Uber was a sociopathic company awash with venture capital dollars that operated at and beyond the edges of the law to get access to markets, to minimise regulatory and police investigations, and to maintain growth rates. It also co-opted leading politicians to help promote its growth. Emmanuel Macron in particular comes out badly, and George Osborne’s fading reputation dims yet further.

Via Oversharing, here’s a snappy Washington Post summary of the business model:

While users saw a handy app that brought rides to them with a few taps on their smartphones, executives behind the scenes were pumping billions of dollars of investor cash into an explicit corporate strategy of taking down “Big Taxi” and defeating other rivals. The money paid for hefty driver subsidies that were withdrawn once Uber got established, undermining earnings of those who had bought or leased cars to work for the company. It also offered deep, but temporary, discounts for passengers who grew accustomed to convenience offered at an artificially low price.

In theory, at least, Uber positioned itself as a “market-making” platform that connected customers to independent drivers. In practice, courts in multiple jurisdictions have found that this account of the Uber model was not the case. And at the same time, Uber was going to extreme lengths to rig the market that it had ‘made’.

124,000 documents

With 124,000 documents and a whistleblower to back up the story, some of the detail of how Uber operated, and the scale of it—and the active promotion of illegality by senior managers—still has the power to shock. Reading some of it, the culture reminded me of how the culture of Enron was described by whistleblowers. (For clarity this point is about business culture: I am not suggesting that Uber has committed financial fraud).

And although Ali Griswold at Oversharing, who’s been reporting ting on Uber for years, was obviously disappointed that there wasn’t more that she didn’t already know about, her summary of the things she was reminded of about Uber is still quite the list:

I started reading through story after story about how Uber broke rules and laws to expand its service globally; poured money into lobbying and public relations efforts to codify the ride-hail model; deployed venture-capital subsidies to lure drivers onto its network; and worked with prominent academics to build a body of research that legitimized its business and policy goals. All of these things are important chapters in the Uber story and vital to understanding how powerful and well-funded Silicon Valley startups can do business…. One of the big storylines from the Uber files is how Uber broke laws, lobbied governments, and leveraged violence and social unrest to advance its business model.

That was then

The company’s response was much as it has been ever since Kalanick was forced out of the company and replaced by Dara Khosrowshahi five years ago: that was then, but this is now. This is the official response:

“We have not and will not make excuses for past behaviour that is clearly not in line with our present values. Instead, we ask the public to judge us by what we’ve done over the last five years and what we will do in the years to come.”

Kalanick, who was instrumental in shaping the Uber culture, and who appears in many of the worst stories in the Uber Files, has issued the usual denials and special pleading via a spokesperson, probably carefully vetted by legal counsel.

So it was useful to see a piece in the Conversation by Jimena Valdez, who has researched the relationship between Uber and governments, assessing what has changed and what hasn’t.

What hasn’t changed

What has changed: the company is less aggressive:

The company has generally moved away from its original service towards one where licensed drivers use vehicles with specific permits to hail passengers (in other words, a taxi for the smartphone era), and introduced a food delivery wing, Uber Eats. It has also taken a calmer and more polite approach to expansion —- moving slower, breaking less stuff.

What hasn’t changed: it sails at the edge of regulation and expects to spend time fighting regulators (and drivers) in the courts:

Uber adapts to existing rules, but only as little as necessary to provide its services. Meanwhile, it continues to fight legislation everywhere – spending billions on lobbying and in crafting political connections – to push existing rules closer to its preferences… Though they are not openly breaking laws anymore, they continue to push for their preferred regulations through the courts or by finding legal loopholes.

The reason for this: the business model probably doesn’t work:

Uber’s executives know their business model might not be sustainable, and even less so if they are forced to classify workers as employees and pay for related rights and benefits. Fighting regulations is a survival strategy.

‘A sovereign state influence operation’

At The Blind Spot, and behind its paywall, Izabella Kaminska—a long standing Uber critic—reviewed the story. I don’t have access to that, but the free-to-air introduction in her weekly email is worth sharing:

(Transport expert Hubert) Horan and I argue that what Uber achieved in the marketplace was far more reminiscent of a political coup than a vanilla business deployment. Some might even say the company operated a sovereign-state-level influence operation, focused on capturing hearts and minds at all levels of society.

Reflecting on the many years I spent covering the company… I am inclined to suggest the company was always aware of the failings of its business model. Does this equate to it having misled investors? I do not know.

Elites love the elite

Some quick notes from me.

One: Rapacious companies aren’t new, and they tend to grow more quickly during the ‘deployment’ stage of a technological surge (borrowing Carlota Perez’ model here). What was different this time around was (1) companies like Uber were operating globally almost from the very beginning, and (2) that there was so much capital around that was cheap to the point of being free—and investors willing to believe any story about growth and returns to investment—that they could subsidise their expansion strategy and not worry too hard about the consequences.

Two: One of the reasons for our rolling political crisis, nationally and internationally, is that elites are closer to each other than they have ever been, and further away from everybody else. This is one of the reasons why there’s such a big ‘trust gap’. This is also a function of globalisation. But it is also a big political economy problem: politicians end up with a form of wilful blindness about the consequences of their actions.

