People aren't mathematical entities. On average, you're a fool, because most people are fools. But if I start out treating you that way, you're right to be upset.

In evaluating a feature, it's important to have a global view. But it's also important to have a local view. If you choose features at random, they're generally used by a small number of people, but they get made because they're important to somebody.

In this case, you can't think just about the people trying to view the video responses. You also have to think about the people who have been making them. Content creators are a vital part of the YouTube ecosystem, and getting non-creators to cross over and become creators is an important engine for content growth.

To somebody who has been using this feature, this is a fuck-you twice over. The first time is shutting down a feature they were using. The second time is the recognition that YouTube was having them do something that didn't work well, but that YouTube doesn't really care.

Can YouTube get away with offending a bunch of people? Sure. But treating people like they only matter when they are currently useful to you? That's for sociopaths and supervillians. When you're building things for people, it's important to treat them as people. They'll remember when you don't.

>People aren't mathematical entities. On average, you're a fool, because most people are fools. But if I start out treating you that way, you're right to be upset.

That's true, but the two cases can't really be compared. Groups of people are, very frequently, in nearly every discipline and line of work, treated as mathematical entities. It's how governments determine how to allocate resources and funding in social programs, and how advertizing and marketing work.

In a public company, one of the main goals is to make as much of a profit as possible. Reducing costs is paramount to keeping profits in the green, and when you're operating on the same scale as YouTube, even a tiny optimization like this can translate to huge savings in the long run.

I agree that people are treated as mathematical entities. And I even said that was fine if that's not all you do. The mathematician's solution to the problem of overpopulation is just to shoot people until you get below the required number.

You are also wrong about the main goals of a public company, and wrong again about how to get there. Those are central dogmas of the MBA worldview, but they're just dogmas.

A great example is cars. American car companies minimized costs and focused on profits. Toyota maximized user value and minimized waste in creating that value. Toyota has been kicking the asses of American car makers for decades. Even when Toyota teaches American car-makers their secrets, they can't adapt, because those and other items of dogma make it impossible to change: http://www.thisamericanlife.org/radio-archives/episode/403/n...

Your point is valid, but your opinion is out of context.

  The team is focused on enabling you to share video links in comments.

They're not removing the feature because people don't use it, but because they're planning to replace the feature with one that improves the original use-case.

Although it's true that the execution wasn't smooth, they should have removed the feature when the replacement was ready.

Right. And your last sentence shows that you are thinking of people as humans, not abstract mathematical entities who it ok to offend.

That's all nice and well, but my point is, no matter what change you make, you're going to alienate users. Youtube can't make everyone happy. It isn't feasible. This change, that affects the view-flow of 0.0004% of users, is an intelligent one. The change Youtube made to subscribed videos (mentioned elsewhere in this thread) might not have been as intelligent.

Youtube has to optimize for the greatest number of users it can. When you're working with a business at scale and there are millions of users (or even a billion), you don't get a bellyache over features that have to be scrapped because they're not used enough. It's a company, not a charity. That's how it works.

No, it does not affect the view-flow of 0.0004% of users. The actual number of users affected is not stated and cannot be inferred from the data given.

My point is that sure, sometimes you must alienate users. But good product management requires you to minimize that number, and the size of the pain. That's just good business.

it just wasn't worth it. Fully agree with that, but was it worth it to remove it before its replacement is ready? Since it was more or less unused it isn't like it was causing a big burden.