Great content, thanks for sharing. Casual take as a small builder: What HAS worked to increase construction productivity is innovation at the component level: power tools, trusses, pex pipes, click-together flooring, digital plans and permits…etc. These improve productivity in every locale, in every market condition, and at every scale (from DIY to Lennar). Maybe someone will crack the code to truly factory-ize housing one day, but I can’t envision it. As long as housing remains so fragmented and inconsistent for the structural reasons pointed out by the author (plus the innate variability of land itself) I think production builders are as factory as it gets. Continuing to innovate at the component level can and will, however, allow us to enjoy capitalism’s magic “more for less” promise in housing.
I note that about 10% of the US population currently lives in manufactured/modular or manufactured housing, which is a niche but far more than nothing.
I worked as a laborer in a modular home plant during the 1970s but it operated in a very different way than these startups:
-the plant had minimal automation...mainly just fixtures, regular tools with central pneumatic power and relatively skilled, non-union labor. Also, we were paid piece-rate (and teams allowed to go home when work was done for the day), so the pace was fast and motivated by teammates, and also labor cost was directly tied to output. However, pay was quite good for the area and time ($20-25k/yr in 1979, similar to average engineer pay) , especially since most days were <8h. However, only young people could keep up with the pace. Also, excess people were laid off during seasonal downturns.
-The main cost advantages were protection from the weather, centralized purchasing and materials close at hand, plus a limited range of options (about 45 floor plans, but also trim options.)
In recessions, these companies would typically restructure or go bankrupt, but the buildings would remain and could be readily outfitted to begin again once the economy improved.
In terms of why these homes fit a niche at a good price point:
-Homes consisted of two sections, each max road-travel width (12") and of varying length. The sections were hauled to site with a semi tractor, backed onto a prepared foundation side-by-side using only the delivery semi (no crane needed). Installation was completed by joining sections together onsite in <1 day by a crew of 3-4 people, since all the internal electrical, plumbing and heating was pre-installed in the factory and designed so the two modules and site connections could readily joined. I believe the whole process, assuming the buyer had land and financing set up, from selecting a home style at a dealer and placing an order until home was installed was on the order of 6 months.
-The factory was in northern Indiana, so in general the terrain is fairly flat plus land plentiful and inexpensive in the area.
The resulting homes were single-story, 1500-2000sf and well-designed and relatively inexpensive, but definitely a step up from an apartment or trailer home.
Great recap. I think the market dynamics of housing cycles makes pretty much any model that requires a factory incompatible. Couple that with the fractured regional requirements limiting the scalability of any design and you've got a recipe for disaster. Gene's comment on innovation at the component level is very astute. I think the path to scale for builders will come through acquisitions and standardization of operations for efficiency gains, keeping the regional knowledge with local builders. All that said, if we want to build more homes, we need innovation of policy and much more support at the government level. Right now there are just too many hurdles and roadblocks.
This was a great postmortem primer! I’ve always been a fan of modular construction and never really saw the VC model as a good fit, but this forced me to really consider what sort of modular business could thrive rather than get bogged down with access to that kind of capital. Seems to me modern-day ARM might be a good analogy (although it looks like a UK company called Modulous basically tried this and failed - but maybe not due to the same reasons listed here)
Coming from the manufacturing world, the concept of regional micro-factories is frankly a pipe dream.
I would say that based on your analysis, what they failed to do in manufacturing terms is establish sufficient model diversity. Of course a design for say, the regional climate of Texas isn’t going to be a great fit for the regional climate of Wisconsin. But certainly we can also imagine that substantially similar models sharing substantially the same or similar components could produce a range of models compatible with various climate/environmental factors, and that such an approach would allow for centralized manufacturing shipped and assembled on site.
Regulatory and inspection are more a political tangle to unknot; I agree that rushing to automate with robots is useless without a functional production line in place first. I’d be curious to know where these failed start-ups were siting their factories? Ideally you’d want someplace with rail transport accessibility (and without congested access to that rail transport), plenty of cheap land, closer to a city with plenty of skilled tradesmen. But I also have a sneaking suspicion they were trying to build factories in California to be close to their VC investors.
My hunch is that the next group of modular startups will be direct to ‘consumer’ Auxiliary Dwelling Units (in-law suites) designed to be sold to existing homeowners / landlords to ‘infill.’ That would sidestep a lot of the financing problems of trad development— the land is already purchased, if financed by a homeowner they might even able be able to use the existing home as collateral, etc. Presumably a two story design built on the equivalent of a double-wide concrete slab could be packed into a single standard shipping container for transport by rail and truck, the repacked into a box truck the equivalent to a large moving truck for access to residential streets for last mile delivery.
