Hacker News new | comments | show | ask | jobs | submit login
Google May Be Stealing Mobile Traffic (alexkras.com)
533 points by akras14 on Oct 17, 2016 | hide | past | web | favorite | 250 comments

"I hope I am not being too hard on Google." -- no, the author is being very, very nice. Google has a strong business interest in users never going to other sites. Between AMP & voice search, the future is going to be very rough for businesses that rely on free search traffic.

In practical business terms I would strongly suggest website owners build ultra-light versions of their sites. If you have international aspirations your site should work on Opera Mini. If you have a big audience it is very reasonable and worth it to have desktop/tablet, smartphone, and ultra-lite versions of your site.

Reposting part of a comment I made a month back on AMP:

#1 There is a big problem with mobile sites. I'm using a recent iPhone and many popular news sites, without ad blockers, are as close to unusable as the worst websites I've ever been to, dating back to using Internet Explorer in 1999. Auto playing inline video ads that slide in to view, just insane. These things clearly kill time on site and reader retention. I have theories about why publishers are ignoring this, but who knows.

#2 Google is using AMP to co-opt publisher's traffic. This means users are scrolling to another story from another publisher or easily bouncing back to the Google results when they land on your content. (See the X in the story link on the animated gif example.) There goes your time on site and long term user retention. If #1 was a problem for you already, you probably don't notice.

#3 AMP & Instant articles are going to put a stranglehold on third party ad networks and represent a very real anti-trust issue. There are a bunch of other privacy implications too, which have been discussed in length. Publishers should be thinking really hard about their future.

Web publisher here...the real problem is twofold:

- Display advertising makes the web suck; while display ads fund a large share of general market content sites, from a purely technical perspective they kill site performance.

- All major market participants suck equally, rendering no alternatives for end-users or advertisers to defect to; this inherently limits any incentive for real change.

AMP / Mobile index (new development from Google) breaks the gridlock: divert traffic to sites who deliver a good (fast) mobile experience and (assume site owners are economically rational - we are) is supported by something other than old school display advertising. Solving the 2nd part is left as an exercise to the student....

Interesting enough, you can make a banner ad load really fast. Just host images on your site and load it directly from your server without all the tracking scripts and separate calls from your advertising network (plus others). Or go completely native and render text links in html. Granted, this requires you to actually market your content and merchandise an offer (affiliate marketing).

Display advertising actually isn't the highest CPM option for a website; I did a study on small site auction data a while back and it was the lowest CPM model (admittedly, laziest to implement):


Google's DFP adserver is the slowest part of the web - they already had a vector to improve everything without AMP.

How do you suggest publishers who are dependent on traffic from Google cope with AMP in the short-term? It seems like there is no way to compete with AMP-enabled content on Google ranking except by adopting AMP.

Therein lies the problem, Google owns 80% of the market. Google is a de-facto monopoly. They can force you to act against your long term interests for a short term gain.

I personally find the gain suspicious, it's already possible to make fast pages.

Soon, we'll all merely be Google content providers.

Stop using amp and find other ways to optimize and minifity their websites for mobile? Foundation has the ability to include only the components you want if you build it using SASS and by including/minifiying the individual Javascript files during your build process.

The more I look at AMP the more I absolutely hate it. This is just another thing to make me not want to use it.

If Google points to a site, they should serve your site. If they're serving from cache, put in a DMCA and tell them hands off your content.

It would be sooo bad if your business suffered a search rank penalty due to the lack of AMP. It would be a real shame.

I don't think if you're running a business there's any way to avoid optimizing for Google in the short-term. You simply have to do it, if you want to be seen. Not optimizing for Google won't hurt Google, it'll just hurt you. Because Google can always just promote someone else.

But you should also support antitrust legislation and judicial action against Google. Don't vote for candidates next month receiving large donations from Google (now referred to as Alphabet[1] on most donor lists). The long game is requiring open oversight of Google's search ranking methodologies and forbidding them from modifying them to benefit their own products and services.

[1] http://www.opensecrets.org/orgs/recips.php?cycle=2016&id=D00...

As an outsider observing the US election the statement of yours "Don't vote for candidates next month receiving large donations from Google" strikes me as ridiculous.

Don't get me wrong, your logic is correct. But, in order for politicians to play the game they need large donors like Google. So, if they don't take money from Google they'll just take it from someone else with an agenda. The real solution is to vote for candidates pushing for campaign finance reform (reading Lawrence Lessig may be insightful)

I just want to point out that it is illegal in the U.S. for federal candidates to take money from organizations, such as corporations like Alphabet, Inc.

When you see campaign contributions listed "from" Alphabet, that money is actually all coming from individuals.

Employees of Alphabet who give more than $200 are required to list their employer on the donation form. The campaigns then submit that data to the FEC, which makes it public, which is where OpenSecrets gets it. So even if the Google janitor gives $250 to Clinton, OpenSecrets will list that as coming "from" Alphabet.

OpenSecrets also shows money that went to candidates from political action committees (PACs). But any money that passes through a PAC to a candidate must originate from an individual as well. The corporation can only pay administrative fees, like providing an office and Internet connection.

The reason that donors are required to list their employer on the donation form is because of how much influence that employer has on the employee/donor.

You may note that I said to avoid voting for candidates with a LARGE amount of contribution from Google. Because this isn't about $250, but, for instance, about the $58,000 Googlers have given to Ro Khanna, which pretty much ensures Ro would never vote against Google's interests if elected.

> But, in order for politicians to play the game they need large donors like Google. So, if they don't take money from Google they'll just take it from someone else with an agenda.

This is funny considering you think "donations" from Google don't have an agenda.

Lets call it what they are without double speak - a fucking bribe.

