Fellow former tradesman and perennial DIY-er here, I think this is dead on.

There is a fundamental mismatch between the economics of startups and the economics of trades. Startups and technology work best when you can achieve scale quickly and without friction by using the same solution repeatably. Software is ideal for this, it scales (nearly) infinitely and can continuously deliver the same result. Work in the physical world frequently requires site-specific solutions. It's more akin to consulting than engineering.

It's not surprising that startups in the space came in overly optimistic and imploded quickly. I think most tech people have a bias against the trades and think it's easily understood grunt work. The reality is that it requires deep subject matter expertise, familiarity with local markets, sites, and housing stock, and a healthy amount of on-the-fly creativity.

I was thinking about the automation when doing earthwork when I was in the field 20+ years ago. They tried some things. Like set up a laser to spin shoot across the site and the dozer blade had a reader that would tell the operator to go up or down. That apparently only works if the laser level was setup up correctly they learned.

https://en.wikipedia.org/wiki/Laser_level

A quick google search of "laser grading" shows its still a thing attaching to heavy equipment.

https://www.starequip.com/construction-equipment-dealership-... https://www.burchlandmfg.com/products/lgx-laser-grader/

Construction Companies are pretty competitive and always looking for an advantage. They have some interesting ideas. (Using radiation to measure soil compaction was my favorite).

https://www.virginiadot.org/VDOT/Business/asset_upload_file7... [pdf]

Laser leveling is very much a thing, I can't speak to grading for construction, but I have several family friends who use it extensively for agricultural fields in California.

These days it is done via GPS not laser. With expensive GPS units (RTK is the term to google) you can get to within a couple cm and that is good enough.

This is such a correctly and thoughtfully phrased insightful response.

> - most of the industry is decentralized small contractors. You can't go viral with construction, you can't build monopolies to extract rents later on.

At least around my area (in the PNW somewhere), and specifically for residential construction, the opposite happens. The builder has a bunch of small contractors he uses -- but they are only legally separate. And disposable, to be crude. They arrangement is a monopsony -- the plumber only has one customer: the builder. Same for the electricians, framers, etc. Even though they're not employed by the builder, they might as well be. Hell, even the realtor who sells their houses is an independent contractor with only a single customer. Guarantee she isn't making 2.5% per sale.

Interesting. The builders I've worked with used lots of contractors like that, but most were worked with more than one builder. Now the contractors did assign crews to builders, and so the same people only worked on houses for that one builder, but the other teams worked for other builders.

> almost none of it is automatable or in any way "disruptible" by a startup. Sure you can improve a bit by investing in automation or better digital tooling, but the ROI is in the single digits.

This is the angle I would attack if I wanted to get into a construction startup. Logistics & data. Single digit ROI might be worth going for if we can keep the rest of the costs under control. Maybe you aren't replacing Lennar 100%. Maybe you get into a B2B partnership with them and gradually go for a bigger play with them as your construction partner. How much would they pay you to reduce their opex by 1%?

If you are trying to revolutionize the art of construction and push the entire market towards some crazy idea involving off-site fabrication, I agree... Total waste of energy. All of the factors you describe point towards "work with an incumbent, don't fight them".

Indeed and there's lots of ways to make money there - but you won't get any easy funding. VCs aren't interested in businesses that won't offer a potential 100x (or better) ROI, and banks don't like funding R&D so your only chance is either being from "old money" yourself or knowing "old money" (aka someone who can afford to sink a dozen million euros as an angel investor).

Make sure you target the right people. Logistics is something the lumber yards handle, not the contractors. The contractor drives there van to the job, unloads, works all day, loads up, and drives home - that is all their logistics. The lumber yards drive to many different sites, and often do have partial loads that could be combined.

I don't think there is much scale you can pull off there though. They have been doing this job for a while and can figure it out.

I'm something of a construction-ist myself. Just kidding. I've gone through 3 major home renovations in different parts of California - clearly not an expert, but I'm confident in saying that the biggest obstacle to real estate construction is government regulation, not that there's a lot of it (most, not all, of the regulation makes total sense), but the process around how it's applied is so ass-backwards it makes my head spin.

I've done all my own construction plan reviews with cities/counties. It's mind-blowing how it all works. You sit down with a reviewer, who looks at your plans drafted by an architect. You go back-and-forth with the reviewer about the length of a rafter and the type of wood, grandfathering existing structure, etc. Then, he pages through big-ass books to find the latest codes for insulation R-values or whatever. And the process goes on and on for hours. It's painful how inefficient it is. One reviewer I sat with actually fell asleep during the review. Seriously. I asked him if he was okay and he said he was hungover because he went to the Raiders game the night before. And the funny thing is, he was the toughest reviewer I've ever had. Another lady didn't even look at my plans; she just talked for an hour about her daughter's college graduation, then "stamped" my plans.

I often wonder how people who do lots of large projects deal with this. I assume they just dedicate staff to it, which seems totally unnecessary. As a software/systems guy who's always striving for efficiency, I wonder why all these various "codes" can't just be plugged into a system, then the "plans" be evaluated against those codes, kinda like a series of unit tests. All these "codes" are documented in the big-ass books; it's not magic. Has anyone heard of companies trying to do something like this (plan approvals as unit tests)? Just curious.

However, I'm reminded that city governments have zero incentive to streamline the process. That was painfully obvious during the plan reviews I was involved in.

Most of what you state is specific to California. Most states don't have all that review, if it meets code you get your permit and build.

But how do you know if it meets code? Isn’t the review what determines it.

I am under the impression that I’ve read about something in Texas where they would outsource this class of approvals to people who were qualified and able to apply the same rules that the government would. Unfortunately, I can’t find it.

Some municipalities outsource it to Architecture & Engineering firms, for various reasons. It's not just Texas, I assume it happens all over, I'm in New York and have done some of this work.

This is why I never want to build a house. I'd seriously pay someone with superior interpersonal skills and construction knowledge to handle all this shit for me. Startup idea?

That's called a general contractor.

its an architect/designer who designs it, not the general contractor who builds it.

If your general contractor is picking wall assemblies you have skipped the planning stage, and you will have poor results because of this.

Is it? I asked a while back and there was a huge discussion as to whether GC does this or not: https://news.ycombinator.com/item?id=34409077. I don't know.

> However, I'm reminded that city governments have zero incentive to streamline the process.

I closed last year on my new home northwest of Houston. The suburb is one that's growing like crazy, but still is run like the small town it was once was. (We didn't get a water bill for some 3+ months due to a ransomware attack).

The final certificate of occupancy was potentially held up due to the one person who signs off on them, who wears multiple hats and apparently only processed them one day a week. If you got on his bad side (and this included trying to bribe him to hurry up), he'll put yours at the bottom of the stack.

Try systemizing and scaling your way around that.

>As someone who actually worked in construction (excavation in my case):

I dont think we could tackle construction as a whole. But if "Tech" ( Not just Software ) could help with all the smaller things. Like somehow make excavation 10-20x faster and cheaper.