Why Software Has a Hard Time Eating Construction
Today: Construction isn’t a standardized industry, Pascal Chazal, Katerra, China, travelling to Bordeaux.
The Agenda 👇
My past views on Google building houses
When The Family met Pascal Chazal
Turning construction into a real industry
Softbank-backed Katerra coming up short
Travelling away from Normandy
Most of what I know about software eating the world, I learned studying the music industry. I built up an entire framework to trace what’s happening in any industry that I modeled after what happened in the music industry (see The Five Stages of Denial). Before I designed that framework, however, I had to go through many different industries so as to spot common patterns. One that taught me a lot, on which I spent quite a bit of time back in 2012-2013, was construction.
I’m not really sure when and how it started, but back then I was interacting a lot with people in the construction industry (the likes of Bouygues, Vinci, etc.), and in early 2014 my attention was piqued by Google acquiring Nest.
The narrative was clear: Google had maxed out its data-collecting capacity coming from us looking at screens on our computing devices. It needed to collect even more data by branching out into the more tangible areas of our lives: that would mean watching TV, driving a car, and, yes, moving around in our home.
Nest seemed like Google’s gateway into the domestic sphere: the first foothold from which an entire home operating system could be deployed and operated: come for the thermostat, stay for the network that will help you manage energy, surveillance, appliances, etc. (Remember this was in 2014, so things like the Amazon Echo weren’t around yet.)
But Google had a problem: it had to convince customers one by one, and then help them find a specialist to install the device and connect it to the heating system. I was already conducting value chain analysis at the time, and so my view was that it was too risky for Google to wait for adoption down the stream. At some point, the logical end game would be Google partnering with construction companies so that you can buy a Google-branded home with Nest and other Google products already installed.
Incumbents in the construction industry have got the construction know-how, but the difference between building a house and managing its operating system is that in the first case you don’t have a real link to the occupant while in the second one you do. And in the battle to divy up added value, it’s always the person with the link to the end user who wins. Especially when that’s a giant like Google.
Google allies itself with a multitude of people, and with that alliance they can turn around on the value chain and say to the builder, look, if you want to build housing with our operating system, you’re going to have to give us a lower price. Then all the builders will be competing to figure out who’s going to be able to build these houses featuring Google’s operating system.
So Google’s arrival is a threat to major construction industry players, since Google could choose different players who aren’t part of today’s major groups to build their connected houses.
In fact, Google’s arrival is a threat to everyone who’s situated upstream in the construction value stream, since if all the value becomes concentrated downstream, then the upstream players will be left fighting over what’s left. That creates a phenomenon of scarcity and pressure on the margins. The Nest purchase is thus announcing a major battle for control of the construction value chain.
I basically left it at that, and Nest went on to not be the smashing success that some (including me) were envisioning. But some time later my cofounder Oussama and I were approached by a veteran of the construction industry named Pascal Chazal.
He had seen our videos and read our articles online and was wondering if there wasn’t an ambitious play to consider for building a full-stack operation in the construction industry—one that would control the whole value chain, from buying land and building a house on it to connecting the house to the Internet and operating value-adding embedded services, both online and offline. That echoed the parallel with Tesla I had considered in my Le Moniteur interview:
One day, Google will build houses because they’ll be tired of being constrained by the slowness and rigidity of upstream players. After contracting with Bouygues, Vinci and Eiffage, they’ll learn to build themselves. The giants of Silicon Valley have so much capital now that they aren’t afraid of losing money in learning how to do the things they don’t know yet.
Look at the designer and builder of electric cars, Tesla, which learned in a few years things that it took traditional auto manufacturers more than a century to learn. The goal is to no longer depend on builders who are going to slow you down. Maybe one day Google will buy up construction SMBs to launch a “Google Builders” profession. But there will be a phase where the major construction players will try to hobble the digitalization of the market through regulation, which could lead to France lagging behind in this field.
Pascal had more or less the same reasoning, and he had a secret, learned during his previous experiences as an entrepreneur in prefab construction. Here it is (like every good secret it’s long been out in the open):
Construction may be called an industry, but it’s merely a group of craftspeople who don’t apply any of the recipes that make an industry: there’s hardly any division of labor, and people in construction barely use standardized parts. And so you can’t build anything that resembles an industrial supply chain, which means you can’t pursue increasing returns to scale. That alone explains why productivity has lagged in construction, why building a house is still so damn expensive, and why workers are so badly paid.
On the division of labor: you can argue that there is some, but (at least in France) it’s been derived from regulations rather than from discovering what best generates increasing returns.