Realistically, you will never get a majority of US legislators to agree that giving US legislators money is a bad idea. Unsurprisingly, when Congress moves to increase the salary of Congress, the bill tends to pass with flying colors. ;)

Therefore, your best bet is to get regulations against Google by electing people paid by someone else, then later get regulations against that someone else by electing people paid by someone entirely different. It's a juggling act of corruption!

> Realistically, you will never get a majority of US legislators to agree that giving US legislators money is a bad idea.

While the electorate is bickering over distractions like this cycle you are correct. However, limiting special interest influence strikes me as a non-partisan issue, so if the citizens spoke up I think they'd have sufficient influence.. But, then maybe I'm being an idealist. Regardless, your short term solution works.

The problem is, it's a non-partisan issue in that almost all our representatives, regardless of political party, are totally comfortable with it. ;) And the reason we're uncomfortable with special interest influence is... that it absolutely freaking works. The power money has on election results is so strong, that almost anyone who isn't accepting it probably isn't getting elected. So the likelihood of gaining a majority of Congresspeople who would move against it is pretty much impossible.

AMP is essentially Google's answer to people creating terrible web experiences. It's been discussed and documented a lot. Common webpages today load content from dozens of different ad, tracking and whatever hosts, take several megabytes to load. All well known, but people don't stop doing it.

Now Google comes along and says: You can't do it, let us do it. Which is perfectly reasonable from their point of view. And when I surf google news on my slow mobile connection I'm always happy when I see that a link is going to an amp target - because then I know it'll be loaded fast.

But you don't need Google to do that. If you don't like AMP nobody stops you from doing the same thing. Limit the amount of stuff you load into your page, reduce the third party content that you include to a sane number of hosts (something like 3 instead of 50), optimize your javascript, deploy HTTP/2. None of that is magic and you can have your fast webpage without any AMP.

> If you don't like AMP nobody stops you from doing the same thing.

The AMP symbol tells users that the site loads quickly. Google doesn't indicate the same thing about non-AMP sites that still load quickly. Guess which site the user will tend to choose. I am not sure AMP will be optional for sites that want to get chosen.

I'm pretty sure non tech users don't even know what AMP is for.

They don't have to though. If there's a symbol and sites with that symbol are faster, they'll learn to look for the symbol, perhaps not even consciously.

But obviously this is only the case if they are faster. If people start saying "we don't like this AMP thing, we'll make our pages fast on our own" then it won't be a symbol associated with things being faster, it will just be "some irrelevant google thing". But if people notice that everytime they click on one of those amp things it loads fast and all other stuff loads slow they'll of course remember that.

This is fallacious as you're not differentiating between "my site" and "all sites".

What you're suggesting is not a matter of me making my website faster than AMP in isolation. You're talking about every non-AMP site in the world banding together and collectively bringing up the average speed of an internet webpage to be faster than an AMP page.

Without that, there's nothing to make my one faster-than-AMP site stand out from the rest of the slow non-AMP web in Google search results.

>They don't have to though. If there's a symbol and sites with that symbol are faster, they'll learn to look for the symbol, perhaps not even consciously.

Totally disagree. I didn't know what AMP is and I keep seeing it. I avoided it thinking it was yet-another Google adware property like DoubleClick, another cleverly named way to trick users.

For the vast majority of users on 4G connections, AMP loads aren't that much faster (When you download 8MB/sec on 4G, a 2MB page is just not a deal breaker) and I doubt that very many users at all in America are noticing that dramatic of difference on their devices, enough to subconsciously learn to like it.

Bandwidth is often not the limiting factor on mobile networks. AMP reduces the number of TCP connections your browser has to open (each of which requires 2 RTTs, and many of which will have HTTPS running over them which requires another one or two RTTs).

Take all those connections and introduce a 1-5% packet loss rate, and suddenly there's a very good chance that any given "page load" (really dozens of requests) takes perceptually "forever". That is what AMP helps with.

I'm pretty sure Google has extensively tested the CTR's of SERP pages with and without the AMP logo. There's absolutely no way they would have implemented it if it resulted in a lower click-through rate. That would run counter to their own stated goals with AMP.

Google tests and re-tests the crap out of things like this.

> Totally disagree. I didn't know what AMP is and I keep seeing it. I avoided it thinking it was yet-another Google adware property like DoubleClick, another cleverly named way to trick users.

But when you're rewarded by a quick page load for the first time, and the second time, you'll start clicking it more because you know what reward to expect.

They'll know soon enough that the AMP sites load faster. Even if they don't know what the AMP actually means. Likely they'll think AMP is this big, huge site that just has faster content.

Google also prioritizes AMP content over regular content in their news ticker/strip.

This makes a significant difference in traffic.

It would be great if Google could set some minimum standards regarding average page load times (which they track, check Webmaster Tools) that you could meet and get the same quick loading icon in SERPs. My site loads very quickly, but I feel pressure to go AMP because of the icon and ranking increase.

What's stopping Google from promoting fast webpages without AMP? I have spent weeks improving load times for my websites, with A A A A A ratings op webpagetest.org as a result. AMP is just a fraction faster, why should I have to create a seperate AMP version?

Also, my websites are essentially all dynamic content. People come for the comments, like on HN. I gues I'll have to use some cache-busting javascript if Google caches my content?

They do promote fast pages: https://webmasters.googleblog.com/2010/04/using-site-speed-i...

"we've decided to take site speed into account in our search rankings"

Sure, but then they put a lightning bolt next to any AMP page, even if your non-AMP page is faster.

I really doubt that that factor has any real influence on ranking. In my niche there are notoriously slow websites at the top that aren't even responsive in some cases.

Just like the promised benefits of HTTPS. I have yet to hear about a case where search traffic increased after switching to HTTPS.

The AMP Cache is an option that you can opt in to. The wordpress plugin by default opts in. The author said as much.

The problem isn't that Google "steals" traffic in the sense it's going against the wishes of the site publisher, but that the author didn't know what he was buying in to and thus was surprised.

