Enshittification, a term coined by Cory Doctorow in 2022, describes a predictable lifecycle in platforms built on two-sided or multi-sided markets: a service first allocates surplus to end users to establish lock-in, then shifts that surplus to business customers once users are dependent, and finally claws back value for the platform operator itself once switching costs are high enough on both sides.
The mechanism is not moral decline but a systems property – it’s what happens when a platform has enough monopoly or monopsony power, weak enough regulatory constraint, and low enough interoperability that users and business customers can’t easily leave even as quality degrades.
For engineers, the useful move is to map this onto architecture and metrics: enshittification shows up as growth-team incentives optimizing engagement or take-rate metrics disconnected from user-reported quality, as APIs deliberately left unstable or undocumented to blunt third-party interoperability, and as dark-pattern friction inserted specifically to raise the cost of leaving rather than the cost of arriving.
For product designers and technical leads, the concept is worth treating as a design constraint rather than a rhetorical complaint, because the same four levers – end-user lock-in, business-customer dependency, competitive moat, and regulatory or technical exit cost – are levers you are pulling (or resisting) in every roadmap decision, whether or not you use Doctorow’s language. A team can audit its own system by asking where switching costs are being engineered intentionally (data portability, API access, algorithmic transparency) versus where they’re emergent side effects of technical debt, and by tracking whether product changes are shifting surplus toward the platform at a measurable cost to the two sides that made the platform valuable in the first place.
The concept’s durability comes from being falsifiable at the level of a single decision – a given feature either narrows or widens the gap between what users get and what the platform extracts – which makes it a sharper design-review lens than “don’t be evil” style heuristics, and one that maps directly onto concrete engineering choices: rate limits on third-party clients, contract terms with advertisers or sellers, and the granularity of data export tools.
Key Takeaways for Engineers and Product Designers
- Enshittification is a structural outcome of market power and low interoperability, not a moral failing of individual product decisions.
- The sequence always moves from subsidizing users, to extracting from business customers, to extracting for shareholders once both sides are locked in.
- Switching costs are a design surface: teams can audit whether they’re emergent technical debt or deliberately engineered friction.
- Engagement and take-rate metrics can improve on a dashboard while the underlying product experience degrades, because the two are not the same measurement.
- Interoperability – stable APIs, data portability, open export formats – is the single lever most directly opposed to this pattern.
- Dark patterns raise the cost of leaving rather than the cost of arriving, which is a distinct design failure from poor usability.
- Competitive pressure and credible exit options are what keep the extraction phase from being sustainable, regardless of operator intent.
- The pattern applies as much to enterprise SaaS lock-in as to consumer platforms with network effects.
- A single feature or API change can be evaluated against this framework directly, which makes it usable in design review rather than only in critique.
- Reversal is rare without external pressure, since internal incentives keep pushing toward further extraction once it has started.
Our very related full articles, sometimes with a similar humor as the video below:
Introduction, “A Day in the Life of an Enshittificator”
As our introduction to enshittification, a brilliant advert from Forbrukerrådet, the Norwegian Consumer Council. An extremely fun advert to make all realize the truth of some aspects of current times, technology and marketing.
The movie is produced by NewsLab: https://www.newslab.no/
Script & direction: August Jorfald
Photographer: Alexander Herlev Christoffersen
Thumbnail art: Axel Bjellås Gilje
And the wording and all the references hit so well the modern product design, marketing and trends, that we couldn’t resist not include them here (all copyrights to the author above):
- You can call me a shitmaker. My job is to make things shitty. The official title is Enshittificator.
- What I do is I take things that are perfectly fine – – and I make them worse.
- More specifically… I make it shitty. I come from a long line of people making things shitty.
- My father was an Enshittificator, his father before him and his father before him. So yeah, it’s a proud tradition in our family.
- Making things shitty. But it is hard work. And it’s not very cost efficient to do it like this.
- I can only be in so many places at one time.
- So I needed a way to to scale up the business, – – without working myself to death.
- And just when I needed it – – the solution lands right in my lap. Enter the internet.