‘Exploitation and corruption’

Three (related): Uber’s ‘innovation’, to the extent that it existed, wasn’t primarily technological. It was in connecting a sea of venture capital money with the ‘influence operation’ described above by Izabella Kaminska, which in turn made it harder for regulators to enforce regulation. Jimena Valdez says this:

Apps like Uber and the hundreds that followed promised innovation. Instead, they have brought a barely-disguised version of the exploitation and corruption that has always characterised capitalism. Given the allegations in the Uber files, one also has to wonder if there will ever be consequences for tech entrepreneurs with a taste for rule breaking.

Four: One doesn’t always want to blame consumers for their choices, since often they are locked into poor systems. In this case, though, we have known about most of this for about a decade, and there are decent alternatives to Uber, which have all improved over the last few years. As with Amazon, affluent consumers seem to be willing to overlook a lot in exchange for price and convenience.

Five: This is all another sign that the long digital wave is running out of momentum. (See also the collapsing fortunes of the online delivery companies.) If you are running a tech-enabled business, and your business model doesn’t work by now, you really ought to wind the thing up before you sustain more losses.

A version of this article was also published on my Just Two Things Newsletter.

Portrait of Andrew Weatherall by Tom McShane. Via Wikimedia Commons, CC BY-SA

One of the best ways to tackle climate change is to educate girls. It reduces population growth and reduces pressure on the environment. As well as being a good thing in its own right.

Ultra processed foods now account for half the calories we eat in the US and the UK. They are convenient, affordable, and very profitable. They also cause obesity and are bad for our health, says Bee Wilson in this long and compelling essay.

The fashion industry is being wrecked by the Coronavirus. Because: ‘the major labels depend on attracting Chinese consumers’.

Susan Fowler named sexism at Uber. She became a target for harassment, hackers and private investigators seeking to discredit her.

We need new words for the climate emergency. At least, that’s the ambition of The Ecotopian Lexicon. (Thanks to Ian Christie).

Machine learning writes fairy tales. These short tales are charmingly off-centre.

Andrew Weatherall, the producer who defined ’90s acid house, died suddenly. There were shockwaves in the DJ community. Collaborator Lee Brackstone remembers a friend and mentor. Here’s a playlist. His radio shows are online.

“The whole world’s at sixes and sevens, and why the house hasn’t fallen down about our ears long ago is a miracle to me.” (Thornton Wilder)

If you see an article worth sharing on Sixes and Sevens, please add it to the Comments.

 “The whole world’s at sixes and sevens, and why the house hasn’t fallen down about our ears long ago is a miracle to me.” (Thornton Wilder)

The short book falls into three parts. There is a context-setting chapter on ‘The long downturn,’ which will surprise no-one who has read, for example, Wolfgang Streeck’s book and articles on the end of capitalism. The second chapter analyses different types of platforms and their strengths and weaknesses. A final chapter, ‘Great platform wars’, looks at the prospects.

Official futures

One of the strengths of this approach is that it provides some valuable context. One early observation is that the tech sector remains small by value and even more by employment. In the US it represents 7% of the value added by private companies, and 2.5% of the employment. However, it is also a pervasive technology, both systemically and ideologically.

The ‘official future’ (my phrase) of the digital economy, as he observes, is that

cities are to become smart, businesses must be disrupted, workers are to become flexible, and governments are to become lean and intelligent. (p5)

The critical view, in contrast:

[W]ith a long decline in manufacturing profitability, capitalism has turned to data as one way to maintain economic growth and vitality in the face of a struggling production sector. (p6]

Platforms as data engines

This is where platforms come in. If data has become a massive new raw material for capitalism, then platforms are the engines that allow it to process this data. Srnicek identifies five types of platform (p49), described below. For me this is a valuable piece of analysis, since it allows greater clarity about their differences and their strategies, rather than bundling them together as if they are a single phenomenon. What they do have in common, though, is that “collecting massive amounts of data is central to the business model.” (p89)

• Advertising platforms (e.g. Google, Facebook), which extract information from on users, analyse it, and use the outcome of that process to sell advetising
• Cloud platforms (e.g. AWS, Salesforce), which own the hardware and software used by digitally-dependent businesses and rent them as needed
• Industrial platforms (e.g. GE, Siemens), which build the hardware and software necessary to transfer traditional manufacturing into internet-connected processes
• Product platforms (e.g. Rolls-Royce, Spotify), which generate revenue by transferring a traditional good into a service
• Lean platforms (e.g. Airbnb, Uber), which seek to connect buyers and sellers of a service while maintaining a minimum of assets. (I have adapted Srnicek’s definition slightly here to emphasise the *market-making* nature of such platforms.)