There is the other thing which is the AMP Cache is probably more local and faster than your site. Small sites usually dont use CDNs so there it is.

No, it is not optional. If you publish a valid AMP page and Google finds out about it, it will be cached.


> Can I stop content from being cached?

> No. By using the AMP format, content producers are making the content in AMP files available to be cached by third parties. For example, Google products use the Google AMP Cache to serve AMP content as fast as possible.

So cache-busting Javascript will be needed, which will kind of defeat the purpose.

> Small sites usually dont use CDNs so there it is.

I think it's easier to use a CDN like CDN77 than to convert the average site to AMP.

Every time I look at a website its shocking how no one even does basic image optimisation using PS on the images.

This is such a trivial optimisation it makes you wonder about the level of competency of the average web development company.

=> "makes you wonder about the level of competency of the average web development company"

Our competency on projects where we are being paid a full domestic consulting rate is quite high, thank you.

Now, if you're asking us to compete with cut-rate bids from some online freelancer exchange, we will reduce the level of service and QA to match our competitors in that space....

Eg. you may have Nordstrom's service for a Nordstrom's price. Wal-mart service for a Wal-mart price. But don't expect Nordstrom's service for a Wal-mart price; that usually isn't going to fly....

Is there an alternative that doesn't require a monthly subscription?


You can insert ImageMagick somewhere into your production pipeline. It's excellent software.

Well other than being riddled with security holes. It has its place, but never use it on user-submitted images.

Well, the person I'm responding to is using Photoshop to resize images. Presumably that isn't for user submitted content.

While that may have been true in 2014, today it's hard to find an exploitable bug affecting a major image format parser in ImageMagick. Assuming you aren't using unsanitized user-provided parameters on the command line with it, it should be fine unless your attacker is very motivated.

There was an RCE vulnerability this year: https://imagetragick.com/

ImageTragick did not affect major image formats; it was a vulnerability in the parser for ImageMagick's scripting languages [1]. The real problem was that support for scripting was enabled by default, and there was no obvious big red button to disable it.

[1] Like MVG (http://www.imagemagick.org/script/magick-vector-graphics.php) and MSL (http://www.imagemagick.org/script/conjure.php)

I'd argue that you should simply sandbox it. If performance is an issue, throw it in a VM or container with a simple socat + fork + TCP-LISTEN on a socket and pipe data in/out over TCP.

You can get some pretty good images out of almost any image editing software, even GIMP. First, know which format to use, and size your image appropriately.

For JPEGs: use chroma downsampling, and don't go above 80% quality unless absolutely necessary. If your software has a preview function, use it. Encode with MozJPEG if you can.

For PNGs: compress the hell out of them. There's several tools that can compress them further.

Save For Web is a useful plugin for GIMP to this end.

ImageOptim is better than PS: https://imageoptim.com

You still need to resize your images appropriately but ImageOptim will do the rest. It passes images through a series of image optimization libs.

There is jpegtran and other cli based tools - but PS does tend to perform better.

If you are using a Mac I recommend this little known gem: Graphic Converter. It supports hundreds of image formats. First used it on System 7, such a long time ago! It has of course been ported to MacOS X / OS X / macOS.


Would you rather have an internet where every website has a design like an article from Wikipedia? Google is probably in their right to highlight pages using AMP (it's their search engine) but probably unfair.

It would be amazing if every website were designed like Wikipedia. With the exception of MediaViewer, Wikipedia is one of the best-designed websites out there. It's light, content-prominent, progressively enhanced, HTTPS-only. Even their mobile site is pretty good (though basically all mobile experiences are bad nowadays). What exactly is wrong with Wikipedia's design?

Wikipedia's mobile site is fantastic, IMO. I really wish all mobile sites were like Wikipedia's. The pain of browsing the web on mobile would go away completely.

I think it's mostly OK. A lot of my problems with mobile would be alleviated if I could get a browser to reflow text on demand. It used to be you could set your zoom level, double tap and the page would reflow to the new zoom level. This doesn't seem possible now, so I'm always scrolling side to side if I want to be zoomed in a bit more.

I have a few other minor problems about how Wikipedia's mobile page handles without Javascript enabled, but one big thing is that there's a minimum zoom level set. I can zoom in, but I can't zoom out, which means if I want more text per line, there's nothing to be done. This again feeds in to the lack of reflow in mobile browsers. I have no idea why this feature went away, but it completely ruined the mobile web for me.

Mobile Firefox does this fine. JumpGo does it, too, which means that it's certainly still possible in Android System WebView.

I use mobile firefox - how do you get it to reflow? I would love that feature if I knew how to activate it.

Similarly, I use mobile FireFox and would LOVE to know how to reflow text. That was my all-time favorite feature in any phone browser I've ever used, but the last browser I had access to with auto text reflow was the Silk browser on an original Kindle Fire.

How can I make this happen on mobile FireFox?

My mistake. I thought it did support it, but apparently not anymore.

I find it laggy when browsing to section anchors or more generally to big pages. Other than that, you're right, the mobile site is nice.

Maybe I am in the minority here, but I like to enjoy design, the placement of the content, the choice of fonts and custom article designs for featured contents, etc. My point is that efforts should be focused on optimizing the speed but still maintaining the richness and individuality of the Web.

Oh God yes please. So many sites would actually be readable. This is how you get things like Readability getting popular

not every site has SEO as its main traffic source

> If you don't like AMP nobody stops you from doing the same thing.

Yes and no. Google ranks AMP-ed pages differently, as well as showing that lightning bolt icon. It is definitely not a level playing field.

Diet pill mentality. People don't want to put in an effort to lose 25 lbs, they'd rather take a pill. Same here. People don't want to put in the effort to make something better, they want Google to do it.