- All of a sudden I didn’t have to move around to make things shitty.
- I could just sit in one place and make things shitty – – for way more people at the same time. And that’s when things really started to move.
- Got a favorite website? Let’s fill it with pop ups. Preferably scam. Enjoying a video?
- Uupsi! Ad break! Oh, you bought an expensive phone?
- Great! Here’s an update – – to make it shitty. Let’s say you have a thousand pictures of your deceased mother.
- Oh, so sad. What do I do? I make you pay for it. The memories may be yours, but the photos, they’re mine.
- So it’s very important, it needs to be nice at first. But then… Yes. Oh, yes.
- What do you say? Okay! That’s what I’m talking about! You’re not happy with the way something you really liked turned shitty.
- What do you do? Contact customer service. Here, talk to a completely useless chatbot.
- “No human available.”
- Here’s my favourite part: Let’s say someone starts an alternative service. What do we do? We buy them.
- Make people dependent enough, – – and then … MAKE IT SHITTY!
- You. Sell me … A car. Well, it’s great and ehm.. Sold! I’ll buy your car.
- Now what? I don’t have any more money. You have me in the palm of your hand. You can crush me. So what do you do?
- Updates. Radio. Premium feature. Heating? Premium feature.
- Now I have a shitty car and I can’t afford a new one. But what I can afford is a small monthly payment.
- MAKE IT SHITTY! The future?
- The future’s looking bright. Braking is now a premium feature.
- Please enter your credit card details.
The Enshittification Concept by its “Inventor”, Cory Doctorow
Once of the great talks on enshittification modern trend by Cory Doctorow, and what we can do against that
Outro: how the Enshittificator got even more efficient
In the same serie, again by the Norwegian Consumer Council, how the Enshittificator partnered with Death to be even more efficient.
Agency: NewsLab at https://www.newslab.no/
Death: Frida Homlung Enshittificator: Bjørn Asgeirsson
Written & directed by: August Jorfald
Cinematographer: Adam Enochsson
Edited by: Nicolai Gundorff Asmussen
Sound on set: Ivar Kvalvik/ Boombay
Production manager: Ida Solheim-Olsen
Advisor: Snorre Bryne
Thumbnail art: Helen Mehammer
Contributors from The Norwegian Consumer Counsil: Marianne Andenæs, Elin Volder Rutle, Andreas Framnes, Hanne Fossaa Eriksen, Tove Bø Laundal.
Related Concepts
- Two-sided market dynamics: pricing and value allocation decisions on one side of a platform directly determine leverage and extraction potential on the other.
- Vendor lock-in: switching costs engineered through data formats, integrations, or workflow dependency rather than through superior ongoing value.
- Network effects: a product’s value to each user increases with the number of other users, which is what makes early-stage user subsidization rational.
- Interoperability mandates: regulatory or technical requirements that let competitors or third parties build on top of a platform without operator permission, directly reducing lock-in.
- Data portability: the practical ability for a user to export their history, graph, or content in a format usable elsewhere, distinct from nominal export compliance.
- Dark patterns: interface designs that increase the cost of an action the user wants to take, such as canceling or leaving, rather than the cost of the action the business wants.
- Loss-leader pricing: subsidizing one side of a market below cost to build share, with the expectation of recouping margin later through a different lever.
- Algorithmic feed curation: ranking systems whose objective function can shift from relevance to engagement to monetization without any visible change in interface.
- Take-rate creep: incremental increases in the percentage a platform retains from transactions between its business customers and end users.
- Attention arbitrage: extracting value from the gap between what a user’s attention is worth to advertisers and what the platform pays back to the user in utility.
- Multi-homing cost: the friction of using several competing platforms simultaneously, which determines how easily a user can hedge against any single platform’s degradation.
- Churn suppression: retention tactics aimed at lowering measured cancellation rates without addressing the underlying decline in product quality driving them.
- Regulatory capture: a condition where the entities meant to constrain a market’s dominant players instead operate in ways that protect those players’ position.
