Thanks for this post Brian. You touch briefly on the common problem with the construction innovations in your list--that they're lateral improvements on products, and have practically no impact on construction productivity. A fiberglass door takes as long to hang as wood door. In the case of vinyl windows, and by extension all modern clad-composite windows, proper installation actually requires an increase in labor inputs. PEX is a little faster to install than copper, but any productivity gain gets more than canceled when homebuyers demand more plumbing fixtures.
In theory, SIPS offer the most productivity gains on overall shell construction, but they create an additional planning burden on design and site management. They also can slow down some of the trades because of the challenges associated with running wiring and plumbing.
In my brief career in single family home architecture the highest impact innovation has probably been LED lighting. Even in that area, easier and faster installation hasn't moved the needle on construction costs or speed.
There is quite an extensive literature about the diffusion of innovations sitting at the intersection of the disciplines of economic history, and productivity studies within economics (JEL code D24, mainly). Being economists, the economists often look at proxies like patents and royalties rather than directly at individual innovations, though. The historians are better.
Vaclav Smil has summarised some research about the time required for innovations, with a focus on what he calls "prime movers", augmentations of muscle power such as wind power, water power, steam engines and turbines, electric motors, internal combustion engines and jet turbines.
His recent book "Grand Transitions" covers these, although I found his earlier "Energy In Nature and Society" to be better for some of them. The latter book also has charts of (USA) adoption curves for telephones (landlines), radios, refrigerators, color TVs and more.
In almost every case the adoption cycle falls into this 20-to-40-year duration that you have identified. Some of his innovations, like nuclear power, also saturate well below 100%.
Incidentally, the tractor is not a good example of an innovation for reasons of new capabilities. Pretty much everything that tractors did initially--pulling things, ploughing, harrowing, seeding, etc.--was done with horses and horse-powered machinery beforehand. In "Energy In Nature and Society" there is a photo of a team of 20 horses pulling a McCormick combine harvester. Tractors were also preceded by traction engines -- the same thing, except steam powered and consequently larger and heavier. New capabilities--front-end loaders, spraying equipment, etc.--came after tractors were established. Initially a tractor was a low-latency, tireless horse: you didn't have to spend an hour fetching it and harnessing it before starting work, nor did you have to manage its workload so carefully.
For tractors, the biggest difference between a horse and a tractor is that a tractor can be used as a power source for other farm equipment. This functionality was added very early (you can see it on this 1917 Ford tractor, for instance https://en.wikipedia.org/wiki/Fordson), and makes them pretty different than just a tireless horse.
I'm not sure to the extent to which this functionality existed on steam powered traction engines, or what the penetration of that technology was (I looked a bit into this and couldn't find any easily available literature), so I'm not sure what the traction engine -> tractor transition looked like.
Thank you for all of the writing your regularly publish Brian. It has greatly informed my thinking around construction innovation and I enthusiastically look forward to reading each new post.
I've tried to identify the 'dimensions' of construction innovation and I would love to know if I'm missing something:
- Cost
- Complexity
- Time
- Functionality
- Environmental impact
- Aesthetic value
Further to these dimensions of innovation, how would you compare their contribution to the success of specific historical construction innovations. For example, does an innovation that decreases cost more likely to succeed than an innovation that decreases the time to completion? In this post about innovation uptake, you mentioned that innovations in construction rarely increase functional capability (like the arguably unsuccessful smart glass); I'm curious how successful smart glass would have been if it did not also increase cost and complexity. I'm also curious about how the dimensions interact, such as with innovations that decrease cost and decrease complexity (eg. PEX) - are there multiplicative effects? I'm not sure if you take requests for future posts or have time time to answer my query but I'd love to hear your perspective on the idea I've put forward here.
I wonder if the lack of uptake has to do with the limited way that they are implemented.
For example, if all a developer does to implement factory-built modular construction is to build a factory and install some heavy duty cranes to deliver the large modules into place in a building, it's hard to see how that will result in any time and money saved.
It seems to me you'd have to consider transportation networks - including the well-known constrictions of too narrow flatbed trucks that have to fit on roads. You also have to consider running the factory 24/7, not just when the sun shines on the site, and maybe some kind of storage facility until the job site is ready to accept the modules, or, better yet, maybe, to install kleig lights at the site so it can run 24/7 too, if that's allowed. Of course, this comes with tradeoffs in terms of labor costs for night shift workers, but the upside of getting a project done in 1/2 or even 1/3 the time ought to outweigh that.
Innovation may not be working simply because it is not innovative enough.
Many of the innovations you discuss are already in use, *where they provide benefits that outweigh the costs.* But that's not often, again due to the nature of construction.
For example, I've had jobs where they run 24/7, either because we allowed it or, in a few cases, required it. Now, my experience is in heavy civil works for the federal government, which is going to be something of a different animal than what Brian is discussing.
However, a few of the main considerations still apply. The first, of course, is cost. Night shifts are more expensive than day shifts, and indeed you've identified one place (providing Wacker lights) where that's obvious. There's also a burden on our personnel to provide oversight, which means we (the Government) need to spin up more shifts as well. Secondly, it's more dangerous to work at night, lights or no, and construction is already one of the most dangerous occupations in modern America.