I don't agree with this analogy. It's not like google is offering to redo your site in AMP for you. You still have to do the lightweight optimization yourself. You're just using a library that will guide you in how to do it by imposing restrictions like not loading from too many sources.

It's more like people don't feel like following a weight loss program until a trusted professional outlines an easy to follow and precise method.

The whole idea of AMP seems like it's really the wrong way to solve the problem. If you remove all the third party JavaScript, fonts, large images and "like" buttons you'll have effectively the same. I guess it already messes up ads and analytics (although I'd assume Google's services still work), so what's the problem?

Kind of related: I recently switched my blog from Wordpress to Hugo. I found a minimal theme, but the amount of junk it was loading was shocking. I created a stripped down version if anyone is interested: https://github.com/lucaspiller/hugo-privacy-cactus-theme

In the last 5 years there was NO WORDPRESS THEME OR PLUGIN that I didn't have to modify. I never saw any WP theme that comes with minified CSS/JS, handles sidebars correctly - sidebar created after post content, generated by PHP not JS, generates correct <a> tags, even one "modern" theme had links in sidebar with "nofollow"... "SEO-friendly" plugins that make blog generation time 3-5 times longer than before, of course JS not minified, jQuery and fonts linked with ?version=xxxx (prevents server and browser from caching it), social buttons that were downloading ~1MB of scripts and CSS (for Facebook, Twitter, G+ and Pinterest).

Whole WP ecosystem is a huge pile of crap, each plugin adds more and more mess to it. I tried Hugo and Joomla, but they were even worse (Joomla had a few security issues that developers refuse to fix, Hugo was more bloated than WP).

After trying and modifying over 20 acceptable themes I decided to start working on my own blogging platform that will be SEO-friendly and clean.

Agreed on the WP part. At the same time, when you make a WP theme from scratch, with Twig and without all the crap, your WP becomes surprisingly fast.

Hugo is a completely different thing, don't compare it to WP.

Why not compare Hugo to WP? That is what people do who want to create a new website.

They take a different angle, but the goal for both is to allow for easy maintainable websites.

I completely agree about most commercially available WP themes... but it's eminently possible (and quite easy) to roll your own very good theme, if you like the WP interface.

> Joomla had a few security issues that developers refuse to fix

Care to elaborate on this?

Is your blogging platform available as a service or do you sell the themes? I was looking for simple and clean website templates just a few days ago, but couldn't find anything.

Writing my own blogging engine that will work for my needs. Not themes. Not related to anything existing.

I think you are missing one point though - Having that AMP message with the lightning symbol next to your search result in the rich snippet as well as coming up as a result in the carousel at the top of the search results (search for some news terms and you will see what I am talking about) increases organic search traffic as well as CTR. Almost all of the media companies are doing it for that reason and actually making pages fast.

You have a good point, but how Google approaches this problem is wrong in my opinion. Why can't they just rank the page by the amount of junk it has in the first place, instead of actively looking for an AMP-optimized page?

If they were doing that, the media companies (and any other site which isn't in the deep web) would be working towards making their content minimal at the core, not branching off with an "optimized" version of the page. This would make the browsing experience better for all devices, not just mobile users.

Page speed is still a ranking factor. Optimizing pages and reducing bloat does have a positive effect in all SERPs. By my understanding though, AMP is more specific to mobile. A light page may or may not render well on mobile, but AMP pages almost always will.

I'm still undecided on if I like AMP or not, but I do see the reason for the distinction.

I also don't get it.

Anyone that wants to provide a responsive web site (as in fast to use and load) should stick to plain HTML/CSS, no need for JavaScript or "patterns".

It's like people who develop those ideas can't notice when they dug themselves into a hole, and just think all the problems can be solved by more digging.

In my experience, it's the terrible ad servers that are to blame. Run with an ad blocker turned on for a while and the whole web feels amp'd.

If you say to thousands of developers "make your site faster", you'll have very little luck. How do you make it faster? Which optimisation is appropriate to make? What about all the 'tech debt' our pages have accumulated.

AMP is a single 'framework' you can work with in making your site faster - it's just one thing to learn and reimplement your article pages from scratch.

This is of course ignoring the "ecosystem" capture Google gets as well from this, and the pressure/persuasiveness they put on media companies to implement AMP.

Just stick to PageSpeed/YSlow recommendations, don't rely on JavaScript (and AMP relies on JS).

There's validation, so you can run a checker as part of your deployment routine. Validation and routine checking is great.

I don't understand how validation requires the parts of AMP I dislike, though...

I don't (and never will) use AMP and I feel that it is patently unfair to penalize pages for not using some non-standard tech. Google should index the web fairly, not make use of their tech a factor. Webpage speed is a fair measure (and I really wouldn't know what I could do to make my pages any faster, AMP will not make a measurable difference).

I would expect an un-biased search engine to rate pages with and without AMP equal and to not show 'badges' based on whether or not a page uses tech by the same vendor of the search engine.

In my opinion, the problem is that your version of unbiased doesn't add much value to a user (and is actually biased for the website-owners). For example, SourceForge started adding a lot of misleading 'Download' buttons all over their page, and they might ask the same thing about 'an unbiased ranking' (Google's always been criticized for being the 'Web Police'). As a user however, I'd want Sourceforge's rankings go down the drain because it uses (directly or through its advertisers, doesn't matter) such a bad practice of getting users to click on the ads. AMP is similar. They're giving better ranking to pages which add value to the user, and are probably less concerned about the webmaster. In countries with slow internet, AMP adds a lot of value. You click, and the page is there (no waiting icons and 'downloading' for a few seconds)

(Edit: Sure, Sourceforge was improved a lot after the new owners, it was just used as an example)

I work with people who have limited Internet literacy, or limited literacy, period. I have seen a marked improvement in their experience when Google started to police the web more. Simply ensuring people can surf safely is a service to lots of people who do not frequent Hacker News.

