I build software inside the system that produces the products people complain about. I have watched good ideas get killed not because they were bad ideas, but because the spreadsheet said no.
The word people reach for is "enshittification." It is satisfying to say and it captures a real feeling — that the things we use keep getting worse on purpose. But "on purpose" gives too much credit. Most of the time, the product gets worse because of arithmetic.
Take a company with 200 engineers, each averaging £80,000 a year in total compensation. That is £16 million a year in engineering costs alone, before office space, infrastructure, management, legal, HR, and the rest. That team needs to generate enough revenue not just to cover its own cost, but to justify its existence to investors who expect a return.
Every feature decision passes through that filter. Will this generate or protect enough revenue to be worth the engineering time? Innovation, which is uncertain by definition, loses to optimisation, which is measurable. A/B tests beat conviction. Incremental improvements beat bold bets. The product gets 2% better at extracting revenue and 2% worse at being something a person actually enjoys using. Compounded over years, that is how you get a product everyone uses and nobody likes.
This is not a conspiracy. It is a budget meeting.
The larger the company, the more layers of approval any idea needs to survive. By the time your proposal reaches the person who can greenlight it, it has been translated through three levels of management, each adding their own context and priorities. The person who says yes or no may never have spoken to a user. They are looking at a dashboard and a quarterly target. Your idea either moves a number or it does not.
If it does not, it dies. If it does, it ships — but shaped by the dashboard, not by the original insight.
I have been in rooms where the right product decision was obvious to everyone present, and the wrong decision was made anyway because the right one could not be justified in the next earnings call. Nobody in the room was wrong. The incentives were.
So when people ask why everything sucks, the answer is boring: large companies push every decision through financial filters that have nothing to do with quality. The product is not the point. The product is the vehicle for the revenue model. You see it everywhere once you know what to look for — in the streaming service that buries its best content under algorithmic slop, in the game that ships half-finished and monetises the rest, in the app that used to be simple and now has seventeen upsell modals.
The standard advice is to go independent. Build your own thing. Be small, lean, free. I understand the appeal. I have felt it. But I also know what it actually costs, and the advice is easier to give than to live.
Going independent means trading one set of constraints for another. You escape the committee, but you inherit every job the committee used to do: marketing, sales, support, infrastructure, compliance. Your runway is your savings. Your safety net is your employability if it fails. The indie game that took three years in your spare time still needs to compete for attention against products backed by marketing budgets larger than your lifetime earnings.
Some people make it work. They tend to have savings, a partner with stable income, or a skill that translates directly into a product. The advice to "just build it yourself" assumes resources that are not evenly distributed, which is the same structural problem the enshittification argument describes, just pointed at individuals instead of companies.
None of this means you should not try. It means you should go in honest about what you are trading and what you are not. The bureaucracy is real. But so is rent.
What I keep coming back to is that people at these companies do care. Most of them, genuinely. The problem is that caring is not a line item. It does not compound quarterly. And in a system where everything that matters must be measurable, the things that make a product worth using — taste, craft, restraint, the willingness to leave money on the table because the product is better for it — those things have no column in the spreadsheet.
The product gets worse because the spreadsheet gets better. That is the mathematical reality. And I do not know how to fix it from inside the room. But I know that pretending the answer is "just go indie" is the kind of comfortable fiction I would not accept in any other part of my thinking.