As far as I can tell, Google have the power to essentially end web bloat at one stroke: introduce severe Google ranking penalties for bloated pages. Websites would soon get the message and cut down on bloat.

Presumably the reason Google doesn't do this is that they'd have to punish many of the most popular websites, which might be seen as damaging the quality of their search results (at least in the short term).

Some of the bloat is very specifically Google's own ad and tracking code, so they are very much working at cross-purposes within Google.

Google's ad/tracking code is generally measured in the tens of kilobytes. Compare that to e.g., the tens of megabytes that news sites routinely waste autoplaying videos on completely different stories to the one you clicked on.

Google clearly already knows whether a website is primarily a text-based site or a video-based site, as it displays the two differently in search results. If they immediately blacklisted any text-based site with autoplay video from ever appearing on the first page of search results it would cut about 50% of web bloat overnight.

Yeah they have competing interests when it comes to truly minimising unnecessary resource usage, but there's so much good that they could do without going anywhere near cutting into their own analytics.

> introduce severe Google ranking penalties for bloated pages

That's literally what they're doing with the AMP requirement change, no? Instead of giving priority to AMP pages, they're giving priority to any pages which have good performance.

I'm not sure. The article does say:

> If you want higher rankings and more traffic from search engines, you need to optimize your site for a better, more performant and faster user experience.

But wasn't this how things were meant to work before AMP? Google search never had harsh enough penalties to seriously deter bloat, and I don't know that they're going to change that, they're just going to remove the preferential treatment for AMP.

AMP pages are incredibly bloated with all the ad assets that slowly load in. Media sites browsed with aggressive JavaScript blocking are significantly faster.

I've repeatedly browsed AMP and non-AMP pages (without blocking as a normal user) - this is basically a total lie.

The amount of crap on media pages (while they wail about privacy) is absolutely staggering. How many trackers do these folks need?

I got to MSBNC - a place looking to take down this tracking panopticon system and they are shoving

demdex taboola scorecardresearch tvpixel chartbeat sail-horizon condustrcts imrworldwide hotjar connect.facebook.net womanear.com mparticle.com

etc.

I mean, seriously - why not just use one (like google) and be done.

Can anyone explain why the need so many beacons on a page?

The biggest reason is that google doesn't provide the utility that each of these individual libraries brings to the table. Even if they did, there's nothing that says the total "omega tracking bundle" that you'd be downloading from Google would be any smaller than the aggregated total of these libraries. You'd definitely have fewer network requests but I'm not sure that would really move the needle as much as you'd want.

While there is probably a bit of overlap, many of the above tools have very different use cases. For example, chartbeat is commonly used more on the editorial side for writers to track article performance but imrworldwide is a subdomain own by Neilson that they use to serve their sdk which offers metrics for preroll video ads. Hotjar provides user heat maps but sail-horizon is part of sailthru and used for email marketing.

I interpreted this as the downside of competition in the ad network space. Similar to “why do we need 4 cell towers on the top of this building” or “why does Boston have so many hospitals”.

I don't think that users complaints actually had any impact. It seems more likely that avoiding regulatory scrutiny was G's motivation in scrapping AMP.

User complaints do drive regulatory scrutiny, and Google will point to a lack of user complaints to try to justify its behavior.

Why not both?

There are companies pushing the boundaries every day, with governments generally failing to even investigate unless there are enough complaints to raise attention.

Complaints by themselves depend only on shame, which most companies seem to avoid easily. Complaints that catch the attention of governments, on the other hand...

Perhaps, or they were seeing an uptick in DDG usage on mobile.

It was the wrong long-term solution for sure. But I think it forced publishers to reevaluate their priorities with respect to bloat and loading times, whereas prior attempts at quietly calling attention to the problem apparently didn't make a shred of difference...

AMP pages no longer get preferential treatment explicitly, but I'm guessing time-to-load is still a signal that is used by the algorithm.

I wonder if they have hard guidelines? Something like "your page should load and render in 1000 ms" on a broadband connection.

1000 ms for application/html. How far we have come...

When I remember getting BBS results faster... sigh.

  while (true) {

    Every available channel will fill with every available amount of content until the SNR gets so low that a different channel is created.

  }

I agree, and whenever I bring up that web designers can do anything but wrong, I've been piled up on before.

I'd still take a mildly broken AMP page to read an article over the "intended experience" with ads and trackers everywhere and any attempts to block them would break the page further.

Agreed - all the claims that AMP sites are slower / more bloated then non-AMP sites seemed like total nonsense to me. Maybe HN folks with blocking capabilities - but average folks like my mom, AMP was the place to be.

> and any attempts to block them would break the page further.

The fun part about ads and trackers is that they do not contribute anything functional to a page, so blocking them generally does not break anything.

I get the sense that the only reason this happened is because amp sites were returning less advertising revenue for sites implementing it vs regular web. If the money was the same or better then I can't assume it would have ended up this way.