The best solution to combat bloat is less bloat, not faster ways to deliver bloat (all that will do is increase the budget for bloat).

Why not both?

very interesting, so now it actually boils down to

- whitelisting (AMP bolt); OR

- blacklisting (down-rankings)

So now it depends on our perspective, is the web default good or bad? default good means blacklisting & vice-versa

Alternatively, maybe DDG can do a blacklisting of sites based on some criteria (bad-practice / malicious / content-farm / etc... )

> I would expect an un-biased search engine to rate pages with and without AMP equal

Well there's your problem :-). Google isn't an unbiased search engine, they're a commercial entity with a financial interest in feeding you particular search results. They're absolutely going to be biased in the results they provide you.

They also have near-monopoly status on search, so they must not be anticompetitive.

But it will take years for the government to do anything about it.

Longer, perhaps, if Hillary is elected with Eric Schmidt's help.[1]

1: http://qz.com/520652/groundwork-eric-schmidt-startup-working...

Why do you think that "person X supports candidate Y" implies "candidate Y will help person X commit more crimes without penalty"? Is that true of all people and all candidates, or only some?

I don't like the word "supports" for providing services at a high level in the campaign, but yes, I do think that in general, politicians attack their enemies and protect their friends.

An admittedly cynical view I've developed after years of watching partisan prosecutions, impeachment votes, and pardons.

> I don't (and never will) use AMP

You may intend this to mean that you will never consciously publish your content in AMP, but there is a good chance that you will consume content that is published in AMP.

It is already something I find myself noticing, especially on mobile. If you see an article work on mobile from a news site and it was fast, it was probably AMP.

Yeah, you can't really argue that AMP isn't a good product, from a technical perspective. AMP sites load instantly for me.

Does anyone know how to disable AMP?


You can't. That's probably not an oversight.

As mentioned–you can't, but I noticed that if you have an iPhone with 3D touch (6S or 7) you can skip the AMP page and go directly to the real website by hard-pressing the link. That brings up the preview modal thing, which you can hard press again to make full screen.

It's not your site any more. You're just a free content provider to Google now.

When I heard about AMP, I didn't get warm fuzzies. I didn't exactly know how or why, but something told me that the main benefactor with this technology was Google, not the user, and certainly not the content provider.

I can make pages faster without AMP by stripping the needless bloat and using a proper CDN. Why do I need this tech again, besides the artificial goad of page rank?

> I can make pages faster without AMP by stripping the needless bloat and using a proper CDN

You can. In theory. But people don't, as is evident with the web we have today. So Google comes along and tries to force people to do it. AMP wouldn't exist if people would've optimized their webpages on their own.

To be fair, as a user... I absolutely love AMP.

(That being said, I dislike it as a content provider and as someone worried Google has too much power)

That's your experience. In my experience, AMP pages are horrible and I avoid clicking on those.

Why don't you like them as a user? They load extremely quickly and you can scroll through the different AMP pages by simply swiping left or right.

For me, at least, it's being unable to load it in a background tab or copy the actual URL from the URL bar. It breaks the web.

Plus who needs to lose even more screen space to that stupid bar at the top plus the fact that the browser chrome doesn't hide itself when you scroll like it does on a normal page?

The scrolling is messed up and I cannot share the original link easily. And non amp pages load just as fast if I just disable js.

I haven't had a super awesome experience so far with it on my iPhone. I've seen a lot of results where I click on them and they spin forever, never loading. But if I long press and open the result in a new tab, it loads instantly and works fine.

Yep, in the short-term it's a good thing for users because it means they can access the content they are looking for more quickly in a consistent format.

In the long run, it might be bad for users, as it could result in media producing less valuable content, and one big company having more control over distribution.

But content isn't really valuable anyway. If there's less of it, then I'll make breakfast a few minutes earlier - that's all.

Since I've searched for the content, these are not ads but real content that match my search query which site owners have asked Google to render. If, despite matching my search query, the content is not what I'm looking for, then it helps me as a user to know that sooner rather than later.

What are these searches? I mostly see AMP on sites found through Google News (or other aggregators like HN).

If you're willing to put in all the work to do it yourself, and also pay for CDN (though it probably won't be as good as Google's), then feel free to do it.

But yeah, nothing's free. They give you this great technology, free access to their world-wide CDN, and better organic search results, at the cost of having the X button going back to search.

Though I have no idea how the author things that the "X" button should go back to the home page honestly... You open an article, you close the article. If you want to go to the homepage, you click the top banner with the logo.

>I have no idea how the author thinks that the "X" button should go back to the home page honestly

He's not suggesting the "X" leads to the home[0] page, but to the actual "deeplink" page (blogpost, etc).

I'm not the author, but here's my idea:

Normal webpage: click result in Google, page loads (for a long time), I click reader view, read some, close reader view, explore the site.

AMP page: allows me to skip the loading time & the click on 'reader view' and start reading directly (so far, so good), but then I close "(AMP) reader view" and want to explore the site. Instead I am back at square one, exactly where Google wants me to be.

[0] https://xkcd.com/869/

You hit the nail on the head. The cost to the publisher is added discoverablility of their other content and the ad revenue that comes from it (it also hits networks like Taboola hard).

Their success becomes ever more tied to Google for the next hit.

I also didn't understand this. If there's a quick view with a close button, where would it go but back to the search results?

It's pretty straightforward to provide a link in the page content that goes to the non-AMP version of the page. Multiple other news sites already do this.

But why would you ever want as a user to go from an AMP article to the same article but non-AMP? I could see the user going to the homepage of that site to see more similar articles (or clicking on other articles on the side), but I don't see reason why they would want to see the exact same content but slowly loaded.

And again, definitely not by pressing X. Pressing X means you're done with the article. And if you're interested, you'd press the websites logo at the top to go home, not X.

