Why I don't love Light Rail Transit

Snippets 2, Episode 22

One problem I think about a lot (as a clear non-expert) is how land use and transportation patterns are evolving into a new kind of inequality wedge in North American cities. There’s no doubt that many cities here in the United States and Canada are enjoying an urban renaissance: industrial and waterfront areas are getting revitalized; downtown is a cool place to live again; walking or biking to work is a desirable goal.

Meanwhile, there’s this chasm of inequality of opportunity that’s opening wider each year, especially in these trendy tier 1 cities that feel like they're arranging themselves into economic caste systems. Commuting has something important to do with it.

In the 60s and 70s, remember, cities were seen as undesirable. The goal was to live in the suburbs; farther away was more desirable than closer in. As I’ve talked about before, this spatial arrangement had a useful side effect: commuting distance was a part of the cost you had to pay to live the good life. But now, we’re going through this big inversion where downtown is cool again, and we’re losing the self-corrective mechanism we used to have. 

Now if you can’t afford to live downtown, you get forced out towards cheaper real estate, and you’re the one who has to pay the cost of commuting distance. What used to be a negative feedback cycle between commuting and residential desirability has flipped into a positive feedback cycle. (European cities have looked like this for a long time. The really poor areas have been out in the forgotten suburbs for a century.)

This positive feedback cycle, I suspect, is acting as a real economic inequality wedge. It used to be that anyone could move to a city and increase their earning power; nowadays, the economic advantage of cities is only really available to people on the top half of the income spectrum. Most of the discussion around this phenomenon is around specialization of jobs and skills, but I think there’s another factor increasingly at play here: if you can’t afford to live in the attractive part of cities, you’re going to have to pay a “commute tax” that’s going to be a real drag on your ability to prosper and build a good life. High earners consciously pay a premium to avoid this.

Meanwhile, public transit has become popular again. I say “popular” not because public transit ridership is increasing (in many cities, it isn’t) but rather because it’s become a socially virtuous, progressive thing to care about. In the past couple decades there’s been a lot of interest among city planners, the downtown crowd, and the “urban rejuvanists” around a new public transit form factor: the LRT, or Light Rail Transit. 

LRT lines are propping up all over the place, and you’ve probably taken some: cities like Denver and Minneapolis have invested a lot into building out LRT networks, and even Phoenix – the ultimate land of the car – built a line recently that’s been quite well received. They’ve come to represent a particular kind of “popular environmental urbanism”: a mix of eco-mindedness and sophistication. Light rail is one of those things you’re just supposed to embrace; or else turn in your Progressive Citizen credentials and admit you’re an eco-terrorist.

But not everyone is excited about them. The Koch brothers, of all people, funded an anti-LRT campaign in Phoenix, and cities like Nashville have voted down initiatives to put in light rail lines. And you know what? I have some issues of my own. 

My worry is that LRT lines, in many cases, are a good example of what progressive people think lower-income people want. But their impact is the exact opposite: they gentrify public transit into something that magnifies inequality of opportunity rather than equalizes it. In practice, they’re a luxury option for rich people, rather than something for everyone. 
How so? Before we get into what’s going wrong, we need to first get anchored about what a successful public transit system looks like in a North American context. 

A lot of North American public transit advocacy can feel like a perpetual lament that North American cities aren’t laid out like European ones. So it’s worth looking at an example of a North American city that’s laid out pretty typically but where the public transit system still works well: Toronto. Although it’s currently going through a huge population and density growth spurt, for the better part of the 20th century Toronto was an average-sized, sprawling, suburban, mid-continental city that didn’t really stand out in any way. It’s a city that’s generally organized around the car.

However, if you look at Toronto’s transit system (the TTC) and how many people use it, it doesn’t look like a classic sprawl city anymore. The subway network doesn’t look all that impressive at first glance, as there are only two “real” subway lines (with two more branch lines on the outskirts). But it has huge ridership, for North American standards. Total ridership is behind only NYC and Mexico City. No other city except Montreal comes close; which makes sense, given Montreal is much more European in its layout and mindset around public transit. 

Where are all these riders coming from? Subways are great for moving large numbers of people from one place to another, but they’re a lot less useful if you don’t live and work within walking distance from stops. In a dense city like NYC or Montreal, that’s not a problem, since there’s a critical mass of people living in any given square mile to justify digging an expensive underground subway stop there. But in a low-density, sprawling city like the rest of North America, you can’t easily justify the cost of extending a subway through low-rise neighbourhoods of single family homes. 

However, as the subway network was getting built out in the 60s and 70s, the regional Toronto Metro government (whose influential constituents were largely suburban homeowners) got upset that the developing suburbs were left out of the downtown network, and demanded good transit coverage. And they got it: an extensive network of frequent, reliable bus service that covers essentially every major street in the city with 10-minute or better service, all day. 

Buses are a good form-factor for serving North American cities outside of their downtown cores, it turns out. They’re flexible, inexpensive, and when properly funded and managed, can be a really solid transit option that genuinely competes with car ownership for getting downtown, or to the subway. They’re not fancy, but they get the job done. It turns out when people say they don’t like taking the bus, what they really don’t like isn’t the bus itself, but mostly bus schedules that are unreliable or infrequent. High frequency, prioritized bus service is actually quite popular when done right. 

The extensive bus coverage lets transit planners squeeze every inch out of the subway, too. Practically every subway stop in Toronto outside of the downtown core primarily serves customers who do not walk to the station, but instead take a bus for their last mile. So suburban subway stops still get real ridership, because they can attractively serve a large catchment area. A recent Citylab piece used the York Mills subway stop as an example. It’s next to a golf course, in a low-density neighbourhood. It should be a terrible spot for a station. But the extensive bus connections feeding into it draw in well over 10,000 riders per day, which is more than most subway stops in Brooklyn and even some in Manhattan. 

The buses matter. Well-run bus routes can cover a lot of territory, and they can move a lot of people so long as there’s somewhere sensible to take them. Virtually everybody in Toronto lives within walking distance of a solid bus or streetcar route: not just people in gentrified areas, but also everybody in the lower income northwestern and northeastern shoulders of the city that downtown people rarely visit.

Remarkably, out of the 23 surface transit routes in North America that move 30,000+ people per day, 12 of them are in Toronto. Even if you exclude the downtown streetcars, which operate basically like larger buses on rails, there are 26 distinct bus routes in Toronto that move more than 15,000 people a day each. (And another 59 routes that move more than 5,000 a day, which would be a hugely popular bus route in almost any other city.) 

Anyway I’m going through all of this to say two things. First, you can move a lot of people by bus if you need to; the TTC is a good demonstration that large numbers of people will take bus routes for their regular commute if the network is well-run and gets them to where they need to go. Second of all, unless your city is really dense or you have infinite money to spend, any public transit system that wants to actually provide coverage across the entire city is going to need a good bus network.

Fortunately, it’s relatively easy to set up a bus route that works well. You need buses, drivers, and some paint to make dedicated lanes and transit priority intersections. The cost just isn’t on the same magnitude as expropriating land, ripping it up, putting in rails, and then running trains all day. 

So this brings us to LRTs and their recent popularity. “If we want more people to take public transit, we have to make public transit nicer!” goes the argument. I’ll happily admit: LRTs are very nice! They’re smooth, comfortable, and definitely a superior rider experience over the bus. They get credit for helping to spur rejuvenation (read: gentrification) of main streets and central areas. 