Overall, for typical jobs this often doesn't pencil out to do night work. Where we (on the owner side here) tend to allow overnight work is where we have what we think is a relatively tight schedule, but we'll let the contractor do their scheduling and then ask for night work if they think it needs it to meet schedule, allowing a balance of work schedule vs. cost depending on what resources the contractor can access. Required overnight work is typical for emergencies, where there isn't a schedule other than "unf**k what's broken as soon as possible." It's also sometimes required for routine construction in e.g., dam construction, where you want to keep concrete placements going around the clock. For example, the repairs to the Oroville Dam spillway fit into both these buckets: it was an emergency, and you shouldn't starve the RCC placements. They did that work around the clock.
For storage, that's also possible. But again, it's more fraught than it seems in a blog comment. One big thing is that buildings are, well, building sized. You need a big, huge, yard to store an entire building. You very rarely have that anywhere near the building under construction--laydown space on site is *extremely* precious on the vast majority of projects--so you'd need to be hauling your components from far away, anyway. For the on-site work, it may not actually make a big difference whether these modules are coming directly from the factory which is producing them as-needed, or if they're coming from a yard. The storage yard in this case is more acting as a buffer for the factory to allow it to be utilized more efficiently than something speeding up the on-site work.
One of many other considerations for storage is security. Your building modules are going to have plenty of pilferable materials in them. Of course, security patrols are a mitigation (at an additional cost), but you really are going to have these modules just sitting for a very long time while you finish, and there are going to be things that are hard for them to check. What happens when you go to set your module containing the crappers and find that all of the valves have been stolen in the last year while it was sitting in your yard? I mean, you replace them obviously, but that adds additional on-site work, and on-site work that is difficult to plan for, since you can't know for sure what might get stolen ahead of time. Not an insoluble problem, but one that eats into the cost efficiency of using storage to shorten timelines. (Even current construction methods have theft issues, but to a certain extent they can minimize it by reducing the amount of stuff you keep on hand at risk of being stolen. By definition, building a huge portion of the building and keeping it in storage means you have a huge portion of the building on hand, in a half-finished state, at risk of being stolen.)
I don't want to say that these two examples (night work and long-term storage to shorten on-site work) can't be done--they're done frequently. But they're not silver bullets, just tools that in some circumstances make sense but usually don't.
To be fair, comparing innovation in construction (or any other industries for that matter) to that of information technology is a bit unfair. The rate of innovation in information or computer technology is unprecedented in modern history. Since 1970 you could think of an important innovation happening in the sector every year.
Thanks for this post Brian. You touch briefly on the common problem with the construction innovations in your list--that they're lateral improvements on products, and have practically no impact on construction productivity. A fiberglass door takes as long to hang as wood door. In the case of vinyl windows, and by extension all modern clad-composite windows, proper installation actually requires an increase in labor inputs. PEX is a little faster to install than copper, but any productivity gain gets more than canceled when homebuyers demand more plumbing fixtures.
In theory, SIPS offer the most productivity gains on overall shell construction, but they create an additional planning burden on design and site management. They also can slow down some of the trades because of the challenges associated with running wiring and plumbing.
In my brief career in single family home architecture the highest impact innovation has probably been LED lighting. Even in that area, easier and faster installation hasn't moved the needle on construction costs or speed.
There is quite an extensive literature about the diffusion of innovations sitting at the intersection of the disciplines of economic history, and productivity studies within economics (JEL code D24, mainly). Being economists, the economists often look at proxies like patents and royalties rather than directly at individual innovations, though. The historians are better.
Vaclav Smil has summarised some research about the time required for innovations, with a focus on what he calls "prime movers", augmentations of muscle power such as wind power, water power, steam engines and turbines, electric motors, internal combustion engines and jet turbines.
His recent book "Grand Transitions" covers these, although I found his earlier "Energy In Nature and Society" to be better for some of them. The latter book also has charts of (USA) adoption curves for telephones (landlines), radios, refrigerators, color TVs and more.
In almost every case the adoption cycle falls into this 20-to-40-year duration that you have identified. Some of his innovations, like nuclear power, also saturate well below 100%.
Incidentally, the tractor is not a good example of an innovation for reasons of new capabilities. Pretty much everything that tractors did initially--pulling things, ploughing, harrowing, seeding, etc.--was done with horses and horse-powered machinery beforehand. In "Energy In Nature and Society" there is a photo of a team of 20 horses pulling a McCormick combine harvester. Tractors were also preceded by traction engines -- the same thing, except steam powered and consequently larger and heavier. New capabilities--front-end loaders, spraying equipment, etc.--came after tractors were established. Initially a tractor was a low-latency, tireless horse: you didn't have to spend an hour fetching it and harnessing it before starting work, nor did you have to manage its workload so carefully.
For tractors, the biggest difference between a horse and a tractor is that a tractor can be used as a power source for other farm equipment. This functionality was added very early (you can see it on this 1917 Ford tractor, for instance https://en.wikipedia.org/wiki/Fordson), and makes them pretty different than just a tireless horse.