The same can be said about AMP. Google is not going to port your site to support AMP for you. You have to put in all the work yourself.

I think it was a way for Google to get newspapers and other publishers to stop smothering themselves in ads. Google has been preaching the gospel of the fast web for years now as publishers add yet another box for an ad server to fill.

Then Facebook started to court publishers and suddenly the best way to read a lot of web content was inside of Facebook which is obviously terrible for Google. So Google came up with Amp as a way for publishers to provide the same experience on the web.

> but something told me that the main benefactor with this technology was Google

Yep, and that something was that it was an initiative by Google.

They are not a charity. They would not do this is it didn't benefit them (or inconvenience a competitor, though Google seem to err on the side of the former rather then aggressively going for the latter) in some way. The fact that is offers benefits for others is secondary, or at best has equal footing in their collective mind (many individuals within the company may be more altruistic individually, but the company as a single entity won't be).

The author is regretfully realizing that in this day and age when so much of the discoverability of the web is dominated by Google, Facebook, and a handful other players, content is a commodity that these gatekeepers compete over for the privilege of serving as entry-points, and with it, getting analytics and ad network data. But on the plus side, it's a free CDN, if you're passionate about your content.

Actually realised last week I hardly use google search anymore. And cant remember the last time i saw - let alone clicked on an ad (clickbait excluded). Not that I've switched to another search provider as much as search isn't the entry point to the internet it once was (especially with no google on tor sites).

That techies like me dont rely on them any more and non techies all share stuff via facebook must be hurting their page and ad hits.

Makes me wonder if this is more a move of desperation on their part. Food for thought even if it is a little too forward thinking.

> That techies like me dont rely on them any more

You're going to have to cite some data for that bold assertion. Lot's of techies outside of your bubble still use (and rely) on Google.

In fact let's face it, half of being a techie is knowing how and what to Google

... or 'duck' a la duckduckgo.com

And knowing "how and what" to search is really teasing out of the index the magic combo of jargon, phrases, and synonyms through repeated refinements on your query. That approach works just as well on Google or ddg for me and the bang syntax on ddg for domain specific searches has really sped up some of my searches over what I used to do on Google.

I'm drawing a distinction between use/rely and how it was previously.

Does anyone still have google search as their home page? And if they do, do they use it?

When was the last time you got past the first page of results? (and btw, its so rare that that people do now did you know if you get past the first couple of pages it checks whether you are a bot).

Im contrasting how it used to be (reading google results making up a good majority of the time online) to how it is now (no more than 1 or 2 pages a day, sometimes not even that in a week)

You are wildly extrapolating from your own experience. I'll provide a counter anecdote.

Google is my home page, I use it so often I don't think about using it. I can't think of one thing I've searched for this morning. But opening my history reveals half a dozen things I've searched for. Most of the time I "knew" my destination but google gets there quickest.

And I'm not sure I ever remember spending a long time reading google results, for me google became my go to precisely because it was so quick I didn't have to spend long just get to the results.

It's true that social media provides a lot of traffic to places around the web but search is still huge.

I said "techies like me".

Thats not extrapolating anything.

Obviously much less experienced techies still need to look up how many sides a square has at least twice a day.

> I said "techies like me"

A data point of one isn't particularly descriptive.

While i'll take the compliment that you think i am that unique.

I'm pretty sure I'm not.

I'm pretty sure no one is implying any compliments.

Yeah, definately picking up a "omg he doesnt google ten thousand times a day" vibe.

Kinda guessed as much from the start (hence the - probably to forward thinking conclusion).

But probably worth pointing out. The reason i noticed I hardly used google any more last week, was because I did a search and realised it had been so long since i saw google search results on a desktop i almost forgot what they look like.

And yeah, i know that is because i spend most of my time now on places like hn, closed networks with all the technical manuals you could ever need on hand, specialist forums or tor sites that google doesn't even know exist let alone provide search functionality.

But i am definately not alone. And it definately looks like the future (or at least more like what the internet used to be). Infinately more valuable information. No spam bots or corporate advertising. Just experienced techies sharing knowledge, skills and resources.

But all that is irrelevent to the point that unlike historically techies arent earning google revenue and non techies really dont use google that much.

We went through a time when adsense completely dominated the ad market.

Yet there is another google "dependancy" I havent seen for a very long time.

And that i share with pretty much everyone techy enough to install an ad blocker.

> did you know if you get past the first couple of pages it checks whether you are a bot

I still go past the first couple pages somewhat regularly and I don't get any sort of CAPTCHA.

Could be its changed since the last time i used it.

for example. just did a search for "trump" on mobile and google search stops offering next after 13 pages.

> let alone clicked on an ad (clickbait excluded)

If you've clicked on clickbait then you are part of that problem, especially if you've clicked on clickbait but not other adverts.

What do you use instead when searching for something?

eBooks. The manual/documentation. Asking someone else working on a similar problem (who will then usually send over an ebook or other documentation I didn't have).

>That techies like me dont rely on them any more and non techies all share stuff via facebook must be hurting their page and ad hits.

The fact that you even think this makes me wonder if you're even a "techie".

It's odd, I feel like we're going into a Google vs Facebook webs. Each trying to control its part of the virtual world.

I'm really really curious about how it would be if we had "free" mesh networks under ipfs like content. It wouldn't need to be high capacity, just text, mail, json; enough to communicate.

If we had that, someone would come along with a flashier technology, content providers who could afford to try that technology would, consumers who want more would use that technology, and the web would evolve much along the same lines as it has.

Citation: The "web" used to be just text and <form>s; we were already there.

Wired broke this story [1] in February 2016, soon after Google announced it'd start directing results to the AMP Cache on the mobile site. In the meta-writeup the next day, Wired wrote "Google's AMP Is Speeding Up the Web By Changing How It Works" [2], while noting that this was a necessary step to compete with Facebook's Instant Articles and Apple's equivalent tech.