But here’s the thing: they’re not actually moving that many people. LRT lines are getting daily ridership in the tens of thousands, which sounds great, until you remember that a well-designed bus route can easily carry that many people. In North America, compare a sampling of popular LRT systems’ average ridership per line to the more popular TTC bus routes:

(A few notes on methodology here. First of all, for people who know the TTC, I separated out the 504A and 504B so that they’d fit on the graph; otherwise it’d break the y axis. Second, for the other North American LRT and surface rail systems I did a highly quick and imprecise method of taking their total number of daily boardings divided by their total number of routes (in brackets); this is obviously not a very precise way to graph them but it does give a general sense of ridership loads being handled. The point here is that these LRT lines are handling passenger loads that a well-designed bus route could handle no problem.) 

These LRT lines, as nice as they may be, are not really being designed or used as building blocks for true higher-order transit systems. They’re more like… a luxury bus. And I don’t want it to sound like I don’t like the idea of a nicer bus! But we should be clear that the real civic purpose of these LRTs isn’t building out the kind of high-capacity backbone trunk routes that you need in an actual at-scale public transit system. Nor are they doing the job a bus network does of crisscrossing the entire city and providing continuous service to every resident. What an LRT line accomplishes is nice, clean, lovely service along the length of the line – but that’s it. 

Unfortunately, these LRT lines soak up a lot of money from cities’ and transit agencies’ capital and operating budgets. This is money that could be far better spent, on a per-rider basis, on operating good bus routes; but instead is going to be spent on beautiful, brand new lines that only serve a relatively small catchment area unless they’re already integrated into a dense transit network (which is rarely the case).

The area around the LRT lines definitely attract investment, but if you look at who actually uses the line a few years in, it’s mostly rich people. Why? Because they’re the only people who can afford to take it – not because the fares are too high, but because real estate in the immediate walking area around stations becomes too expensive. 

There’s an argument you hear a lot: “The ROI on these lines are great; just look at the increase in property values that follows!” Yeah, exactly! Spending money on LRT lines means spending money in order to make public transit less accessible, not more. For public transit to be accessible to you, it needs to be a real option. Unless you have bus routes or other last-mile ways of getting to the LRT, then it’s going to be a public transit option that’s only available to people who can afford to live nearby. And the nicer you make the line, the higher an income threshold that’s going to require. (Unless you do the hard work of actually integrating the LRT line via last-mile bus routes into all of the other neighbourhoods that aren’t gentrifying.) LRT investment on its own doesn’t expand public transit; it gentrifies it. 

What really pisses me off about this is that the biggest proponents of these light rail lines are usually progressive, urban, “forward-thinking” brunch people who can be quick to lecture you about driving or using too many plastic bags or whatever. They love the idea of public transit – except when it’s a bus, of course; you’d never catch them dead riding the city bus.

The LRT lines they advocate, in practice, are essentially a highjacking of the public transit system: reprioritizing funds away from buses and transit networks that actually serve a diverse ridership, and into these Luxury Rapid Transit lines that look nice and pretty and socially progressive, but actually shut out public transit to anyone who can’t afford access to it, forcing them back into cars and into traffic. 

It gets worse, actually, if you look at cities like Phoenix who build LRT lines that initially connect affluent communities and then try to expand lines into lower-income areas. Business associations and homeowner associations will stop at nothing to block that kind of progress – “it brings the wrong kind of people into our neighbourhoods”, yeah mmhmm. We’ve heard that one before.

The proof is honestly in the ridership numbers. If an LRT line is carrying a daily load of passengers that could be handled by a bus route (as nearly all of them could be, and are, here in Toronto) then its purpose for existing isn’t “higher order transit”, it’s “luxury public transit”. Again, spending money on public transit infrastructure is something I’m generally going to agree is an attractive idea. But a lot of these LRT lines are being built for the wrong reasons, and they’re having the wrong kind of impact if you care about having a public transit system that actually serves regular people. 

The broader lesson here is that a lot of our progressive-minded civic actions and priorities, which seem to locally be something that helps out everybody, are actually having the opposite of the intended effect in the long run. In theory, an LRT system is something that anybody can use; in practice, they start to become virtue toys for rich progressive people. When you look at the LRT itself, it looks great! It looks like everything we’re supposed to be doing. It’s a public transit investment, it gets people out of their cars, it in theory ought to help increase interaction between people they wouldn’t otherwise encounter. 

But in practice? If you’re taking public transit, look around, and notice that it’s all rich people – well, you might be witnessing the gentrification and luxur-ification of a public good in real time. The current urban renaissance is amazing, and I’m certainly a beneficiary, but I hate to see it become this “cities as luxury goods” trend. If you ask people what they want and they say “I wish the bus came more frequently”, sometimes the best thing to do is take them at face value, rather than counter with, “Oh, what you actually want is this beautiful expensive train.” 

Even here, with our historically humble but high-functioning transit system, we’ve gotten excited about new theoretical expansion while letting our existing networks burst at the seams under wear and overuse. (At least the crosstown LRT being built is projected to handle 150,000+ people per day, which is a much more real number for higher-order transit, at 50% more projected use than the entire 6-line Denver LRT network.) 

The best news in public transit here isn’t in Toronto; it’s actually in the suburbs. The city of Brampton, west of Toronto and home to a half million people in its own right (it’s in fact the second fastest growing city in Canada) has seen their own public bus network growing at double digit percentages, serving over 100,000 riders daily – despite the fact that it’s as pure car country as you can possibly get. Other cities in the greater region, like Mississauga, are seeing starting to see similarly impressive ridership gains after decades of car dominance. It’s simple, it works, and it’s popular. Who said public transit couldn’t work in sprawl city? Maybe the key is to build stuff that actually reaches people where they are, and not just vanity transit gentrification projects. But what do I know. 

Permalink to this post is here:

Why I don’t love light rail transit | alexdanco.com

A quick note here - if you liked last week’s post on Positional Scarcity, I bet you’ll like the book I’m writing! It’s called Scarcity in the Software Century. I’m currently writing it week by week (albeit a bit spottily over the last couple months as we adapt to life with a newborn, but I’m getting back into the swing of things), and you can support the project by subscribing to it here. This past week, we looked at another kind of emergent scarcity, Integral Scarcity, that naturally and inevitably emerges with abundance:

If positional scarcity was “in conditions of abundance, your place in line becomes scarce, and line-management is great business”, integral scarcity covers a whole other set of issues: “abundance leads to complexity, complexity leads to problems, and those problems are great businesses.” If you’re interested in learning more, I’d love to have you as a subscriber!

Here is a really interesting finding from the biology world:

Off-target toxicity is a common mechanism of action of cancer drugs undergoing clinical trial | Ann Lin et al., Science Translational Medicine

The short version of what this study is and why they did it (as explained by the senior author on Twitter; and while we’re at it, WOW is Twitter great for research! I can’t imagine how things might have been different if academic science twitter was a thing when I was in grad school. Ah well…): 

So, you probably know that most big clinical trials for new drugs fail. It’s especially bad for cancer drugs, where 97% of all hypothesized drug treatments, which were all backed by lots of basic science and had a plausible mechanism for action, do not bear fruit when tested for real medicine. Our understanding of cancer is mostly organized around genes associated with cancer, which we call oncogenes. A certain gene might code for a protein that, when expressed, tells the cell to start dividing and never stop; that’s a cancer gene. So we try to go after those genes with drugs that, through some logical biochemical mechanism, disrupt the cancer process. Hopefully all this makes sense. 