- Platform governance: the rules, APIs, and policy enforcement a platform sets for third parties operating on it, which can be tightened unilaterally at any time.
- Moat erosion strategy: deliberately investing in a competitor’s switching costs or ecosystem lock-in as a defensive move rather than a product-quality move.
FAQ
What is enshittification, precisely?
A term coined by writer Cory Doctorow for the predictable decay pattern in two- or multi-sided platforms: value is first allocated to end users to build a user base, then reallocated to business customers once users are locked in, then reallocated again to the platform operator once both sides face high switching costs.
Why does it happen in a specific sequence rather than all at once?
Because each phase depends on leverage the previous phase created – you can’t extract from business customers until end users are locked in, and you can’t extract from the platform’s own shareholders’ priorities dominating until business customers are dependent on the distribution the platform controls.
What makes a platform vulnerable to this pattern?
Three conditions: monopoly or monopsony power over a market, weak external constraints (regulation, competition, or organized labor), and low interoperability, meaning users and business customers can’t cheaply move their data, audience, or workflow elsewhere.
Is this just a complaint about corporate greed?
No – the mechanism is structural, not psychological. A platform with genuine competitive pressure and functioning exit options can’t sustain the extraction phase even if its operators want to, because users and business customers simply leave.
How does this show up in day-to-day engineering decisions?
As deliberate choices to weaken interoperability: unstable or undocumented APIs, rate limits aimed at third-party clients rather than infrastructure load, and data export formats designed to be technically compliant but practically unusable.
What’s the role of metrics in the process?
Growth and engagement metrics get optimized for engagement or take-rate, and once those metrics are disconnected from a direct measure of user-reported quality, teams can improve their dashboard while the actual product experience degrades.
How is this different from a company simply raising prices?
Price increases are visible and can be evaluated against alternatives. Enshittification is typically implemented through quality degradation, feed manipulation, or terms-of-service changes that are harder for a user to notice, quantify, or act on before they’ve already sunk switching costs.
What is a “switching cost” in this context, concretely?
Anything that makes leaving a platform expensive: lost social graph, lost purchase history, incompatible file formats, retraining time, or the absence of an equivalent alternative that supports import.
Where does regulation fit in?
Doctorow argues that antitrust enforcement, interoperability mandates, and consumer protection law historically constrained this cycle; their weakening over recent decades is presented as a primary enabling condition, not a side detail.
Can a platform reverse course once it’s extracting value?
Rarely, without external pressure, because the internal incentive structures – investor expectations, revenue targets already built into the business – keep pushing toward further extraction rather than restoring user surplus.
Is enshittification unique to Big Tech platforms?
The pattern is described in the context of digital platforms because network effects and low interoperability are unusually strong there, but the underlying two-sided-market dynamic exists anywhere a business intermediates between two groups with asymmetric bargaining power.
What’s the difference between enshittification and ordinary product tradeoffs?
Ordinary tradeoffs balance competing legitimate needs and can go either way. Enshittification specifically describes a one-directional transfer of value away from the groups that made the platform valuable, toward the platform operator.
How would a product team audit their own roadmap for this pattern?
By checking whether a given feature narrows or widens the gap between what users receive and what the platform extracts, and by separating switching costs that are incidental technical debt from ones that are deliberately engineered.
What technical countermeasures reduce a platform’s ability to do this?
Data portability standards, open or stable APIs, algorithmic transparency requirements, and interoperability mandates that let users or competitors build on top of the platform without the operator’s permission.
Does enshittification apply to internal or enterprise software, not just consumer platforms?
Yes – the same dynamic appears in enterprise SaaS, where a vendor cultivates deep integration and data lock-in with a customer before increasing prices or degrading service quality once migration costs are prohibitive.
Why has the concept gained traction among engineers specifically, rather than just critics or journalists?
Because it gives a falsifiable, decision-level lens – a single feature, rate limit, or API change can be evaluated against it – rather than a values-based judgment like “this company is evil,” making it usable in design review and architecture discussions.
