I think the most basic and true explanation is that the companies we are thinking about started out with unusually high-quality products, which is why they came to our notice. Over time, the conditions that enabled them to do especially good work change and their ability tends to regress to the mean. So then the product gets worse.
Related ideas:
High-quality product design is not very legible to companies and it’s hard for them to select for it in their hiring or incentive structure.
Companies want to grow for economy-of-scale reasons, but the larger a company is the more challenging it is to organize it to do good work.
Of course, doing nothing at all seems ridiculous, particularly so for companies whose investors all invested on the premise of dramatic growth.
In many cases, a company probably originally designed a product that they themselves liked, and they happened to be representative enough of a potential market that they became successful and their product was well-liked. Then the next step is to try to design for a mass market that is typically unlike themselves (since companies are usually made up of a kind of specific homogeneous employee base.) That’s much harder and they may guess wrong about what that mass market will like.
I think an important part of this model is that “Doing nothing doesn’t look good to your boss.” Lets imagine ten websites offering the same service, and (for the sake of argument) assume one of them has the absolute best UI it is possible for any website providing that service to ever have. That one grows, the other nine die. Do the managers/UI programmers/visuals people at the successful website company decide “lets never change anything ever again”, or do they change things. Changing things is more fun, more rewarding, and easier to justify to your boss than doing literally nothing. By assumption is was perfect, so every change makes it worse.
The need to justify things to your boss, rather than to your user, is a result of institutional incentives generated by having a centralized reporting structure that justifies itself to the company’s main authority, typically investors.
I think the most basic and true explanation is that the companies we are thinking about started out with unusually high-quality products, which is why they came to our notice. Over time, the conditions that enabled them to do especially good work change and their ability tends to regress to the mean. So then the product gets worse.
Related ideas:
High-quality product design is not very legible to companies and it’s hard for them to select for it in their hiring or incentive structure.
Companies want to grow for economy-of-scale reasons, but the larger a company is the more challenging it is to organize it to do good work.
Of course, doing nothing at all seems ridiculous, particularly so for companies whose investors all invested on the premise of dramatic growth.
In many cases, a company probably originally designed a product that they themselves liked, and they happened to be representative enough of a potential market that they became successful and their product was well-liked. Then the next step is to try to design for a mass market that is typically unlike themselves (since companies are usually made up of a kind of specific homogeneous employee base.) That’s much harder and they may guess wrong about what that mass market will like.
I think this answer is correct.
I think an important part of this model is that “Doing nothing doesn’t look good to your boss.” Lets imagine ten websites offering the same service, and (for the sake of argument) assume one of them has the absolute best UI it is possible for any website providing that service to ever have. That one grows, the other nine die. Do the managers/UI programmers/visuals people at the successful website company decide “lets never change anything ever again”, or do they change things. Changing things is more fun, more rewarding, and easier to justify to your boss than doing literally nothing. By assumption is was perfect, so every change makes it worse.
The need to justify things to your boss, rather than to your user, is a result of institutional incentives generated by having a centralized reporting structure that justifies itself to the company’s main authority, typically investors.