale of construction to construction sites and have things queued outside the city that come into a site at just the right time, they're not there early, they're not there too late, and they get to the person that's doing the work, we think that scale of innovation is very ripe right now for development. We call them land drones. Any time in a factory, you take something overhead in a gantry, you have to shut down the whole factory. When a crane is

lifting heavy stuff. Because you're using that airspace so much. So littleer things on the ground that can help, you know, deliver things when you need them and deliver goods and help you build in an efficient way. I think it's very interesting. Are we seeing that kind of scale play out in textures, too, and materials? How are we thinking about finer grained things or different ways of using materials in that part of the process? So one of the heroes of lightweight construction was Buckminster Fuller, who said that you should judge the quality of every building by how much it weighs. Oh, my gosh. And it really does work.

If you want to save money, if you want to save time, and if you want to build more efficiently for the environment, just weigh what you're building. In every other industry, you'd want to have lighter, faster, stronger. So if you take a tall building, the biggest load on the building is it bending in the wind. It's actually not how it's carrying a load down. It's how much load is up in the air. If you can take 10,000 tons of steel out of the top of the building and replace it with carbon fiber, better. You're saving tons and tons and tons of material

and also saving a lot of time. And you can move that stuff more easily and people can actually lift it without machines. And so... So it trickles down in all sorts of ways, the cost savings. It's a very simple principle is just build. light, like be efficient and even use exotic materials that might not seem like they're sustainable because actually if you just use a little bit of them like plastics and instead of throwing them away

in a cup, build them into the built environment. When I was in high school, I had a brief internship with Michael Graves. And I remember Michael Graves used PVC columns to make these kind of like pillar like, you know, these Italian renaissance-looking things. They were just PVC tubing. He was a big collector Bakelite. We both shared a love of plastics. But no, like brick, which you think of as a natural good material, is the worst thing you can build with because you have to fire it with all this energy, you have to truck it. It's heavy. Trying to get weighed out with the little vehicles. You want to save energy, just make them lighter because you don't have to move as much stuff. You know, with your batteries or with your gasoline engines, make it lighter is really the first thing to do

before you do anything else. Then things get put together faster. You can move bigger, pieces. Everything gets stronger because it's consolidated. And it's happening more and more that composites and things from the 1960s, high performance boats and airplanes are now starting to show up on job sites. It's again a problem of getting it permitted and fireproof and there's some obstacles there. But it's a very interesting space of innovation right now. So we've talked a lot about sort of coordination and information flow and then about physical movement. But what about these visual technologies that are changing how we see things or changing how we can model and imagine them with like AR? How is that changing the workflow or the coordination or the

construction or the imagination, all of it? No one thinks in 2D plans and specifications. A blueprint is not a natural language for anybody. You can put a blueprint in front of three different people and they're going to see three very different things. And when you have new kinds of visualizations, Suddenly, you have new kinds of clarity around what people are seeing and what the problems are. The kinds of tools that we're talking about here that are at the forefront of construction and design really offer that opportunity for more creativity at the point of the problem. Something that always gets specked into larger projects, let's say anything over $50 million or more, is the concept of a mock-up room. They'll take the most common rooms, let's say bathrooms or in a hospital exam rooms. and they'll build a physical template of it called a mock-up.

That you can walk in, too, to scale. That you can walk in, and what they'll do is bring the end doctors and nurses in to confirm. And especially for hospitals, right, that's notoriously difficult design problem into a hospital room because there's so much. Right. There's ergonomics. There's like everything needs to be in the right place. Exactly. And so these mock-ups are physical, and they cost a lot of money.

And more often than not, they'll build. build them off of the side of the construction site, and then once everyone approves it and signs off on it, they'll just demolish it. Wouldn't it be nice if we could just put on a headset and then see it and have the building owner or these doctors and nurses walk through it in the model, confirm. This is exactly where I want these medical equipments and then call it day. When for me, what underlies all of these technologies is a very simple concept that is localization, whether it's the new iPhone or an autonomous car. What they're all doing is building. building a three-dimensional map and simultaneously locating that camera inside that three-dimensional

space. Yeah. And so it's not so much about the autonomy of a vehicle being able to get to the place. It's that there's a three-dimensional map, and everything in that map has a sense of where it is. So if you put on a HoloLens, you can suddenly see that's where my duck needs to go. You get to see where something is supposed to be in space localized. You know where it is. And that's huge for the construction industry.