In particular, the division of labor between those who design and those who make provides a specific category of people (architects) with a lock over the entire value chain, and the end result is that standardization is impossible: culturally, an architect considers themselves to be more an artist than an engineer, and they tend to consider each house as unique.
Combine that with customers’ passion for customization and the fact that the rest of the value chain (those who actually build) is scattered across a vast field of small businesses, and you realize why the power in the industry resides with two constituencies: real estate developers (those who buy the land and sell the houses), and architects (those who design the houses).
Meanwhile, Pascal had been busy as a pioneer in wood frame prefab construction. His company Ossabois, which was later acquired by Bouygues, was focused on building bungalow houses for large customers such as Blackstone-backed Center Parcs and others. Indeed these were typically cases where industrialization made a difference:
Lots of small units, so it makes sense to standardize up the stream (considering the volume) and to prefabricate in a factory (considering the small size).
It’s for tourists on holiday, so you don’t have to deal with someone who wants their house to be unique/different and finds a willing ally in the architect.
Pascal thought that the unit economics of prefab could be extended across the entire industry and that it would turn the customer experience from a nightmare into an attractive bundle of services—all thanks to the increasing returns made possible by prefab.
However, after having tried various approaches, his feeling was that you had to build the whole value chain yourself—a full stack approach—, which is why he was looking for allies in the tech space. We started reflecting on a project named ‘Woodstack’ (a clever combination of ‘wood’ and ‘full-stack’ 😜). Alas it didn’t go far: Pascal had to take care of other parts of his business; we had a hard time bearing with the slow pace in the space while startups around us were racing forward like crazy.
In the end, the project faded away, but Pascal kept working on pushing his vision.
When he realized how far behind things were in turning construction into an industry, he decided to refocus on more basic things: evangelizing and education. First, he realized that we didn’t even have a proper name in French to describe prefab construction. ‘Préfabriqué’exists in construction, but it has a bad connotation: it’s something temporary, typically in plastic, that is set up so as to provide shelter while the main house is refurbished (the same kind of connotation that ‘mobile home’ has in the US).
Pascal finally settled on ‘hors-site’ (literally, ‘off-site’), and decided to push the concept toward a larger audience. He launched a magazine called Hors Site, which has been published since 2016, and related industry events. More recently, I saw (on LinkedIn) that the magazine has expanded into education, with a line of business called Campus Hors Site. It seems we’re still at the stage of building a startup community before we can change the industry as a whole.
However, there has been progress in other parts of the world. A few things came to mind when reflecting on the current state of things:
Bovet and his researcher found that developing a building system wasn’t the main challenge. The bigger problem they needed to solve was labor. There have been shortages in labor and skilled tradespeople in the homebuilding industry for years, as workers have fled construction jobs tied to the volatile housing market in the years since the great recession and shifted to higher-paying jobs in other sectors. More than 80% of builders have reported shortages of framing crews and carpenters, according to the National Association of Home Builders. Availability of labor remains builders’ top concern.
So Bovet began looking for ways to make a process-based industrial building system that could be assembled even without skilled craftspeople. “Because you spend three or four years to plan your project, you get some architects, you’ve bought the land, you’ve got your financing, and now 80% of builders cannot get carpenters to put it up,” he says. “You don’t have the labor. And even more so now, with COVID. So what do you do? You scrap your project? Or you try to find a platform. Well, the platform is what we offer.”
Meanwhile, there’s a US-based company called Katerra that has raised more than $2B, including $1B from SoftBank. Basically Katerra echoes the idea of combining offsite construction with technology so as to deliver affordable quality at scale in construction. Unlike what we and Pascal had in mind 6 years ago, it seems to operate more as a real-estate developer than as a service provider focused on the end user experience. Also, like most SoftBank-backed businesses, it’s not going well: there were changes in top management in the Spring after news of financial and operational difficulties as early as last year (meaning before the pandemic). Have a look at Warped lumber, failed projects: TRD investigates Katerra, SoftBank's $4B construction startup.
Finally, go further with the Boston Consulting Group’s The Offsite Revolution in Construction (2019).
I’m currently in Amboise (Touraine) with my son Ferdinand. Since there’s a 2-week holiday (what is known in France as vacances de la Toussaint, or All-Saints Holiday), we’re on our way to Bordeaux, where my father lives, after which we’ll head back up to Le Havre to visit my mother. And that means that for the first time in a long time, some of the next editions will be sent from places other than Normandy!
Meanwhile, Laetitia and Béatrice are in Munich, criss-crossing the city by bike to spot the best neighborhoods, and they will start visiting houses in the coming days.
If you’ve been forwarded this paid edition of European Straits, you should subscribe so as not to miss the next ones.
From Amboise, France 🇫🇷