I'm not sure to the extent to which this functionality existed on steam powered traction engines, or what the penetration of that technology was (I looked a bit into this and couldn't find any easily available literature), so I'm not sure what the traction engine -> tractor transition looked like.
Thank you for all of the writing your regularly publish Brian. It has greatly informed my thinking around construction innovation and I enthusiastically look forward to reading each new post.
I've tried to identify the 'dimensions' of construction innovation and I would love to know if I'm missing something:
- Cost
- Complexity
- Time
- Functionality
- Environmental impact
- Aesthetic value
Further to these dimensions of innovation, how would you compare their contribution to the success of specific historical construction innovations. For example, does an innovation that decreases cost more likely to succeed than an innovation that decreases the time to completion? In this post about innovation uptake, you mentioned that innovations in construction rarely increase functional capability (like the arguably unsuccessful smart glass); I'm curious how successful smart glass would have been if it did not also increase cost and complexity. I'm also curious about how the dimensions interact, such as with innovations that decrease cost and decrease complexity (eg. PEX) - are there multiplicative effects? I'm not sure if you take requests for future posts or have time time to answer my query but I'd love to hear your perspective on the idea I've put forward here.
I wonder if the lack of uptake has to do with the limited way that they are implemented.
For example, if all a developer does to implement factory-built modular construction is to build a factory and install some heavy duty cranes to deliver the large modules into place in a building, it's hard to see how that will result in any time and money saved.
It seems to me you'd have to consider transportation networks - including the well-known constrictions of too narrow flatbed trucks that have to fit on roads. You also have to consider running the factory 24/7, not just when the sun shines on the site, and maybe some kind of storage facility until the job site is ready to accept the modules, or, better yet, maybe, to install kleig lights at the site so it can run 24/7 too, if that's allowed. Of course, this comes with tradeoffs in terms of labor costs for night shift workers, but the upside of getting a project done in 1/2 or even 1/3 the time ought to outweigh that.
Innovation may not be working simply because it is not innovative enough.
Many of the innovations you discuss are already in use, *where they provide benefits that outweigh the costs.* But that's not often, again due to the nature of construction.
For example, I've had jobs where they run 24/7, either because we allowed it or, in a few cases, required it. Now, my experience is in heavy civil works for the federal government, which is going to be something of a different animal than what Brian is discussing.
However, a few of the main considerations still apply. The first, of course, is cost. Night shifts are more expensive than day shifts, and indeed you've identified one place (providing Wacker lights) where that's obvious. There's also a burden on our personnel to provide oversight, which means we (the Government) need to spin up more shifts as well. Secondly, it's more dangerous to work at night, lights or no, and construction is already one of the most dangerous occupations in modern America.
Overall, for typical jobs this often doesn't pencil out to do night work. Where we (on the owner side here) tend to allow overnight work is where we have what we think is a relatively tight schedule, but we'll let the contractor do their scheduling and then ask for night work if they think it needs it to meet schedule, allowing a balance of work schedule vs. cost depending on what resources the contractor can access. Required overnight work is typical for emergencies, where there isn't a schedule other than "unf**k what's broken as soon as possible." It's also sometimes required for routine construction in e.g., dam construction, where you want to keep concrete placements going around the clock. For example, the repairs to the Oroville Dam spillway fit into both these buckets: it was an emergency, and you shouldn't starve the RCC placements. They did that work around the clock.
For storage, that's also possible. But again, it's more fraught than it seems in a blog comment. One big thing is that buildings are, well, building sized. You need a big, huge, yard to store an entire building. You very rarely have that anywhere near the building under construction--laydown space on site is *extremely* precious on the vast majority of projects--so you'd need to be hauling your components from far away, anyway. For the on-site work, it may not actually make a big difference whether these modules are coming directly from the factory which is producing them as-needed, or if they're coming from a yard. The storage yard in this case is more acting as a buffer for the factory to allow it to be utilized more efficiently than something speeding up the on-site work.
One of many other considerations for storage is security. Your building modules are going to have plenty of pilferable materials in them. Of course, security patrols are a mitigation (at an additional cost), but you really are going to have these modules just sitting for a very long time while you finish, and there are going to be things that are hard for them to check. What happens when you go to set your module containing the crappers and find that all of the valves have been stolen in the last year while it was sitting in your yard? I mean, you replace them obviously, but that adds additional on-site work, and on-site work that is difficult to plan for, since you can't know for sure what might get stolen ahead of time. Not an insoluble problem, but one that eats into the cost efficiency of using storage to shorten timelines. (Even current construction methods have theft issues, but to a certain extent they can minimize it by reducing the amount of stuff you keep on hand at risk of being stolen. By definition, building a huge portion of the building and keeping it in storage means you have a huge portion of the building on hand, in a half-finished state, at risk of being stolen.)
I don't want to say that these two examples (night work and long-term storage to shorten on-site work) can't be done--they're done frequently. But they're not silver bullets, just tools that in some circumstances make sense but usually don't.
To be fair, comparing innovation in construction (or any other industries for that matter) to that of information technology is a bit unfair. The rate of innovation in information or computer technology is unprecedented in modern history. Since 1970 you could think of an important innovation happening in the sector every year.