The Why and How of Google AMP at Condé Nast

We had to pause that project and drop it all to go work on AMP...

(so you could fairly say I'm a bit bitter about that)

edit: HN is another great example, of course :)

> AMP’s new speed gains in Google Search are due to several key optimizations that we made to the Google AMP Cache, such as server-side rendering of AMP components, and reducing bandwidth usage from images by 50% without affecting the perceived quality. We also used a compression algorithm called Brotli that Google launched a couple years ago, resulting in a reduced document size by an additional 10% in supported browsers.

So, no. Just stripping down the page would not make pages load as quickly as AMP. Then, there's also the fact that you get Google's CDN for free.

[0] https://www.ampproject.org/latest/blog/turbocharging-amp/

Well...of course. Limitations are empowering in their own way. HTML grew, and grew, and grew to the point that soon middle managers are demanding the most abusive tactics in a desperate tragedy of the commons. Users -- tired of pop-overs and subscription boxes and notification demands and location monitoring and janky scrolling and slow loading -- start to prefer silos like Facebook instant articles or Apple news. AMP comes along and says "we'll limit these to the minimum to elegantly provide text content and put an icon to let you know" and users love it.

Could the site optimize themselves? Of course but they won't.

Every time AMP comes up I remark that we really need a new HTMLite specification, and must demand the same promotion/iconography, or optionally allowing users to turn their browser to HTMLite mode where it will only accept validated HTMLite content. All to achieve the same user-benefits without the central control. Instead everyone just pretends that we all just need to behave as developers and it isn't needed.

What sucks up space that you can't get rid of? It's just a web page, same as any other page.

The reasoning behind consistency and experience which we talk about in the post are mainly with regards the AMP content itself. Valid AMP content (passing the AMP validator) from different publishers behaves very similar to each other. Meaning you won't have any surprises as the content loads, and it will load very quickly. The experience benefits revolve around the AMP Runtime loading advertisement and embeds on-demand without blocking the main content itself from displaying properly or causing it to re-layout as the user starts reading.

What are you referring to? Safari adds a feature requested by Google that if you share an AMP page it shares the canonical URL and not the AMP url. Is that the "anti-AMP"?

The screeds and anger about AMP on here absolute baffle me.

And what freedoms will be lost when we are all required to run AMP website?

So ranking #2 in search results on average correlated with an increase in click through rate. Oh, btw, just happens that you gotta use AMP to get the SERP gain.

Meanwhile, "monetization benefits are less decisive and pending further analysis of our current implementation."

Publishers find a number of benefits to AMP -- revealed frankly in the article -- but the only reason they use it in the first place is because Google is where a lot of those eyeballs originate anyway. If Some Obscure Search Engine with ~1% marketshare reworked their mobile page to prioritize pages that adhere to some criteria of limitations, used an iframe to frame the content so that it loads inside their page, and threw in a free CDN to sweeten the deal, who'd retool their publishing infrastructure -- as excellently detailed in this post -- to publish in this format?

AMP enables an alliance of convenience between one of the top sources of traffic (Google), and the authors of long-form textual content reliant on advertising revenue. To the AMP team's credit, it does so by aligning closer than its competitors to how the web actually works, despite the Google-Search-on-Mobile page then acting as a captive newsreader.

Most user criticism of AMP is focused on Google-Search-on-Mobile stealth-morphing into a captive newsreader, while most publisher criticism of AMP is from sites outside of its target market trying out the tech and alarmed at the lack of traffic reaching their server, forgetting that offloading traffic at high volumes is a good thing. AMP isn't for people who are upset that Google is stealing your content; it's for publishers who want Google to expose their content to wide audience and hope to generate revenue through means that correlate to that reach.

And, like many engineering blogs about employing Google products, seems like it might've been suggested by a Google marketing team.

Then Google will make their next move.

This is exactly how Froogle/Google product search went.

>> in this case it was clearly a net positive

Not so fast. When I was there, the earlier experiment was that its was a "mixed" success. This post also suggests strongly to me that that hasn't changed too much ("Monetization benefits are less decisive and pending further analysis").

For many publishers it comes down to they cant get slightly more traffic with AMP (because of the AMP carousel in Search results) at a lower CPM which is a positive. However, as more publishers come on board that gets tricky all over again and risk of it becoming another race to the bottom of lower CPMs.

It's a way for major publishers to get a dominant/favored position and crowd out "fake news" and it's a way for google to get a monopolistic publishing platform.

So it is a "win-win" for google and some of the major publishing industry. It is a lose-lose for everyone else.

Publishing/media has been waging a war against social media and decentralized internet. They have to losing audience due to the fractured nature of the internet. And they are trying to get back their "captive audience".

Whether it works or not, we'll see.