And it's also huge to know that the machines you're using are seeing the same thing in the same way that you are. So suddenly the workers can now work with the robots, where to put all the screws in the drywall. And a red dot turns to green when that's where you're supposed to put the screw in to hang the drywall. And then you record it and send it to the inspector. And they know that you're the one that put those screws in on that schedule like you're supposed to do in the specification. But it's a real collaboration between robotics and people in a way that you're not really. replacing them in a factory somewhere, but you're bringing that level of technology onto a job site that's...

Well, you're projecting that light table, basically, and seeing through it, right? Those layers... It's definitely the future, and construction is this industry that has to be obsessive over safety. Every single job site in all of U.S. will have a billboard on the gates of their job site with the number of days since the last safety incident. Right. And one of the challenges with these new devices that are, you know, wearable on your face is that OSHA's not okay with it. What are the objections to it and how does regulation need to change? These glasses that I have, just regular old glasses, I have to have these

plastic clips that go on the side just for me to walk into a job site because they're not safety glasses. It's very bureaucratic and slow. And so most of the codes don't change for like seven to ten years at a time. And then there's a lot of interpretation that goes on. So it's going to be challenging. But with that said, there's easy wins, right? Like digitizing of information. There's this massive correlation between an industry's ability to digitize their information and this massive increase in labor productivity. And when you walk into a job site, even on projects that are paperless, we still see paper. It's crazy. The real question is, though, who's incentivized in all these stakeholders?

If it's agriculture, that's a vertical industry now. If it's mining, that's a vertical industry now. So there's an incentive. Whereas in the construction industry, the 30% waste and the, you know, 20% of late schedule, it's baked in from the finance all the way down to the smallest subcontractor. So who in that chain is going to go ahead and develop the AR device that's OSHA certified and why? It's true. It's construction is an industry that there's a reason why they're slow to adopt technology because their margins are so slim. The top construction firms, the ones who employ over 10,000 field workers, hire only 5% of all builders.

Whereas in services or manufacturing, the top firms, they're employing like 25% of the industry. So what that means is in construction, it is ran and built by a lot of regional small mom and pop contractors who are all competing against each other. So their margins for a general contractor are something like 1 to 3%. That's incredibly low. What other industries has such slim margins. Restaurant is probably the next one, and even that's like 10 to 15%. So because margins are so slim, there's not like this bank roll of cash in the bank

where they can invest in technology to do their jobs better. It's easier for the industry to hire more workers, and then if things aren't going well to fire them, then to make this capital investment on new technologies. Do you think this level of consolidation is going to change? change in the construction industry. I mean, here we have an industry that has a very old way of doing it. There's several very large players in the industry, you know, transnational, multi-billion dollar companies.

And yet, when people want a solution, a building solution delivered, they often have to go and get multiple pieces of that. I mean, think about any other product that we want to buy. If we want an automobile, an airplane, we actually don't go and contract the designer, the engineer, and the builder, the manufacturer together, right? We ask for the thing itself. And, you know, when some of these technologies were first emerging, people were talking about the fact that a large-scale manufacturing concern might really integrate and deliver integrated solutions. Do you think that's possible? In certain totalitarian countries with high labor costs, like Singapore or the Emirates, there's an incredible investment in innovation and construction.

And that includes materials like building in carbon fiber, like the Apple headquarters was all built in Dubai because they built a facility to build carbon fiber at the scale of buildings there. It's high labor cost and they could permit everything and make it legal because it's, You know, you just ask one person and it's all smooth. Other than that, I think it's very grassroots. And most of the action, which there's a lot of, is smaller companies that are innovating little chunks in terms of getting robots onto construction sites, making vehicles smart, solving this whole problem of a high fidelity 3D model, delivering it to a construction site. There are a lot of small players doing things. Construction is the infrastructure of this world. And if we're unable to figure out how to solve this productivity problem, we'll be unable to deliver the infrastructure that this world needs, not in the next 15 years and the next decade.

It's really important that we get this right because the building sector in the United States accounts for 5.50% of energy consumption in the U.S. And so if we can make even small improvements in the efficiency, both in the construction process and in buildings themselves, we can have a huge impact on the future of the planet. The way to do that is to bring these kinds of innovative solutions into the point of the problems that are faced in construction and allow for more creative and energy efficient design to happen at lower costs. It might be that we don't look to the biggest engineering and infrastructure companies to innovate. It might be that we look at, you know, somebody in eastern Washington State doing amazing things with timber construction that's sustainable, fast, affordable, and very high tech. And that's kind of where the action is. And I think also in scale, focusing on the little things from hand tools to wheelbarrows is also where a lot of the action is. So I think we focused on like the big global, you know, industrial problem at big scale.

scale. And it might be in the construction industry. You've got to actually shift if you want to make an impact to finer. Yeah. Thank you guys so much for joining us on the A16Z podcast. Thank you. Thank you.