Interesting article. I had not heard of any of those residential modular factories, I guess because I don’t hang out with the “tech bros” that funded them. The author writes like the modular industry was ripe for “disruption”, but the fact is, inventor types have tried and failed to revolutionize pre-fab construction for decades. It sounds like these companies could fail or succeed; didn’t matter, so long as the managers were in on the gravy train. The results would indicate that anyway.
These truths we hold inalienable:
Overhead kills factories
Air is cheaper than wood
There is no cheaper construction than a 2x4 stud with T1-11 on one side and drywall on the other
Get the water off the roof
Modular residential construction can’t compete with site built housing tract construction because 1. The housing tract jobsite is the factory. No extra overhead to cover. 2. The carpenters on piece work can put up the framework so fast your head will spin. 3. Homebuyers want a site built look and that means a steep roofline, which has to be built on site. 4. The sales tax on modular is more than the sales tax on site built, which is just an extra cost with no benefit to the buyer.
What is working right now in modular residential is multi-story in-fill buildings. But that appears to be all taxpayer subsidized work to solve the homeless problem. Companies like Autovol, Full Stack and Iron Town Modular are getting major projects done. Guerdon Modular is also doing repeat business in multi-story hotels, so that must be working even without subsidies.
We have prefabricated components - the construction industry has shifted to consolidating systems over time. The reason we haven’t seen it succeed is we are looking in the wrong place - the idealized volumetric solution has yet to break the market, but the construction industry has already been productized and panelized. It simply needs more vertical integration, not a start from scratch solution.
Why not mention Clayton Homes of Berkshire Hathaway.? Their key success lies in offering a product in areas where land valuations can still support a small house investment. The real failure of the companies you mentioned is the US population growth has pushed up land valuation significantly and a modular manufactured home can’t exist the same way a 30 mph scooter can’t navigate the majority of highway systems today vs 50 years ago.
Great content, thanks for sharing. Casual take as a small builder: What HAS worked to increase construction productivity is innovation at the component level: power tools, trusses, pex pipes, click-together flooring, digital plans and permits…etc. These improve productivity in every locale, in every market condition, and at every scale (from DIY to Lennar). Maybe someone will crack the code to truly factory-ize housing one day, but I can’t envision it. As long as housing remains so fragmented and inconsistent for the structural reasons pointed out by the author (plus the innate variability of land itself) I think production builders are as factory as it gets. Continuing to innovate at the component level can and will, however, allow us to enjoy capitalism’s magic “more for less” promise in housing.
I note that about 10% of the US population currently lives in manufactured/modular or manufactured housing, which is a niche but far more than nothing.
I worked as a laborer in a modular home plant during the 1970s but it operated in a very different way than these startups:
-the plant had minimal automation...mainly just fixtures, regular tools with central pneumatic power and relatively skilled, non-union labor. Also, we were paid piece-rate (and teams allowed to go home when work was done for the day), so the pace was fast and motivated by teammates, and also labor cost was directly tied to output. However, pay was quite good for the area and time ($20-25k/yr in 1979, similar to average engineer pay) , especially since most days were <8h. However, only young people could keep up with the pace. Also, excess people were laid off during seasonal downturns.
-The main cost advantages were protection from the weather, centralized purchasing and materials close at hand, plus a limited range of options (about 45 floor plans, but also trim options.)
In recessions, these companies would typically restructure or go bankrupt, but the buildings would remain and could be readily outfitted to begin again once the economy improved.
In terms of why these homes fit a niche at a good price point:
-Homes consisted of two sections, each max road-travel width (12") and of varying length. The sections were hauled to site with a semi tractor, backed onto a prepared foundation side-by-side using only the delivery semi (no crane needed). Installation was completed by joining sections together onsite in <1 day by a crew of 3-4 people, since all the internal electrical, plumbing and heating was pre-installed in the factory and designed so the two modules and site connections could readily joined. I believe the whole process, assuming the buyer had land and financing set up, from selecting a home style at a dealer and placing an order until home was installed was on the order of 6 months.
-The factory was in northern Indiana, so in general the terrain is fairly flat plus land plentiful and inexpensive in the area.
