[buboard]

from USERS PERSPECTIVE, amp is meh. yes it's faster, does it really matter if you re getting an article in 1 second when you -obviously- plan to spend 2 minutes reading it? i dont know anyone who thinks so, perhaps your USERS do

> A strealined conent delivery platform prioritizing UX that the publishers failed to create

what you re implying is, Google failed to improve their algorithms to bypass obvious SEOs, so they are forcing everyone to use a dumbed down platform that is harder to SEO - for now!

[mrtksn]

Actually, it's a pretty well-known fact that users leave websites if they are not loading fast.

People don't go with reading plans to websites, the titles are optimized to bring you there and you don't know what's in the article. More often than not, the text on the website is not what the title made you believe it is. You can't plan ahead, you want to quickly find out what is this all about.

The article themselves are usually garbage optimized for SEO, long paragraphs of sentences that say the same thing but with different keywords. If that's not enough, they try to sway attention with ads and popups. Even if you had a plan about reading an article, the publisher's plan about you is different(tip: it's not about letting you read in peace).

The Web is horrible, it's even more horrible on mobile. AMP is an improvement.

[reaperducer]

Actually, it's a pretty well-known fact that users leave websites if they are not loading fast.

Is it? The only place I've ever seen push that view is Google. I've never seen any non-Google information reflecting that.

[gregable]

A few references listed in this article: https://www.cloudflare.com/learning/performance/why-site-spe...

- Mobify found that decreasing their homepage's load time by 100 milliseconds resulted in a 1.11% uptick in session-based conversion

- Retailer AutoAnything experienced a 12-13% increase in sales after cutting page load time in half

- Walmart discovered that improving page load time by one second increased conversions by 2%

[buboard]

those statements sound like incredible cherrypicking

does 1.11% sound a lot an is it statistically significant?

cut page in half from what to what? from 30 to 15 seconds? or from 2 to 1?

improving one second like how? from 12 to 11 or from 2 to 1? and is 2% really substantial?

[azylman]

Large companies spend a LOT of money trying to increase conversions by even tenths of a percent. Yes, 2% is substantial.

[buboard]

yeah , large companies. like , the 50-100 top sites right? Why would everybody else care about minor speed improvements

[dhimes]

I too am skeptical- except in extreme cases. Maybe if it's some sort of mindless bullshit site that I'm not really interested in I would leave it if it didn't load in, say, 5 seconds. But I don't really give a shit about the bullshit web. Life would be better without that anyway (except, of course, for companies who make a living selling ads on such sites).

But if somebody's leaving a page that has content they need because it doesn't load in 1 second, then I'd say they're a dumbass.

[buboard]

> users leave websites if they are not loading fast

then google is doing a bad job of presenting these sites to users, since obviously these sites should lose their rank, since users leave them. Google surely thinks bounce rates are important, no?

> The article themselves are usually garbage optimized for SEO, long paragraphs of sentences

and how is AMP fixing this? and who is responsible for SEO having these incentives? SEO literally means they optimize for what google wants

[mrtksn]

They do lose ranks[0].

AMP is fixing this because you can take a peek into the content almost instantaneously so you can just see what is this all about and leave.

https://developers.google.com/web/updates/2018/07/search-ads...

[buboard]

i don't see why any webmaster would want users who "peek instantaneously and leave". what a waste of an http request

[Mindwipe]

It's sometimes a little faster but also considerably less functional. The Guardian and Reddit both have AMP pages that are much worse than the actual pages.

[summerlight]

> if you re getting an article in 1 second when you -obviously- plan to spend 2 minutes reading it?

In my experience, 99.9 percentile is more like 2~30 seconds in NorCal with a potentially sub-optimal ISP. I expect this to be even worse in developing countries like south/southeast asia.

[Kiro]

You clearly don't speak for the users in the developing world with bad connections. Speed is anything but irrelevant.

[wizzwizz4]

Google AMP mostly just reduces the latency by pre-loading the pages from Google Search results. That has the side-effect of increasing the traffic on networks using Google Search when Google AMP results come up on the Google Search results, which they probably will given that using Google AMP will improve the site's Google Search ranking.

I haven't benchmarked it, so I can't actually speak from anything but wild speculation, but Google AMP might actually be making this problem worse.

[mrtksn]

I gave it a try. So I searched for "Trump" in Safari mobile, the result page loaded 2.5mb. It loaded a bit more when I clicked on an AMP article and kept loading more as I jump from article to article. It did not load any data without me switching to the next article.

After 10 articles, the transferred data was 28MB. I visited the Guardian's own website and it fetched 3MB.

So, It looks like Google does not preload a huge amount of data.

ps: I use ad-blocking and cleaned the cache.

[arthurcolle]

How are you able to look at how much data was transferred on Safari mobile? As far as I know you can't access dev tools on mobile. Did you just change the User Agent on your computer to simulate?

[mrtksn]

You can! You can use the desktop Safari to connect to your iPhone's Safari and access full developer tools.

Just connect your iPhone to a Mac, open Safari desktop and in the Develop menu, you will see your mobile phone. When you open a page on mobile Safari, you will be able to see it from the Develop menu and when you click on it, the full Safari developer tools will open in a new window. Works just like the regular developer tools.

[wizzwizz4]

I do know it preloads on Chrome on Android on mobile networks, but it's a little surprising that it doesn't on Safari on iOS.

[Kiro]

I know it's not kosher to complain about downvotes but I said nothing about AMP, just argued against the statement that users don't care about speed. Anything that could possibly be interpreted in Google's favor gets downvoted immediately. It's carpet bombing with collateral damage.

[buboard]

slow users everywhere are better served with the simple amphtml page without any of google's scripts

The major problem is the download size anyway, and most of it is google-served ads and trackers. Amp is making this asynchronous, but afaik it doesn’t get rid of the ads