AMP is Google’s thing where if they decide your site is too slow they’ll force you to let them host it or else drop your index rank. It’s a combination of restricted CSS format, small vetted set of JS libraries, and a CDN / cache.

It used to stand for ‘Accelerated Mobile Pages’ but now doesn’t. It used to be run and served by Google but now is its own foundation somehow, but as far as I can tell Google search only links to their own AMP cache. You don’t choose an AMP cache, it chooses you, like a dragon’s egg:

As a publisher, you don’t choose an AMP Cache, it’s actually the platform that links to your content that chooses the AMP Cache (if any) to use.

They claim that it’s for improving webpage speed for mobile users. Remember the comcast call where the sales agent is like ‘help me understand why you don’t want faster speeds’ and the caller is like ‘delete my account so I can hang up’, for 30 minutes? AMP is like that but it’s the web.

Nice site you got, shame if something happened to it.

They claim that the ranking algo ‘applies the same standard regardless of technology’ but in 2020 their ‘page experience’ criterion is a complex mix of compatibility, technology and design guidelines.

Like most regulatory regimes, complexity gives an advantage to players big enough to navigate and negotiate compliance. And Google’s AMP offering has the biggest compliance advantage at all because of their role in developing the standard.

Why was AMP was a good idea for Google?

It gives them a realtime dashboard of all user behavior and ad revenue flows across AMP sites (though they already have some of that from ubiquitous analytics JS).

It lets them control which competing technologies are used on the web, although this is a card you don’t get to play often because it’s legally risky.

In theory, it lets them transition slowly into ‘CMS dominance’, creating layouts and web components that match the way they think of revenue, tracking and content. They can go from being a search engine for ‘just web pages’ to integrating more closely with the CMSes for news sites, which in turn would allow them to display and monetize that content in different ways.

I think the CMS stuff never materialized because AMP was an instant bad smell for every player on the web. Other people have developed CMSes that publish to AMP, though – wapo for example may be making as much from their CMS as from news.

In practice, it seems like an antitrust suit waiting to happen and I’m not surprised they turned it into an ‘open’ foundation (techcrunch, amp press release).

In addition to ‘improving speed’, AMP also mandates a bunch of proprietary standards for ads and analytics, and on the face of it seems to favor Google analytics products (they’re built in).

It seems like AMP is actively trying to avoid the appearance of favoring Google by introducing formats and supporting third parties who conform to them (AMPHTML ad format, for example).

Bing, as always, wtf

Bing has its own AMP cache and is in fact the only other AMP operator according to amp.dev. While researching this I ended up on their webmaster docs for AMP and instantly regretted it:

Make your AMP web pages selected

Websites must match a high-quality threshold required by Bing to have your AMP content selected and considered for the AMP Cache. Like for typically HTML, bing likes unique, quality content. Websites which replicate content, redirect users quickly, or provide little depth often don’t fare well in our results and in our Bing AMP cache.

I like the mix of capitalization of ‘bing’.

Euro galaxy brain moves

If EU antitrust czar Margrethe Vestager and I swap minds for 24 hours as I have been praying for, here’s the 10-dimensional chess I would play to get Europe the homegrown Google it desires:

Compel G to link to a local AMP CDN in each country, regardless of whether said cache complies with the full AMP regulations. Ban them from operating both the search engine and the mobile-optimized content cache.

Once you have homegrown AMPs in France and Belgium (et al), two things happen:

First, you can favor local tools. Or at least not disfavor them. There’s a chance G has been suppressing (intentionally or unintentionally) Euro products and this gives you a whole local lumascape.

Second, and bigger, you can disrupt search. The constant stream of inbound search hits and on-site user behavior will let AMP operators reverse engineer how to build a search engine, and also bootstrap their search index. AMP isn’t just an in to disrupt AMP (which, let’s face it, might not be a money-making business unless you favor your own ad platform). AMP is an in to disrupt search.

There’s an argument that search engines are ‘of the past’ in that they were hot 10 years ago, and Google is desperate to move past search and ad revenue as their only golden egg, and wants to be an AI company. But still, search pays for the research into the other stuff and it doesn’t hurt to have one based in your country or region. And there’s another argument that Google sucks as a product and will get disrupted purely on the strength of quality and features, and search is about to have a renaissance.

Don’t get me wrong, I don’t think it’s right to use government power to favor hosting platform. But Google has already used its insitutional power to favor their hosting platform, and should be opposed. And Google’s overreach has opened them up to this kind of counterstrike. . If the end result of this is ‘nobody gets to have an AMP’, I’m happy.