Illustration by Vincent Kilbride

How we’ve enshittified the tech economy

The latest threat to our online platforms is as bad as it sounds
October 4, 2023

Science fiction author and internet activist Cory Doctorow did the world a favour in early 2023, introducing a concise explanation of why so much of the internet seems to get worse every day. Doctorow’s insight is not only straightforward and powerful, but poetically named­­. He calls this process of platform decay “enshittification”.

Here’s how it works: many of the services we use online are “two-sided markets”—they match buyers and sellers and profit from the connection. Millions of companies sell products on Amazon to hundreds of millions of consumers. At first, Amazon was a great deal for consumers. Using money from the venture capital markets, it kept prices and shipping costs artificially low. So many customers flocked to Amazon that sellers felt compelled to offer their wares there, and initially Amazon treated them well, giving them access to vastly larger customer bases than they could reach before.

But as Amazon has captured more buyers and sellers, it has worsened the experience for both parties. Searching for products on Amazon as a consumer now involves wading through pages of ads for products that may not be what you want, with no guarantees of suitability or quality. Those ads have become an increasingly important revenue stream for Amazon, and are part of the massive fees—45 per cent of the sales price!—that vendors now pay Amazon for the privilege of reaching its customers.

A similar phenomenon is playing out across the digital economy, as tech-powered giants who surfed the digital wave to success abandon the practices that made them popular with consumers in the first place. Having done that, they then turn on their suppliers as well, in a bid to claw back all the value for themselves. Whenever this happens it doesn’t end well for anyone.

Enshittification shouldn’t be possible in a competitive marketplace. If buyers or sellers feel Amazon is ripping them off, they should be able to move to another platform. But Amazon—fuelled by investor cash—expanded so quickly and captured so much market share that it’s hard to think of a competitive “everything store” one could switch to. An early and enthusiastic Amazon user, I’m now trying to break away from the platform, category by category—I buy most of my books through used book dealers, but often find myself wrestling with Amazon’s terrible user experience to buy clothes or electronics.

Doctorow argues that buyers and sellers will have increasingly bad experiences with these platforms until they collapse into utter uselessness, devouring themselves like a rapaciously capitalist ouroboros. But on a visit to Colorado this summer, I caught a glimpse of another path forward.

I was invited by Nathan Schneider, a scholar of online communities and governance, to speak to a conference at the University of Colorado Boulder on “local tech ecologies”, technology projects aiming to serve local Colorado users. The highlight of the conference was a series of presentations from local nonprofits and businesses, many organised as cooperatives, who were taking on global-scale megabusiness competitors.

Uber is a classic example of enshittification. Initially, it provided friendlier drivers, cleaner cars and more convenient ride-hailing than many taxi systems. Drivers liked the flexibility of working when it suited them. By flagrantly ignoring local licensing systems, harming drivers who’d invested in medallions (the permit system for cities like New York) and those who’d undertaken intensive training (like The Knowledge, required of London cabbies), Uber won the custom of many riders and drivers. Again, the firm used venture funds to keep user costs artificially low.

Now having achieved, with competitor Lyft, a duopoly in many cities, Uber has proceeded with stage two of the process. Riders have complained of unpredictable surge pricing, which charges customers extra at moments of high demand. Uber now folds these increases into its “dynamic pricing”: the company will simply charge you what it believes the market will bear, giving you no sense for what a “normal” fare for your route might be.

University of California law professor Veena Duval argues that Uber is doing the same thing to drivers, offering different rewards for the same journeys. If truly abused, this differential could possibly be used to attract newer drivers to work for the platform with high rates before lowering them for drivers dependent on Uber. Drivers report they are making less money, with New York’s Taxi and Limousine Commission publishing data showing that app-based ride-sharing drivers now make less than taxi drivers.

Enshittification shouldn’t be possible in a competitive marketplace

The Drivers Cooperative, Colorado aims to reverse this trend. Following similar projects in New York, Los Angeles and San Francisco, the Colorado group aims to give drivers 80 per cent of rider fares, keeping 20 per cent to operate the cooperative and its ride-hailing app, which would work out as a far better deal for drivers compared to what they report receiving now. Two hundred drivers have signed up to participate, and will likely follow a model like New York’s cooperative. They will focus first on non-emergency medical transportation and scheduled, government-funded trips which are easier for a small team of drivers to cover, rather than competing head-to-head with thousands of Uber and Lyft drivers for real-time ride hailing. New York’s cooperative is close to breaking even, paying up to 9,000 drivers a $30/hour minimum wage.

Another Colorado project, Nosh, is using cooperative models to challenge another industry that’s become a bad deal for customers and suppliers: food delivery. Restaurants in Fort Collins, Colorado saw Grubhub (an equivalent of Deliveroo) increasing delivery fees, leading to customer complaints and decreased business. Restauranteurs joined together and built their own service, which saves an average of $4.90 on orders compared to established services. The locally owned service now claims 50 per cent of the city’s food delivery orders.

Cooperatives are likely to work better when goods and services must be produced locally, rather than being sourced from anywhere in the world. Amazon can negotiate with suppliers in China, Vietnam and Taiwan to provide millions of pairs of headphones at low prices, but there’s a finite number of drivers Grubhub can hire in Denver. And fields where service is a differentiator, rather than pure cost, may benefit from cooperative models, where happy, well-compensated drivers likely provide better rides than drivers struggling to survive.

Schneider, who convened the Colorado conference, and Trebor Scholz, a scholar of cooperative economies around the world, have written books documenting coops’ successes and shortcomings thus far. One major problem for cooperatives is accessing the capital that companies like Amazon and Uber have been able to raise, which allowed them to increase their market share. Building functional software for ride-hailing may be possible for groups of a few thousand drivers, but raising money for a spectacular advertising blitz probably is not. This can have the effect of limiting the visibility of cooperatives only to those already seeking alternatives to the large platforms.

If platform cooperativism, a term put forward by Scholz in 2014, is to rival the established platforms, enshittification may be the turning point. Many participants in the Colorado conference concluded that international companies would inevitably abuse both customers and suppliers in an effort to increase returns for shareholders. Turning to local cooperatives feels like a practical and ethical alternative. Maybe we can hasten the collapse of enshittified platforms and bring about the transformation Doctorow predicts as a conscious choice, not just the inevitable consequence of a harmful business model run amok.