The authors of the study looked at this 97% failure rate and thought, gee, that’s a really high failure rate. Is it really because we’re so bad at drug development? Or could there be some other explanation: like, maybe a lot of those genes associated with cancer actually have nothing to do with cancer? So they used CRISPR, everyone outside of bio’s new favourite bio trend, to systematically go through and re-test all of those drugs in cancer cells, but with the genes that they were supposed to be targeting knocked out. In the majority of cases, knocking out the gene didn’t change anything! Uh oh! So all of these drugs were still doing something; just not what we thought they were doing this whole time. 

The study is so great because it’s so simple - “hey, let’s use modern tools to replicate all of this prior work, just testing some fundamental assumptions carefully” - and yet, these kinds of replication studies aren’t done nearly enough. Incidentally, by the way, the researchers stumbled upon a new, interesting mechanism of cancer-fighting that they’re now pursuing. If that’s not good science karma justice, I don’t know what is. 

Speaking of science, here’s a classic paper from a few years ago: 

One in five genetics papers contains errors thanks to Microsoft Excel | Jessica Boddy, Science

Apparently Microsoft Excel is particularly prone to misinterpreting and re-ordering lists of genes, because it interprets certain sequences of numbers as calendar dates. You can just hear the palms hitting faces in labs everywhere. 

The more educated you are, the more closely your opinions conform to one political party | via Tom Wood on Twitter

The fight for seed | Kate Clark, Techcrunch

The streaming wars: its models, surprises, and remaining opportunities | Matthew Ball, Redef

Great news from a Tesla-affiliated lab on the battery front:

A wide range of testing results on an excellent lithium-ion cell chemistry to be used as benchmarks for new battery technologies | Jeff Dahn

And just for fun: an oral history of February 26, 2015, one of the greatest days in Internet history: it started with the runaway llamas, and ended with The Dress. 

2/26: the day everyone came online for the internet’s perfect storm | Charlie Warzel, Buzzfeed

Have a great week,


Positional Scarcity

Snippets 2, Episode 21

Each day on Tech Twitter, we get up in the morning, open up the website, and then go see what it is we’re mad about. A few days ago, it was this:

The concept of “pay to get a better place in line” when it comes to advertising or product placement is hardly new. Walk into your local drugstore you can see it literally on display: CPG companies like P&G spend good money to make sure their products appear in good shelf position, like eye-height and in prominent display slots, rather than inferior ones. A more modern example you’ll also know well is Amazon, where merchants can pay to appear higher in search rankings, or gain badges like “Amazon’s Choice” that indicate a premium place in line, and accordingly boosted sales. Curiously, in the latter case we call this expense “advertising spend” whereas in the former we call it “marketing spend”, but it’s really all the same thing: in conditions of abundance, relative position matters a great deal. 

In general, there’s a name for this phenomenon: positional scarcity. Positional scarcity comes in a lot of different flavours. There’s curation: an abundance of people creating music creates demand for record labels, DJs, and other tastemakers to do the job of picking, “Out of all these options, which song should get heard?” There’s prestige: an abundance of prosperity makes it harder to preserve and distinguish high status: “Now that everyone has a car, how do I make sure that my car is the most impressive, and everyone knows it?” There’s access: an abundance of people all vying for each others’ attention and patronage creates a high premium for moving to the front of that line: “How much would you pay to skip the congestion and get to where you need to go?” 

The most interesting forms of positional scarcity combine several of these elements, with the ultimate positional scarcity business model at the middle of our Venn diagram: an interesting industry called the Loyalty Business.

What these challenges all have in common is that they’re symptoms of abundance. In environments of real scarcity, these problems don’t exist. But in environments of abundance, when some new technology or paradigm has created a huge bounty and variety of new stuff that we enjoy, positional scarcity inevitably emerges, creating new bottlenecks and new opportunities.

In my view, positional scarcity is one of the two primary kinds of emergent scarcity that we always, inevitably get in conditions of abundance; the other being a concept I call Integral Scarcity that I’ll write about some other week (or you can read about in my ongoing book project, Scarcity in the Software Century). It’s a direct consequence of abundance: with abundance comes not only great consumption but also great variety, and that variety and velocity of consumption creates several new kinds of problems that all get rolled into this general concept of positional scarcity.

In this post I’m going go try and categorize some different kinds of positional scarcity, with some special attention to all of the in-between categories.

Let’s start up top with Curation, Prestige, and their overlap: Legitimacy

Curation is a form of positional scarcity that we get whenever there’s an abundance of choice of something, and the average consumer needs help finding what they’d specifically like. No one can listen to every song ever made, let alone every song made this month. So record labels, radio DJs, and playlist curators do the job of sifting through it all and finding a few good ones we’ll probably like. Sometimes the problem is sheer volume, like with music; other times the consumer wants an expert to guide them through a consequential choice. When there are 100 different cars on the market and you aren’t a car expert, you’d like somebody to go through every car on the market and tell you which ones are safe, and which ones are good value. So we have curators like Consumer Reports, or someone like Gartner in the business world, to help consumers decide: “out of all these options, which one is the right one?” The job of being a curator becomes more important as abundance and variety increase.

Prestige, or alternately Status, is a different kind of positional scarcity. Standing out, differentiating yourself and wielding status are age-old needs that people have, and they get more challenging in conditions of abundance. Status and prestige emerge as scarce whenever some form of abundance creates a rising tide that lifts all boats, so to speak, and so the handful of people who used to stand out by owning boats are now compelled to make sure you still know that their boats are the best ones. 

In the early days of automobiles, owning a car as a recreational vehicle in itself was a form of prestige. But nowadays cars are ubiquitous; if you want to stand out, you need a nice car; which is why we pay a premium for a Porsche hood ornament even if it’s slapped on a car that’s basically a VW. Sometimes tiny status signifiers gain huge cultural relevance, like the distinction between blue versus green iMessage bubble that has probably singlehandedly done more to protect iOS marketshare than any other feature among the coveted 25-and-under crowd. 

Curation and Prestige are both pretty straightforward, but their intersection is more interesting. I call it legitimacy. In an environment of abundance that’s full of complexity and choices, there’s a really interesting convergence between the ability to curate and the power to wield prestige. 

Angel investing is an exercise where rich people spend their money (and occasionally see a return) in order to be able to say, “Look what I curated!”, and earn prestige in doing so. If you were an early investor in Stripe or something, that’s a huge status symbol - not just because it means you’re rich, but because it means you demonstrated excellent curation ability, and that’s a high-status thing to have. Fashion is similar: it’s an opportunity to earn and preserve status by demonstrating your excellent taste and curation skills. In both cases, successful curation confers prestige, and people seeking prestige will go play at being curators in order to try and earn some. 

The third entry in this segment, “McKinsey ass-covering”, is the inverse of fashion or angel investing but still the same basic thing. When a company faces the dilemma of an abundance of choice, its executives (who don’t want to get blamed should their choice turn out poorly) often seek out a consultant, who acts as a curator - someone who can help them sort through, “out of all these options, which one should I pick?” The curators that they hire - consultants like McKinsey or BCG - use the prestige of their reputations to sell their clients a form of reassurance, or legitimacy, that “you know you’ve made the right choice, because you can say that we advised you.” McKinsey sells a little bit of their prestige to the client, and in doing so solves their curation problem by making any option they choose feel like the best one.