Questions of scale

For the detail of these you’ll need to turn to the book, but pulling them apart like this helps understand their differences. Advertising platforms have been hugely profitable, to the point where they don’t know what to do with their cash, but are running out of road in terms of growth and monetisation. Cloud platforms generate good returns (AWS is by far the most profitable part of Amazon) but require significant scale and investment, as well as robustness. Industrial platforms are specialist versions of cloud platforms; they need scale, and to be able to handle prodigious amounts of data. (Srnicek reports that GE’s natural gas platform alone is collecting as much data as Facebook.)

Product platforms, perhaps better thought of a service platforms, own the asset that they lease to the end user (this is the difference between Zipcar and Uber), although in the case of Spotify that “ownership” takes the form of a licence from the underlying rights holders, which is a problem for its business model. There’s a useful discussion here of the aero-engine market, now moved more or less completely to a service model: effectively this allows the engine companies to integrate higher margin maintenance with the lower margin product by connecting them through their data.

Venture capitalist welfare

Finally, lean platforms need to own the relationship with both buyers and sellers. They are lean because they outsource everything else. “It’s hard,” writes Srnicek, “not to regard the new lean platforms as a retrogression to the earliest stages of the internet-enabled economy.” Or further, I’d say, to the piecework of early industrial capitalism (pdf). As Srnicek notes, “the most notorious part of these firms is their outsourcing of workers.” (p75,76).

It’s worth taking a small detour here on lean platforms, party because Srnicek suggests that despite the hype about them, not to mention that tedious conference trope, which was apparently created by IBM, they are the most minor of these five.

The profit-making capacity of most lean models likewise appears to be minimal and limited to a few specialised tasks. And even there the most successful of the lean models has been supported by VC [venture capitalist] welfare rather than any meaningful revenue generation…. these models seem likely to fall apart in the coming years. (p87-88)

Building on this, Airbnb appears sustainable, because it can charge a larger price per unit than the other two, and therefore take a larger cut itself, the units have lower running costs, and often do represent genuinely marginal capacity, and the balance of power between seller and platform is more equal. Uber, on the other hand, has relied completely on “VC welfare” to undercut conventional taxi prices—to the point where the scale of its price advantage matches almost exactly the losses per trip funded from its reserves. I’m always surprised that competition authorities haven’t paid closer attention to this aspect of Uber’s business model.

Platforms need to expand

In the final chapter, he uses an analysis of the ‘tendencies’ within platform capitalism to identify some of the long-run trends. Since platforms are based on the extraction of data and the creation of network effects, they also have certain dynamics: “expansion of extraction, positioning as a gatekeeper, convergence of markets, and enclosure of ecosystems.” (p98)

Taking these in turn, platforms need to expand because they need more data. Because they need to expand, they tend to encroach on the same markets and the same types of data. One can see this most clearly in the competition between Google and Facebook, for example. Srnicek asks whether this tendency goes beyond competitors around similar platform spaces (e.g. Amazon and Alibaba) to platforms in general. I suspect this is unlikely. The skills and competencies needed to run an industrial platform like GE’s Predix, not to mention the customers and sales channels, seem to me to be fundamentally different to a consumer platform like Spotify or an advertising platform like Google.

Locking in customers

The final tendency, about the enclosure of ecosystems, is because the acquisition of data does not, finally, ensure a relationship with the customer, and so they have to be locked in, as much as this is possible without irritating them. Facebook has been effective as this; Apple too. In the industrial space, it is even more acute; it is possible to migrate your sales and customer data from Salesforce, but it is not something you would do lightly. Apps reinforce this tendency towards closed systems.

There are challenges here. Srnicek mentions the limits of outsourcing, a particular problem for lean platforms, the dependence on loose finance capital, and, for advertising platforms, a dependence on advertising growth. Platforms do continue to grow, but to do this, they have “to extract rents by providing services.” On this basis, “Amazon is more the future than Google, Facebook, or Uber.” (p126)

What is to be done? Srnicek suggests that in the short term national governments (and by extension, cities) might take more interest in platforms and their anti-social side effects:

Antitrust cases can break up monopolies, local regulations can impede or even ban exploitative lean platforms, government agencies can impose new privacy controls, and co-ordinated actions on tax avoidance can draw capital back into public hands. (p127)

Beyond that, there might be scope for publicly owned and publicly controlled platforms, creating new public utilities. There are co-operative platforms out there already.

Reaching limits

My own take on this is that while there should certainly be more controls on what platforms do, they might, nonetheless, be reaching their limits. Given that they need data to fuel their expansion, this comes from two places: growth in the number of customers, and second, growth in how much you can get those customers to pay for (monetisation).

In most markets, growth in customer numbers is plateauing, or close to it, in both consumer and B2B markets. And in a world of long-run slow economic growth, the prospects for further monetisation are slowing too. This may lead to some unpleasant effects as they try to maintain growth by gouging customers or playing faster or looser with their data. This also suggests that platforms that do not already have a sustainable business model are unlikely to get to one.

This means, in turn, that the loose finance capital (e.g. venture capital money) that has been keeping such platforms afloat are likely to cut their losses and move on to other areas of investment, probably not even in the digital sector.

Platform Capitalism is published by Polity Press. The image at the top of the post is by Andrew Curry, and is published here under a Creative Commons licence.