Hacker News Reader

Comment by Zelphyr | Hacker News Reader

It's great that you guys are receptive to these issues but I can't help but wonder if, instead of solving this with what is clearly unpopular technology, we all would have been better off if Google had encouraged better practices?

If, as the article implies, Google will give me the same rankings if I do things to get the same performance as an AMP page then I would rather do that.

View on HN

> It's great that you guys are receptive to these issues but I can't help but wonder if, instead of solving this with what is clearly unpopular technology, we all would have been better off if Google had encouraged better practices?

Unpopular where? Between web devs on HN which are the main cause of the web site bloat plague?

Assuming all other factors are constant — a drop of conversion rates by 70% implies that real users are also getting tripped up with AMP.

That's quite a big assumption from one usecase without much evidence backing it up. I'm not saying the author is lying but I'd need more data before saying this is a repeatable scenario...

Even the author couldn't narrow down what exactly was the cause other than maybe:

a) the domain name being different

b) the addition of chrome w/ a link to Google support + a notification "You're now logged in as [user]@gmail.com"

c) minor CSS changes as a result of minimizing static assets

None of these screams an obvious problem, especially given the other benefits this process has given in return (which of course could be accomplished without AMP).

A conversion rate drop doesn't necessarily mean it's unpopular with users, just that people aren't converting.

> if Google had encouraged better practices?

Google has been using indicators like page performance in their ranking algorithm for quite a while now, so to be fair, it doesn't look like it helped that much. A single label and clear prioritisation apparently make it a far easier sell within companies.

That might be a more believable explanation of why they did it if they weren't hijacking the content and hosting it on their own domain with left-right swipes that take users off of your site. Users don't actually visit your site, but they visit a restricted shell of it on google.com.

I'm not a fan of AMP either; just wanted to point out that it's unfair to say Google should have encouraged better practises when they did just that.

(Not that I'm a fan of Google/any single company having to be that warden, but I digress.)

Except there's NOTHING technically requiring google to serve AMP from their own domain. None of the speed benefits can justify that. AMP could just be another standard where the basic lib is preloaded on the browser.

If you want to pre-load the page in the background in search results without allowing the site operator to track the user, they do need to serve it from somewhere they control.

You could just have an AMP doctype supported natively by Chrome (and whoever wants it) which does the same. Google just did not went that route.

They did announce that they still want to go that route, somewhat: https://amphtml.wordpress.com/2018/03/08/standardizing-lesso...

The destination server would still get a request on preload if there is no proxy operated by Google, no matter how you enforced the standard.

Yes, that's fine by me.

I think the problem is that encouraging websites to load faster, without giving them a tool to do so, would probably result in very little movement.

The article is wrong. Doing those things does not guarantee that your page is safe to prerender.

Reply on news.ycombinator.com