[1] https://www.wired.com/2016/02/google-will-now-favor-pages-us...

[2] https://www.wired.com/2016/02/googles-amp-speeding-web-chang...

    > I was expecting it to cause a redirect to the original
    > article. Instead it redirects back to Google search
    > results. Say What?
Well, what it does is exactly what I would expect; if clicking 'X' redirected to the author's site I would be thoroughly confused.

If I click on a link and it opens in a new tab, I don't close the tab ('X') expecting to go to the home page - I expect to go back to the page from which I opened it.

I agree with your expectation of clicking the 'X' going back to the results.

But it's a bit strange, why have the 'X' there at all when the browser back button will do exactly the same thing more naturally?

I imagine it's to enhance the feeling that the user just quickly popped something open over their search results - designing it to make it 'feel' like a different; faster experience.

Google wants you back in their serps as fast as possible and for you to notice that it's an option.

> Google was injecting a large toolbar at the top of the snapshot encouraging users to get back to Google search results (a functionality already provided by the back button)

I learned, through a series of usability tests my (former) startup ran amongst its users, that most non-tech people do not click the back button and get very confused if your pages don't have their own back/forward/close/menu navigation.

We moved our apps workflow from "use the browsers back button to go back" to having all navigation including back/close as part of the HTML UI.

Fascinating. Shows the importance of actually testing and measuring. I would have guessed the reverse, that everyone is used to the "back" button by now and almost nobody would bother looking for a needless extraneous one actually contained in the site.

I'd be interested in learning what these users do when faced with a site without navigation links. Do they just close the browser and re-type the original URL?

i question the validity of your test and/or it's application in the current context.

what I am hearing is "the average non-technical person incapable of using Google."

These people weren't incapable of using Google or computers. Some users were business-types who are used to using a lot of web-based SaaS tools and quite capable and comfortable using computers.

By non-technical I really meant they weren't developers; they were sales, marketing and management type people.

When asked, it simply didn't occur to most of them to simply press the back button.

Now, its not apples to apples, of course, and since this was in a web app where there may well have been a fear that pressing back would lose the page like you do in some (badly designed) single page applications, its possible that this skewed the results somewhat. Either way, we found that adding in-page navigation improved the workflow for almost everyone.

but how they use your application is not a reflection of how they use the web at large.

i'm not surprised by your study's result, I'm just not sure you can extend your conclusion to believe that people won't or can't click the back button to return to Google after clicking a result.

Oh. That's why Google has the insane "Google and X to back" behaviour. I guess they assume power-users will use the Chrome browser and newbies will use Google and need the redundant X-bar.

Why would I want to give up 10% of my screen space to a useless bar that when I try to exit out instead takes me away from the page? Complete UX madness.

I've asked @amphtml why they're adding it: https://twitter.com/mikemaccana/status/787968936251981824

The URL bar shows "google.com" in a lot of mobile browsers (including the one in your screenshot), the bar shows the original website this content is from

Trying to fix an issue that the AMP system itself introduced.

Thanks, that's a relief.

I think that this whole AMP thing is horrible for the web.

It's like Google is not dictating how one should build/style their websites. Thanks, but no thanks. Stay evil, Google.

Google has an eye on the next billion web users---the ones who aren't on the network yet because of cost and bandwidth. Their experience will be significantly bandwidth-constrained in the short run, and AMP is one of the initiatives to provide a segment of the web that works in such an ecosystem.

If the web were going in a direction to serve that demographic organically, AMP wouldn't be necessary.

That's an interesting point, especially since Facebook's strategy for those same billion users is to bring bandwidth (which happens to be preferentially dedicated to Facebook properties) to them.

But it's actively making the Internet worse for the current users. Is it worth destroying the current user base in order to attract a new one?

How is it making the Internet worse for current users?

Well do you really like the idea that when you go to some search result it effectively is opening an iframe to the website you're trying to go?

Or as a web developer you need to make sure to adopt to their stupid standards like renaming your img to amp-img and making sure it works in amp and non-amp so you'd be doing both?

I see it as complete bollocks.

> Do you really like the idea that when you go to some search result it effectively is opening an iframe to the website you're trying to go?

As an end-user or web developer?

As an end-user, I do not care. In fact, it seems that the AMP solution (which isn't an iframe so much as just serving the data from Google's cache at a Google URL) is faster.

> Or as a web developer you need to make sure to adopt to their stupid standards like renaming your img to amp-img and making sure it works in amp and non-amp so you'd be doing both?

As a developer, if I care enough about my user's experience to be bothering with AMP in the first place, I should be using a toolchain that makes it pretty straightforward to go from my meta-representation of my content to the render target (HTML / css, etc.). Because that toolchain should already be doing things like precompiling my CSS, condensing my image data, etc.

Ah yes, let's continue down the path of bloated web sites loading MB's of JavaScript. If you don't like it then just simply switch to a search engine that doesn't support it.

It makes really long and ugly URLs too.

Fortunately there's only one site I read regularly that uses AMP - Google News

Actually having the site served from the same, already-open, probably pre-fetched connection, already open browser is great for the user.

It even has the advantage of loading inside the Google App, where no adblocking exists, so it might even be good for publishers.

Yes, there's no discovery if you don't design for that, user will just bounce.

> having the site served from the same, already-open, probably pre-fetched connection

So, the bottleneck is really the connection setup time? Seems to me, that http2 largely solves that problem, while html already supports prefetch.

Don't kid yourself, the goal here is a better web locked into the google ecosystem.

http2+quic+proper support in the browser can probably help you here. The thing is all of this is available today inside the google{, news} app.