Next let’s rotate around our Venn diagram at the next overlap: Prestige, Access, and the segment in the middle I call Proximity

Unlike Curation and Prestige, which are both subjective problems, Access is more of an objective, literal challenge. Abundance is great, but it creates congestion: when everyone has something, or everyone’s selling something, or everything’s available somewhere, it can become hard to reach whoever you need to reach. One literal example of an access problem is traffic. When one person has a car, they can drive quickly from the suburbs to downtown, or wherever they need to go. But when everyone has a car, you’re stuck in traffic. It’s too congested. So it may make sense to pay a premium for a tolled road with less traffic, so that you can drive around the congestion and get there faster. If cars were scarce this wouldn’t be an issue; but now everyone drives a car, access becomes everybody’s problem. (Just wait until driverless cars eventually turn traffic routing into a pay-to-play problem, where virtual “premium lanes” zoom rich drivers past everybody else, like in airport security, while everyone else fumes. If you think road rage is bad now...)

Warren Buffett loves to talk about businesses that are like drawbridges: if there’s only one bridge crossing the river in town, and everyone wants to cross it all the time, then you own something really valuable. You can charge an attractive profit in order for people to access the other side; the more people and things and abundance on both sides of the bridge, the more scarce and valuable your bridge and the access it confers. One of Buffett’s classic “drawbridge businesses” was owning the local newspaper. If a newspaper has all the local readership, then advertisers who want to reach those readers need to pay the newspaper an attractive fee in order to make it in the pages - and then another fee if they want to be in colour, or on a good page in a big size, or generally to stand out and be seen. The more abundant the readership and the advertising base, the more scarce the newspaper’s access becomes. (Until they get disrupted by Google, who built a newer, bigger bridge.) 

Then there’s the intersection between Access and Prestige, which I call Proximity. The purest example of this overlap is in residential real estate: how much are you willing to pay to live in a good neighbourhood versus a bad one? A lot, probably. When it comes to deciding where to live, people will pay a premium in order to live next to people who will also pay a premium. The willingness and ability to pay for what is essentially a residential luxury good (a house in a premium neighbourhood, which might be exactly the same quality of physical house, but will sell for twice or even three times as much) acts as a filter to make sure that rich people have easy access to one other, but other people don’t. 

If you want into the world of easy access to the prestigious people and their lifestyle - living on their street, sending your kids to the same school district - you’re going to have to pay to get in. Same kinds of forces are at work in elite gatherings of various forms: country clubs, fancy events, and other such institutions make sure that those with status can gain access to one another, but those without such status cannot. In conditions of mounting abundance and congestion, proximity becomes increasingly scarce.

Rotating around once more to finish out the Venn diagram, we have the overlap between Access and Curation which I’ve provocatively, but deliberately, named Extortion

The conditions for potential extortion arise whenever you have a drawbridge-type industry (e.g. lots of vendors trying to reach lots of consumers and have to go across a bridge of some sort to do so) where the consumers are relying on the bridge operator to make curative decisions for them. One familiar example today is the revolution taking place in online food delivery. If you have a pool of restaurants in the app and a pool of consumers who use the app but do so by explicitly searching for the specific restaurant they want every time, the app is limited to charging a reasonable take rate, because the restaurant is retaining some pricing power (the customer wants specifically them.) It’s an access problem, but it’s reasonable. The extortion problem arises when the app starts explicitly curating options for the consumer, and forcing restaurants to make more and more concessions just to be able to “make the algorithm happy” and reach their own patrons. 

When done reasonably, these kind of curative drawbridge businesses can do something useful for both businesses and their patrons, and we don’t complain about them. But think about the past couple years of drama between Facebook and publishers. If web publishers want to reach an their audience, Facebook is where they are; so you have to do whatever it takes for the Facebook feed algorithm to “curate you in”, so to speak, to your users’ feeds. So whenever Facebook says “Jump”, you have no choice but to ask how high. 

This brings us back to Google and this week’s news. In the early days of Google, we thought of them as a search engine that returned a list of relevant links plus a few ads at the top, and that was great. They were a great aggregator of users and could command high fees from advertisers, but it was fine: they were a pretty neutral access provider so it was all good. But recently, Google has started to use their power as the effectively sole search provider in a few really extortive ways. First, the Basecamp situation above: by making a design choice to stuff so many ads above organic searches for a company’s own name, businesses like Basecamp are effectively forced to pay up whatever Google wants in order to access search traffic for customers searching for them directly. That’s pretty extortive. 

Less than half of Google Searches now result in a click | Rand Fishkin

Meanwhile, more and more Google searches aren’t even turning into clicks at all. As Google gets smarter and learns how to simply answer users’ questions directly, they become a far more powerful type of curator than they used to be: they now have the discretion of whether or not to show users your link at all, because they can increasingly answer the user’s query without one. That’s the power of combining an access business with a curation business. Users at this point blindly trust Google to give them what they’re looking for. So long as Google retains its effective monopoly on search traffic (and controls the Chrome browser from which many of these queries, a lot of which are simply customers typing in the business name rather than adding “.com” at the end or something), they own something even better than a drawbridge. And we’re increasingly looking for an alternative.

Finally, in the middle we have a really interesting overlap between all of these factors. The pinnacle of “in conditions of abundance, relative position is scarce” is something called the loyalty business. In a world of abundant choice, one of the most valuable things you can have is an attractive, repeat customer that is loyal to you. So many companies, like department stores and gas stations but most famously credit cards and airlines, devote a huge amount of their time and resources to the business of loyalty itself. 

As a former ultra-frequent flyer, I can tell you from experience: the Air Canada loyalty program is … really nice! It solves the problem of abundance and positional scarcity for a whole bunch of different people and different angles. In exchange for committing to one airline (and granting them favourable position to me), they then give me back a whole bunch of different kinds of positional scarcity: faster lines, better seats, the ridiculously named but nonetheless coveted “Super Elite” status moniker. Another Loyalty Business that I use frequently is Amex, whose charge card comes with all sorts of preferential access and curation services - along with that really satisfying metal “thunk” gives you that dopamine shot of feeling prestigious for really no good reason. Amex is a fantastic “positional scarcity manager”, if you could call it that. It wouldn’t be too much of a stretch for other programs like Amazon Prime to evolve beyond being a fantastic bundle and into explicit curation and status. Some service like Prime Premium or whatever that gives you the whole prime bundle plus preferential access, curation and status everywhere? Seems appealing, although not necessarily Amazon-y. 

All in all, real positional scarcity businesses take time to develop, especially loyalty businesses that are difficult to get off the ground before a retail or consumption environment is fully understood. But I’ll leave with this thought: as the mobile internet, e-commerce, and the transformation of retail continues onward, I think there’s an incredible opportunity to build a loyalty business in “everyday internet-powered consumption”: taking Ubers, ordering anything powered by Shopify, that kind of thing. Uber itself is certainly trying to build one, with branded credit cards and other attempt to build loyalty programs - I’m not sure how they’re doing, but it’s still super early. Anywhere you encounter a situation that makes you think: “all of this abundance is making my place in line feel scarce”, there’s opportunity for creative business models in there for sure. 

