The Dual Decays of Enshittification

The views expressed here by the author are his own.

Just this morning, as I was grabbing coffee with a couple of friends, one of them made the observation that “there are no good social media platforms anymore,” implying that the major players of the social media space (namely Facebook, Instagram, and X) have lost the characteristic panache that made their early incarnations revolutionarily engaging. These types of complaints are not uncommon, especially for those of us working in the tech space. I am regularly beset by family members asking why Google has “gotten worse,” why they have rolled out AI features, or why their search results have diminished in quality. I have come to find that the answer to these questions is enshittification.

Originally defined by Cory Doctorow on his blog, Pluralistic, the term refers to the sense that online platforms tend to devolve over time—that they tend to get “shittier,” participating in a technological race to the bottom. The idea has gained a great deal of traction in the year and a half since Doctorow initially coined it, having even become the American Dialect Society's 2023 Word of the Year.

Doctorow’s definition centers around the observation that many online platforms’ development decisions ultimately ruin their product for users. To illustrate his point, Doctorow describes how Amazon originally enticed users with the promise of cheap goods (initially sold at a loss) that were relevant to their interests, allowing them to gobble up a massive amount of the online marketplace. Now, with significant market power, Amazon has been able to effectively set prices at will, all the while showing users goods that are no longer the most relevant to their searches but which are instead positioned at the top of the pile due to sellers’ payments to Amazon. According to Doctorow, this evolution is largely due to the evolution of design decisions and who they are immediately “for,” as platforms shift from being designed based on customers’ or users’ interests, to businesses’ interests, to finally, the company’s sole interest, maximizing profit.

Platform decay vs ecosystem decay

While Doctorow clearly identifies a unique phenomenon, his observations are even more acute as we look more deeply at its downstream effects. In his initial post, Doctorow describes a process that occurs on a platform, later having downstream effects on the experience of the platform itself. However, we may find that these downstream effects come in two distinct flavors: platform decay and ecosystem decay.

This first flavor—platform decay—is what Doctorow describes above for Amazon. However, as the process of platform decay continues, the platform’s design decisions also have a secondary effect, one which contributes to the devolution of the broader ecosystem that the platform depends upon.

We may look specifically at the design of Google Search to illuminate this distinction. The recent rollout of the new AI Overview feature in the Google Search product has led to the simmering sense that the service as a whole has begun to decay. Results provided by the AI Overview are often notably off-mark with their suggestions (for example, suggesting that users should cook pizza with glue in the sauce to keep the cheese from sliding off). In response, some have gone as far as to claim that the new feature has ruined the software as a whole. In this case, the worsening of the product that is Google Search is caused directly by design choices and implementations made by the platform itself, i.e., through the implementation of a new AI feature. This process would fall under our definition of platform decay.

However, looking at the resultant process of ecosystem decay, we can see an entirely new dynamic arise. Within this category, the platform’s design decisions worsen the experience of the platform itself, but indirectly. Rather than directly worsening the product through the deployment of explicit design decisions, the product is worsened through the degradation of an external ecosystem (frequently an ecosystem which the company benefits from by co-opting external elements of it as their own product, regardless of actual ownership). In such a scenario, the platform is still met by the roundabout effect of making their product worse as well; however, in the process, they are characteristically taking others down with them.

Looking back toward Google Search, we may see an example of this latter category through reporting that, by virtue of Google’s algorithmic prioritization of websites, the content on websites across the Internet has begun to degrade to fit the mold required to boost performance in search engine optimization (SEO) algorithms. This phenomenon has already been documented perhaps most tangibly among recipe blogs, which have increasingly placed actual recipes at the end of long diatribes or stories in part as a means of capturing SEO attention. This change occurred in such a conspicuous manner that the act has even been satirized in the New Yorker.

In these types of scenarios, the platform’s design decisions, such as decisions as to how to optimize their search engine algorithms, do not directly harm the product. However, by impacting the ecosystem within which the product exists, an ecosystem that the platform attempts to co-opt as a fundamental aspect of the product itself, the platform is worsened through the downstream effects of their decisions. As a result, within our example, as the Internet degrades and becomes less tailored to the needs of its users, the platform that exists to make it searchable degrades in tandem.

Teasing apart this distinction is significant. As it is most often discussed, enshittification is limited to examples of where platforms’ decisions simply affect and worsen their own product. However, the phenomenon extends beyond these companies simply worsening their own technologies but also doing the same to entire ecosystems around them.