The resulting homes were single-story, 1500-2000sf and well-designed and relatively inexpensive, but definitely a step up from an apartment or trailer home.
Great recap. I think the market dynamics of housing cycles makes pretty much any model that requires a factory incompatible. Couple that with the fractured regional requirements limiting the scalability of any design and you've got a recipe for disaster. Gene's comment on innovation at the component level is very astute. I think the path to scale for builders will come through acquisitions and standardization of operations for efficiency gains, keeping the regional knowledge with local builders. All that said, if we want to build more homes, we need innovation of policy and much more support at the government level. Right now there are just too many hurdles and roadblocks.
This was a great postmortem primer! I’ve always been a fan of modular construction and never really saw the VC model as a good fit, but this forced me to really consider what sort of modular business could thrive rather than get bogged down with access to that kind of capital. Seems to me modern-day ARM might be a good analogy (although it looks like a UK company called Modulous basically tried this and failed - but maybe not due to the same reasons listed here)
Coming from the manufacturing world, the concept of regional micro-factories is frankly a pipe dream.
I would say that based on your analysis, what they failed to do in manufacturing terms is establish sufficient model diversity. Of course a design for say, the regional climate of Texas isn’t going to be a great fit for the regional climate of Wisconsin. But certainly we can also imagine that substantially similar models sharing substantially the same or similar components could produce a range of models compatible with various climate/environmental factors, and that such an approach would allow for centralized manufacturing shipped and assembled on site.
Regulatory and inspection are more a political tangle to unknot; I agree that rushing to automate with robots is useless without a functional production line in place first. I’d be curious to know where these failed start-ups were siting their factories? Ideally you’d want someplace with rail transport accessibility (and without congested access to that rail transport), plenty of cheap land, closer to a city with plenty of skilled tradesmen. But I also have a sneaking suspicion they were trying to build factories in California to be close to their VC investors.
My hunch is that the next group of modular startups will be direct to ‘consumer’ Auxiliary Dwelling Units (in-law suites) designed to be sold to existing homeowners / landlords to ‘infill.’ That would sidestep a lot of the financing problems of trad development— the land is already purchased, if financed by a homeowner they might even able be able to use the existing home as collateral, etc. Presumably a two story design built on the equivalent of a double-wide concrete slab could be packed into a single standard shipping container for transport by rail and truck, the repacked into a box truck the equivalent to a large moving truck for access to residential streets for last mile delivery.
Hmm, a lot to think about it. Thanks for sharing your insights on this.
Interesting article. I had not heard of any of those residential modular factories, I guess because I don’t hang out with the “tech bros” that funded them. The author writes like the modular industry was ripe for “disruption”, but the fact is, inventor types have tried and failed to revolutionize pre-fab construction for decades. It sounds like these companies could fail or succeed; didn’t matter, so long as the managers were in on the gravy train. The results would indicate that anyway.
These truths we hold inalienable:
Overhead kills factories
Air is cheaper than wood
There is no cheaper construction than a 2x4 stud with T1-11 on one side and drywall on the other
Get the water off the roof
Modular residential construction can’t compete with site built housing tract construction because 1. The housing tract jobsite is the factory. No extra overhead to cover. 2. The carpenters on piece work can put up the framework so fast your head will spin. 3. Homebuyers want a site built look and that means a steep roofline, which has to be built on site. 4. The sales tax on modular is more than the sales tax on site built, which is just an extra cost with no benefit to the buyer.
What is working right now in modular residential is multi-story in-fill buildings. But that appears to be all taxpayer subsidized work to solve the homeless problem. Companies like Autovol, Full Stack and Iron Town Modular are getting major projects done. Guerdon Modular is also doing repeat business in multi-story hotels, so that must be working even without subsidies.
Is there something that can be learnt from Sweden?
Yes, mainly that we should not emulate their model. It doesn't/didn't translate.
We have prefabricated components - the construction industry has shifted to consolidating systems over time. The reason we haven’t seen it succeed is we are looking in the wrong place - the idealized volumetric solution has yet to break the market, but the construction industry has already been productized and panelized. It simply needs more vertical integration, not a start from scratch solution.
Why not mention Clayton Homes of Berkshire Hathaway.? Their key success lies in offering a product in areas where land valuations can still support a small house investment. The real failure of the companies you mentioned is the US population growth has pushed up land valuation significantly and a modular manufactured home can’t exist the same way a 30 mph scooter can’t navigate the majority of highway systems today vs 50 years ago.
thanks for an excellent post.