Permalink to this post is here: 

Positional Scarcity | alexdanco.com

Debunking the silly “passive is a bubble” myth | Ben Carlson I liked this little piece, mostly because it offers something I’ve long sought-out: a short and simple explanation of passive investing primarily as a consumer innovation, rather than as a market trend. The analogy basically goes; for most of us, when you go to the grocery store to buy bananas, do you go to the counter and submit a bid for bananas? Do you try to do price discovery? No! You just pay whatever they’re charging for bananas, and save the hassle and expense of trying to negotiate a couple pennies. Ordinary consumers don’t have the knowledge or ability to actually negotiate for better banana prices anyway; it’s really nice that we have an easy option to buy bananas at a price that’s pretty fair, really quickly. Index funds are like that. The argument that passive indexing is destroying price discovery is just silly: the fact that consumers just accept the sticker price for bananas doesn’t destroy the ability to find a fair price; nor does a high percentage of passive investors undermine the market’s ability to do price discovery or allocate capital fairly. If anything, it should make active management easier! That’s why I don’t have too much patience for the “passive is a bubble” crowd: if you really believe this, and if you’re actually right, then go make a pile of money active managing and keep quiet about it if it’s so obvious! 

On data, unicorns and granola bars | Neil Constable

Investors and operators: lessons I’ve learned from both worlds | Brent Beshore

The Magic Moment: when Eminem met Dr. Dre | David Perell

Information gerrymandering in social networks skews collective decision-making | Carl T Bergstrom & Joseph Bak-Coleman, Nature

Have a great week,


Misunderstanding "Skin in the Game"

Snippets 2, Episode 20

As the dog days of summer stretch on, I’ve been doing a lot of trail running to keep active. The east end of Toronto is home to a spectacular and unsanctioned, semi-secret network of mountain biking trails that line the river valley ravines. There are over 40 miles of trails maintained by the mountain biking community, and who graciously tolerate runners invading their secret world so long as we follow good trail running practices. 

While it’s not the wilderness (you’re never more than a couple miles from civilization), it can often feel like you’re completely and utterly lost in the forest. The complete isolation you experience in these trails comes courtesy of the the fact that Toronto is effectively San Francisco’s topographic inverse. Instead of hills punctuating the view that can see and be seen from anywhere, Toronto has an extensive network of deep, forested ravines that crisscross the city. They’re virtually invisible aside from the bridges that cross them, and when you’re down inside them, you completely forget you’re in the city. 

Trail running feels like a completely different sport from road running. Once you start running on more challenging trails, you have to perpetually leap, dodge, and dart around branches, rocks, culverts, and all kinds of natural obstacles that threaten you at every turn. You have to dedicate 100% of your attention to not wiping out, 100% of the time. If you zone out even for a minute, you’re going to get hurt. Oddly enough, this kind of intense concentration typically leaves me a lot more mentally refreshed than zoning out for a long road run does. It comes at a cost of occasional skinned knees and twisted ankles; and potentially worse, if you’re not careful. 

Anyway the reason why I’m telling you all this is because I was running with somebody the other day and the concept of “skin in the game” came up, initially in reference to trail running. Compared with regular road running, trail running does indeed feel like it has more SitG: if you stop paying attention, you get removed from the trail - or even the running pool. But the conversation turned to Nassim Taleb, author of books such as The Black Swan, Antifragile and most recently Skin in the Game: Hidden Asymmetries in Daily Life. I think some of you probably know him best from Twitter, at this point; he’s a perfectly unrepentant asshole online (which suits his style and message perfectly), and there’s a bit of a general consensus among people in his Twitter sphere that he’s kind of gone off the deep end over the past couple years, which is a shame. But recent antics nonetheless, his writing continues to resonate. 

The core, unifying theme across all of Taleb's writing - from his early work around randomness and The Black Swan through later works like Antifragility and Skin in the Game - is the logic of risk-taking. Taleb takes great offence at creating work, exercise, or any other kind of activity that’s designed to consciously offload risk from the participant; he’s particularly fond of complaining about elliptical machines at the gym, jogging, or really any kind of exercise that isn’t free weight deadlifting. But I have a distinct feeling that he would be willing to make an exception for trail running, which seems like it’d be a lot more in his ballpark. 

In this vein, Taleb consistently advocates the idea that people generally shouldn’t be put in charge of anything important unless they’re willing to bear the consequences of what happens. Wielding his go-to phrase, "Skin in the Game”, you can often find him separating people into two categories: “Real experts” (e.g. airplane pilots, business owners) who face real consequences if they’re wrong, versus “fake experts” who may not (he particularly likes to pick on economists and academics). 

However, the idea that Skin in the Game leads to good performance and desirable outcomes has been co-opted by the tech industry in a way that’s quite different than what Taleb meant. In tech, we talk about “Skin in the game” a lot, but usually we talk about it in reference to incentives. We say that an employee with stock options might have more skin in the game than an employee paid strictly on a salary. Because a portion of their upside comp is in company shares, they have an added incentive to work hard. Many of these same people take a logical step forward with this thought: Equity-based compensation must be part of the reason why people here work hard; good motivation leads to good performance. In fields that don’t have upside-oriented comp (as in, just about every other field, except finance) then people simply “don’t have the right incentives”. Ah. 

Upside incentive can matter a great deal, but it has almost nothing to do with skin in the game the way that Taleb uses the phrase. When Taleb talks about “Skin in the game”, he’s talking about survival. A small business owner does not have SitG because she’s incentivized to run her business well; she has SitG because if she doesn’t run her business well, then it doesn’t stick around. Over time, if she has survived, then she’s probably good at running her business. 

Skin in the Game isn’t an incentive; it’s a filter. In systems that require SitG, bad performers are eliminated from the system, leaving behind the good ones. Another way of saying this is that people don’t have skin in the game; systems do. Systems with skin in the game are ones where the good performers are more likely to survive than poor performers; it has nothing to do with how people are incentivized. It’s Buffett’s line: In order to win, you must first survive

But in tech we’ve rebranded Skin in the Game to something that suits our narrative better. We have this narrative that upside is a great driver of progress: building an incredible future is simply a question of correctly incentivizing everyone. There’s some irony here. First of all, one of the big reasons that Silicon Valley works is that it’s an environment where failure is more tolerated than elsewhere, and that it’s okay to take risks because you’re not personally eliminated if they don’t work out. Second of all, one of the reasons that startups work so hard is because they’re within an inch of death all the time. Upside motivation is not really a factor when you have two weeks of runway left and need to figure out something. That’s real skin in the game, and it doesn’t have a lot to do with incentives. 

But history is written about the winners. We see the world through the lens of those who happened to survive and win; and we assign the fact that they succeeded directly to the notion that they won because they were incentivized to win. Tech has created this story where it’s a place where Skin in the Game is a great motivator. It’s true that upside can be used to great advantage in motivating people to work hard, but unless you’re in sales and a huge percentage of your salary is commission, then I’m loath to call that actual SitG; engineers making $200k a year and who also have a fat stock option incentive package wouldn’t qualify for Taleb’s definition of SitG any day of the week. 

This new narrative is pervasive enough in Silicon Valley that it’s expanded company-building and into the business models of many of these startups themselves. One such startup is Lambda School, a current tech industry darling. They provide their customers with a computer science and software development education, and in return, take an equity stake in the students’ future income rather than charging an upfront tuition fee. This setup, branded as “Income Sharing Agreements” or ISAs for short, has become one of the big new trends that everyone’s talking about: “ISAs for X” as the new “Sharing Economy for X”. 

Now here’s a question: does Lambda School have skin in the game? It does, actually - but not because they’re "incentivized to do a good job” any more than other startups. They have skin in the game because if they don’t do a good job educating their students and getting them good jobs, they don’t survive. They don’t earn the income they need to make, and they fold. What makes Lambda School work, assuming it continues to do well, is that they’ve built a school that actually does a good job teaching students. The ISA is there because in this particular circumstance, ISAs make a lot of sense: people studying to become software developers are probably in a better position to sell equity in themselves rather than fixed income in order to pay for their degrees. But the SitG for Lambda School here is not a positive motivator - it’s a filter, and a threat.

But a lot of people in the community have looked at the excitement around Lambda School and concluded, “The reason why this is working is because of ISAs. Therefore, let’s go fund businesses that are ISAs for X, and Y, and Z.” This is like saying Superhuman is doing well because they have a waitlist. It’s nonsensical! If you want to start a startup that benefits from this kind of added motivation, there’s a much easier and more effective way to do it: just raise less money! It’ll force you to work harder, learner, and smarter, and it’ll force you to get things right faster, or run out of cash trying. But I’ll bet you that’s not what these people have in mind. 

I think it’s true that people in tech typically work quite hard, especially early employees who have the most to gain if things go right. But in my opinion the reason why those founders and early employees are so effective in the long run is really not that they’re so motivated to be high performers. It’s the fact that they have to work under circumstances that are very close to danger, all the time. They’re like trail runners. In this case, anyway, I’m willing to side with the Taleb notion that it is, in fact, better exercise and will get you better results. But it’s a question of survival, not incentives.

One quick note from last week: wow a lot of you took that little thought experiment seriously! Thanks for all the email. 

The extortion economy: how insurance companies are feeling a rise in ransomware attacks | Renee Dudley, ProPublica This is a really interesting case of moral hazard I’d never considered before. Businesses are increasingly buying insurance against ransomware attacks, where hackers break in and encrypt your files unless you pay them their extortion fee. In many of those cases, the victims are able to fully recover the data by themselves without having to pay the ransom. But the insurers, apparently, often pay the ransom anyway. Why?

Well, because paying ransoms in full is good for criminals, and having lots of profitable criminals is good for the insurance business. Hackers, for their part, want to make that ransom payment as easy and frictionless to do as possible. Hackers and insurers mutually want each other to do well. When an insurer pays out a ransom in full, even when unnecessary, they’re sending a message to prospective clients: See? You need this! To an insurer, paying out ransoms is like paying for CAC. I bet you there’s some sort of really interesting Nash equilibrium in there - if ransomware criminals start charging for too much, the game falls apart; but if they can keep it just right, this looks like a totally sustainable (although obviously illegal) way of skimming money out of businesses’ legal budgets and then splitting them evenly between hackers and insurers. Yikes. 

The Meme Hustler | Evgeny Morozov, The Baffler (2013) This is an interesting long piece from several years ago about Tim O’Reilly. The piece is quite critical of O’Reilly and is definitely coming from one side of the spectrum in the open source community, but I learned a ton that I didn’t know and had to read it twice in order to really soak it all in. Recommended. 

Journalism is an Action | Hamilton Nolan, Splinter

The multilevel metropolis - on the radical origins and mundane deployment of the urban skyway | Jennifer Yoos & Vincent James, Places Journal

Dating apps used to seduce gullible investors | Duncan Hughes, AFR

Have a great week,


Can you lose on purpose?

Snippets 2, Episode 19

Y Combinator Demo Day took place earlier this week, and as usual, the companies generally look really impressive. It’s become this funny ritual in the tech scene twice a year at Demo Day, when we all watch the crazy scramble between YC companies and investors and wonder, “Gee, companies these days are consistently high-caliber enough that a lot of this all looks kinda random. How well could you do as an angel investor at Demo Day if you simply walked around and offered SAFE notes to the fourth, eighth, and each subsequent multiple of four-th founder you met?” Or one more advanced version I saw on Twitter (I apologize for not remembering who wrote this, because it was great): “Go to demo day, point a camera at the crowd instead of on stage, figure out which pitches are generating the most FOMO by watching the crowd, and then just offer term sheets to all those companies, sight unseen."

This got me wondering. In a market for term sheets like YC Demo Day, which isn’t “efficient” per se but certainly feels like it’s found some sort of maturity to it, how much of the ultimate success of any given term sheet offer is due to skill on the part of the angel investor, and how much is due to luck? Clearly, investing in startups is an exercise that requires skill. And the future inevitably won’t be kind to the majority of these startups; that’s just the math. But I feel like if you went to demo day and offered a term sheet to a completely random startup, or even to your pick for the least impressive startup there, there would be a non-zero chance that you’d end up having picked the best of the batch.

You might be inclined to assign this quasi-randomness of outcomes as a sign of YC or startups’ relative immaturity; but in fact it’s the opposite. In the public stock market, the most mature market we have for this kind of thing, stock picking is an activity that trends towards complete randomness unless you can prove the exception. Michael Mauboussin, who has a few fantastic books on the subject including The Success Equation: Untangling Skill and Luck in Business, Sports and Investing, has a great thought experiment question on Skill versus Luck that I think about all the time. It’s a really simple question: 

Can you lose on purpose?

The question is almost deceptively simple: if you can lose a game on purpose, then there must be some skill involved; if you can’t, then it might be an exercise in pure luck. Yes, of course you can lose a footrace, a game of chess, or a business venture on purpose. No, you can’t lose the card game War on purpose. But what about the stock market? If I gave you a thousand dollars with a specific ask: Invest this into public companies with the goal of losing all of the money as fast as possible; can you do it? 

Now, I’m sure many of your reactions were something like, “Well that’s easy. I’ll simply ____ (and then fill in the blank here with “Make 200 trades at a $5 commission” or “Mark my trades so that the IRS takes all my money” or “Buy some options or other derivative contracts that are guaranteed to not work out”) And yes, you’re correct in that all of those stock market-adjacent activities, like market making, diversification, tax efficiency, or creative finance, legitimately require skill. But just the pure act of picking? It’s harder to lose on purpose trading public stocks than you might like to admit. If you were genuinely skilled at picking stocks that underperform the market, then you could simply make the reverse trades and clean house. Most people are not able to do either; not on command, and especially not repeatably. 

The “Can you lose” thought experiment is interesting for two reasons; first, because it’s an interesting perspective on luck versus skill; second, because if the answer is Yes, then you might get some non-obvious insight out of asking, “okay, well then if I can identify a skillful strategy for losing on purpose, then what is the opposite of that?” 

Mauboussin’s thought experiment was restricted to public companies, though, where we could make two key assumptions: one, that the market is deep enough that all prices represent a genuine market, and one motivated trader’s win is another’s loss; two, that betting on a company doesn’t materially affect their chances of success in any direction. (Yes, the stock price going up or down will affect the company’s cost of capital, their executive tenure, and other things. But this is a thought experiment! Come on.) But with startups, and especially early stage ones like in YC, we can’t make either assumption. Raising money has a direct impact on a startup’s chances of success; and is there are no short sellers betting against startups. (Except maybe on Twitter.) 

But in a strange way, the phenomenon still could hold true. Public market companies are all companies that meet a minimum set of standards, and whose price is set more or less reasonably by the investing public, who cannot know the future but does their best on average. YC companies, similarly, are all startups that meet a minimum set of standards, and whose price is set reasonably at “Who knows?!?”, and where the caps on SAFE notes if we’re really being honest are more a reflection on the price of early stage money than they are on the particular valuation merit of a startup. 

Furthermore, the self-fulfilling nature of early-stage investing whereby offering a company a term sheet has some real impact on other investors also wanting to offer them a term sheet, increasing the odds that the round will work out, that the company will make it to Series A, and so on. Sure, offering a term sheet to a dud startup with dud founders probably won’t increase the odds of success if the company was a stinker anyway. But if basically all the founders in a YC cohort are impressive, and all of the startups are plausible, and all of them have followed the right steps to make sure their companies are a genuine shot on goal at scalable success, then it’s worth revisiting our question: can you lose on purpose? What would it take?

I think it’s an interesting thought experiment to imagine what losing on purpose at Demo Day would actually look like. You could only offer term sheets with punitive valuations and deal structures, but then you’d simply get ignored; you wouldn’t lose, you’d simply stay at baseline. You could offer uncapped SAFE notes, throwing caution to the wind and saying “Come at me, dilution”; that would certainly be a great way to ensure you’ll be wiped off a cap table in short order. But then maybe some founders take your offer, and opportunistically leverage your vote of confidence into extra momentum that helps make a difference on their path to success. (Or you could simply pivot into being a media company that monetizes impressively with venture investments made along the way.) Maybe it’ll take some creativity. You could act like a huge jerk to founders and get banned from Demo Day; but again that’s not losing; that’s just forcibly not playing. 

At the end of the day, I can only think of one strategy for effective self-sabotage: by asking the cursed question: “Who else is investing?” Think about it. Suppose you went to demo day and adopted a strategy: you talk to every founder you can, and offer them a term sheet, but only if someone else has already offered them one. This should, in optimal conditions, be a legitimately “skillful” way to gain the worst possible outcome, and find yourself with the highest amount of the worst-quality investments. You’re minimizing your chances of making an offer to any genuinely unique companies; nor will your advances at the genuine rockstar companies be likely accepted. You’ll successfully select for the most average startups, whose founders (who have the most information and leverage in the room) are consciously selecting investors who are offering a strategy deliberately designed to not make high-conviction investments. That ought to work, I think.

So, if an optimal strategy to lose on purpose is to deliberately select for companies that already have other term sheets, it follows that an optimal strategy to win on purpose would be to do the opposite of that: either try to make the first offer a company ever sees, the highest conviction offer a company ever sees, or simply the only offer a company ever sees. The math checks out on all of those, I think. And all three of those activities require real skill. Being first requires skill; it means having the best deal flow, and that takes work and skill. Being highest-conviction (and communicating it successfully) requires skill; it means being a good salesman. So it may turn out that Demo Day, just like the stock market, is tending towards a mature enough market that all of the skill and all of the hard work that matter are actually in all of the activities you have to do peripheral to the market itself; in earning the right to pick effectively; not the picking itself. 

One absolutely surprising read: as you likely heard, David Koch passed away this week. One thing you probably didn’t know about him: he averaged twenty points a game in college basketball. Another thing you probably didn’t know: he survived a commercial airplane crash. Koch was on a 1991 US Air flight from Columbus to Los Angeles in a 737 that collided on the LAX runway with another flight, killing 34 people. Koch survived, and wrote a detailed account of the event that’s remarkable:

Passenger’s account of escape from burning 737 highlights cabin safety issues | David H Koch, Cabin Crew Safety.

Thanks to M.G. Siegler on Twitter for the link.

Some other links this week:

Daryl Morey, GM of the Houston Rockets, on the NBA offseason, the future of eSports, and more | Selfmade with Nadeshot (YouTube)

Nine reasons why Disney+ will succeed (and why four criticisms are overhyped) | Matthew Ball, REDEF (March 2019)

Google is tightening its grip on your website with new AMP updates | Owen Williams

Less than half of Google Searches now result in a click | Rand Fishkin

How TJ Maxx has been able to thrive without costly e-commerce investments | Anna Hensel, Modern Retail

The collapse of a hospital empire, and towns left in the wreckage | Barbara Federalists Ostrov & Lauren Weber, Kaiser Health News

Welcome to Cleveland, where the Browns are contenders | Ben Baskin, Sports Illustrated

Have a great week,


WeWork could win

Snippets 2, Episode 18

So. WeWork, in whose co-working locations I’m sure many of you have spent time, is going public. It’s going for it at a time where – if you believe the news – the world economy is threatening to take a substantial dive. The business is valued like a tech company (in that it is hemorrhaging Softbank money for now) and the IPO prospectus is magnificently bold and vague about tech, mission, aggregation, community, and all sorts of scalable-sounding things.

But it’s quite unlike a tech company in that it’s selling something with distinctly non-software margins: office space. It’s also quite unlike a tech company in that it’s bringing along 33.9 billion dollars worth of lease commitments as it goes public. In its early years, WeWork enjoyed substantial discounts on these commitments (as is common in commercial real estate), but now they’re starting to come due in full and oh boy do they hurt. Can the unit economics work out in the long run? Can this business work?

So I set out this week with a goal of convincing myself that WeWork can plausibly win. And I think I almost got there? I’m not sure I believe they’ll win, but I do genuinely think that WeWork is disruptive – in the real sense of the word disruptive. Don’t believe me? Read on. 

The We Company S-1 is like Tech’s version of the Mueller Report. It’s a Rorschach test that shows you whatever you were looking to see. If you previously believed that WeWork is a burning bag of dog turds, then you probably still think that, except now with numbers. If you previously believed that WeWork is actually a genius power move, then you probably still think that, except now with numbers. 

There are so many S-1 breakdowns out there this week I’m not going to link to them all, save one: Byrne Hobart’s piece What is We!? Understanding the WeWork IPO. This piece is a good place to start, particularly since it dispels some of the hyperventilation around who-sold-what-to-whom that happens a lot whenever tech people read about a business in any other industry. For instance, yes, it’s actually quite reasonable for Adam Neumann to borrow money from WeWork, spend that money buying office buildings, and then lease those buildings back to WeWork. That’s basic separation of high-multiple and low-multiple assets; software people just freak out because they’ve never heard of the concept of a “low multiple asset”. 

But anyway, back to our black and white questions. Is WeWork a bubble, or is WeWork an arbitrage? Is WeWork equity a real estate proposition, or a technology proposition? The WeWork S-1 repeatedly insists that it is a tech company, and ought to be valued like one. But where are its software margins going to materialize from?

Scale? I don’t buy it; not in the “we will achieve software margins at scale” sense. WeWork’s revenue comes from renting office space, and their cost of goods sold is… buying and/or renting that office space. That’s a perfectly fine business model, but it’s not revolutionary. Yes, there are services you can offer along with office space, but unless those services have software margins and can be effectively bundled into the WeWork office offering (a huge assumption) then we aren’t there. Network effects? Also no. There is zero advantage to joining the WeWork network because your customers, competitors, or friends are on it. 

Aggregation? Come on. Commercial office space isn’t like buying diapers on Amazon or watching TV shows where the big variable cost is customer acquisition. Sure, WeWork may have a great brand and therefore be able to acquire tenants more cheaply than some random office manager. But that’s not the same as “WeWork is the choke point through which some critical mass of commercial real estate must pass.” I have seen some interesting ideas around how WeWork subscriptions might be an effective aggregator for other adjacent businesses like health insurance; that’s definitely a neat idea, but it’s not the core business. 

A pretty but unhelpful diagram from WeWork’s S-1 filing explaining how they’re going to make all the money

The prospectus also helpfully includes several other key trends they indicate will drive “The Re-Invention of Work”: Urbanization, Globalization, Independent Workforces, “Flexible Solutions” (by which they basically mean, “shorter leases”), Workplace Culture, and “Sharing Economy”. (What? Who wrote this?) So, what, then? 

I think a reasonable way to look at WeWork’s opportunity is as follows: 

As software and the internet transform the workplace and eliminate friction from the commercial office space market, tenants increasingly resemble one another, and can be served in standardized ways. The more “atypical” a company’s needs (small, fast-growing, multi-region, seasonal, whatever), the more this matters: any customer can now be an easy customer. Shopping for square feet, whether it’s for a day or a decade and for one person or a thousand, is now all pulling out of the same bucket – just like Airbnb transformed short term residential real estate in a way where $200 / night and $2000 / month now have to compete with each other directly; as they should have all along, but never could before.  

There’s an opportunity here to become a global scale low-cost, low-friction provider of office space. Professional office managers have always “used every part of the cow” in terms of carving up buildings and cash flows in order to optimally make everybody happy, but as friction and differentiation between one type of tenant and another type of tenant goes away, there’ll be an opportunity to create a low-end, good-enough, globally ubiquitous class of office space that can make money basically by squeezing as many tenants into as few square feet as possible while being able to juice occupancy by guaranteeing low rates for any duration of lease, for anybody, anywhere. 

Despite the branding and the Instagrammable colour schemes and the beer on tap, WeWork is not a premium product. The fancy fruit water certainly helps reassure tenants that “you’re in a good place; you’ve made it; this is career success”. But don’t fall for it. WeWork is low-end on the main dimension that matters for commercial office space, which is what is the bare minimum of square feet that an employee needs, and what are the cheapest ways to distract employees from the fact that they’re working in an open-concept cattle pen? The lofty rhetoric around community and enlightenment are a key part of maintaining kayfabe. How are employees supposed to keep up the self-delusion if media coverage of WeWork is all about their tremendous cost-cutting efficiencies?

However, that doesn’t mean that WeWork is strictly a low-end product among all elements of performance that tenants care about. WeWork’s competitive advantage is that it’s lower-friction than anybody else. It’s fast, it’s standardized, it works for you, you can get an affordable office space today if you need. They do genuinely live up to their branding as a “tech company” along this element of performance. Software and the internet don’t make WeWork space any cheaper or nicer, but they do make it available faster, with less friction, than any incumbent can match.

WeWork is genuinely “disruptive” – in the original Innovator’s Dilemma sense. They are a low-end disruptor, providing an inferior but good-enough product to cost-sensitive customers who are overserved by traditional corporate office space, but underserved along a new element of performance (speed, flexibility, standardization, “space-as-a-service”, whatever you want to call it.) If you’re an incumbent who makes money by selling square feet, that’s not attractive for you to compete with. The better move for incumbents is to retreat upmarket, and not poison their income statement by renting discounted desks one by one to freelancers.

Ignore everything else. This is the important part of the story. 

Now, here’s where it gets fun: how will WeWork keep up an economic moat in this new world, and sustain an attractive multiple on earnings (once it has them?) If it’s a race to the bottom, what’ll stop other people (particularly incumbents) with cash, experience, and other advantages from simply diving in after them?

Well… would you want to compete with them?

I have a thought here that I can almost say with a straight face. If WeWork can raise a giant pile of cash in its IPO and then not die in a multi-year recession starting any time now, it could find itself in a position where no one else is able to raise and money and get off the ground to compete with them without dying first. All the while it’ll be able to scoop up even more cheap commitments while times are lean, back-loaded in the usual way. That could actually turn into a real economic moat, like the classic example of trying to fund a second storage locker business in a town that already has an existing, massive storage locker facility. There’s just no point. 

That back-loading typical of commercial office space contracts may turn out to be important. If they can get really fast at spinning up offices and filling them with tenants who all pay full freight (cheap as it might be, relatively speaking) then that’s positive cash flow. WeWork gets cash in the door before they start having to pay it back as cost of goods sold to their own landlords. It’s really scary to compete with a business that is operating at revenue break-even but cash flow positive, if you’re starting from behind.

I suspect WeWork is banking on this being a key aspect of their economic moat. If they can get really fast at spinning up offices and get to enjoy a few years of actual, bona fide positive cash flow from each before their COGS comes due, and assuming that speed of spinning up offices for these kinds of businesses is proportional to how big you are (reasonable, I think) then it becomes formidably difficult for anyone else to break into this low-end, good-enough market before too long. (The assumption here, by the way, is that new entrants will not be able to play comparable cash flow games because they don’t have the deployment experience, scale, or order book to actually extract those few years of cash. Their revenues will only materialize as their costs do, too.)  

Note, by the way, that this is crucially different from businesses like Uber and Lyft, who also face criticism for not being differentiated or having an economic moat in any given location. The difference has to do with switching cost for customers. When Lyft enters a city where Uber is already operating, you can start poaching riders and drivers on day one, since there’s effectively zero switching costs for either. But with office space, there’s substantial switching cost. If you want to poach tenants away from WeWork, it’s going to take time as their leases run out and they move spaces; during which you’re losing that key window in which you can generate the positive cash flow that makes the whole thing work.

This is the most un-software like moat possible: “You can’t compete with us because you’ll die first.” Their S-1 might be conjuring up this incredible fiction of software buzzwords, but… I mean, so what? Is the goal for them to conjure up some impossible software-margin pipe dream, or is the goal for them to win? 

It takes a whole lot of stones to say, with a straight face, that you’re raising money for a business that’s going to command 30x multiples for a bunch of fancy sophisticated tech reasons, and then use that money to build a wall out of one dollar bills. But… but… what if it works! Just think about the case studies there’ll be; about how everyone missed this because we read the WeWork S-1 expecting it to show, “no one can compete with us because we’re so great”, but then it actually turned out to mean, “no one can compete with us because we’re so terrible.” That’s real disruption!

This might be one of the most pure, uncut disruption theses I’ve ever seen, the more I think about it.  It’s especially sweet given how Silicon Valley has somehow convinced itself that “disruption comes from the high end now”, or whatever people said to twist the Innovator’s Dilemma into knots just so they could say that Tesla and Uber were disruptive. I’m already enjoying how fun it’ll be to remind people: “and the real disruptor was… WeWork, all along.”

Does it mean it’ll work? lol no. Genuine disruption usually doesn’t work out, especially without some major technological force that differentiates the underdog, rather than pure low-end disruption economics. The majority of smart people reading through this S-1 have concluded that WeWork will blow up as soon as things turn south, and I mean, they’re probably right. But it’ll be super fun to see if the disruption thesis turns out to be right. I’m certainly watching with interest. 

There’s only one real take-home message from this, by the way. If you look at a business and your takeaway is “this business is terrible!”, before you write it off completely, ask: “could this be disruption?” It just might be.

Permalink to this post:

OuiWork? The quick case for WeWork as a disruptive business | alexdanco.com

This week’s special: two iconic regional foods:

How butter chicken roti became a Toronto classic | Karon Liu, The Toronto Star

The cult of the Chopped Cheese sandwich, New York’s most enigmatic icon | Justin Bolois, First We Feast

Other good reading links:

The arc of collaboration: why Slack is an else statement, there is no distinction between productivity and collaboration, and the opportunity to become Discord for enterprise | Kevin Kwok

Three years of misery inside Google, the happiest company in tech | Nitasha Tiku, Wired

Navy reverting destroyer ships back to physical throttles, after fleet rejects touchscreen controls | Megan Eckstein, USNI News

How a vanity plate reading “NULL” backfired big time | Jack Morse, Mashable

Have a great week,


Loading more posts…