It might be locking you in the Google ecosystem, but the project is open source, and if there wasn't a monopoly in search/ads it wouldn't be an issue…

> Consider adding a link at the top of your AMP page, giving user an option to visit the original post

Why should they want to read the post again? The only viable option is to have links to other posts on your site, or maybe a link to the comments.

By the way, what's the purpose of the bar at the top? The back button of the browser already does that, even on mobile.

About analytics: I don't know if those accesses don't show on Google Analytics (I really don't) but what if one uses Piwik? Is there any way to get a report of the access?

> By the way, what's the purpose of the bar at the top? The back button of the browser already does that, even on mobile.

It's to counteract the fact that in some browsers on mobile, it will just show "google.com" in the URL bar, the added bar shows the actual website. (Obviously this is a problem introduced by the whole caching thing but that's why it's there I guess)

> By the way, what's the purpose of the bar at the top? The back button of the browser already does that, even on mobile.

Maybe they just want people to get used to this bar on top, and then start A/B testing it with ads and see what reaction they get.

> Why should they want to read the post again

Or if they want to copy paste an url to email (perhaps that's too advanced a feature...), it looks like a "let me google that for you" site, not what they're actually sending. Yay.

"By the way, what's the purpose of the bar at the top? The back button of the browser already does that, even on mobile."

It has a slide-out animation to it?

I assume it's an effect that plays when one taps the back arrow in the bar. I'm using Opera on my phone, no animation there, so I can only guess.

The are page transitions. Considering that Google controls both the source and destination page I'm sure they can make one without stealing space on screen. That annoys the user more than the owner of the content.

And if the user has a browser that doesn't support page transitions, it's only a waste of screen space.

This feels much more like FB's "instant articles" than what I actually want (as a user) from my web pages.

I still can't get over the fact that AMP breaks the img tag and requires JavaScript.

My long-running browsers instances are bogged down by JavaScript (and I have NoScript on whitelist mode, with it only allowed where necessary!). Requiring it on any fast-page standard seems counterproductive.

And why, why would they break the img tag‽

AMP doesn't "break" the img tag. It's more optimal to use the amp-img tag for AMP pages, but if you use img instead it's not going to break your page.

>And why, why would they break the img tag‽


As far as I can tell the amp-img tag doesn't add anything that img doesn't already support. The page simply mentions some vague statement about needing to understand the page layout, but uses the same attributes as img.

Did you click the support first-viewport preloading link?

embrace, extend and extinguish

AMP in general is rather gimmick and scam. Just normal regular HTML stripped from third party tracking JS--dozen of megabytes--provides pages of the same size. It's only a matter of clean code to make a page mobile friendly.

Introducing yet another additional markup format to serve mobile sites with the only purpose for Google to compete better with Facebook is excessive and additional dev overhead. But it favours Google's interests and Google basically blackmails content providers locking search availability to only those who obey. And additionally it gives Google excuse to serve pages via own domain with backlinking.

There should be just common practices for mobile code simplicity and not another new markup.

I wish they would kill it, personally. It makes sites feel and act weird different.

I get that this is frustrating, but this is exactly what Google said it would do. You can still track your traffic through analytics (per my GA setup, it shows the regular domain's AMP page, rather than Google's site), and you can still run ads against it if that's what you would like to do.

It's not so much a shock as it is a "new normal." Whether you like that new normal, that's up to you.

(For what it's worth, AMP causes a significant bump to my bounce rate, made up by the higher number of visits that come in via users looking for the AMP symbol.)

Well, considering that said traffic originated from Google, and that Google was under no obligation to send it to OP's page, it's a little more nuanced than "stealing". It just means they sent different behaving traffic than usual - traffic that only touches one article and returns to Google.

It might discourage webmasters from adopting AMP though, if they have the expectation to lead the visitor to the homepage or other articles.

I really love the speed of AMP pages, but I don't like the fact that the server request is proxied through Google.

This is akin to an ISP injecting their own little optimization toolbar into pages rendered to the user, except that cert pinning isn't even an option here, since Google is doing the content re-writing at the app tier. They even manage to keep the browser security icon green, creating the at-a-glance perception that the data provenance can be trusted!

If Google edge caching really is critical, then I'd rather see an architecture in which the AMP content is signed by the originating site, and the signature is looked up and validated client-side. This might incur a bit more of a performance hit for sites that are being overwhelmed, but really, that's as it should be -- as an end user, it's surprising that my AMP results are essentially a Google cache. And it would seem that Google goes to good lengths to make the look-and-feel appear browser-like, instead of calling out the cached nature of the response.

> And it would seem that Google goes to good lengths to make the look-and-feel appear browser-like, instead of calling out the cached nature of the response.

Why would a cached website be shown looking different from a non-cached one?

The fact that it's being served out of Google's servers changes the trust model for the content. End users have no good reason to trust Google with the cached data, but people assign trust to them when they click an AMP link. There's no real indication to an end user that that data is coming out of Google's caches rather than from the content provider. I find that misleading.

Typically, cached web pages come out of edge caches, either maintained by a CDN selected by the content provider, an end-user's ISP, or by a box on the local network. In this case, Google -- the search engine -- is fundamentally changing the nature of how the user obtains the third-party search result content.

IMO this is very different than my company or ISP doing some edge caching. It's different because it is much more of an opt-in sort of relationship, and, more importantly, because it's decentralized -- if Comcast starts to muck with search result content, Verizon users will notice the discrepancy. If Google takes advantage of their monopoly position and alters AMP content, nobody except the content providers will be the wiser.

> End users have no good reason to trust Google with the cached data

I'd imagine that end users who don't trust Google aren't using Google.

> There's no real indication to an end user that that data is coming out of Google's caches rather than from the content provider.

Users who care can see the URL and the domain for which the cert is signed. Most users do not care.

> IMO this is very different than my company or ISP doing some edge caching. It's different because it is much more of an opt-in sort of relationship...

I don't see why I can't substitute "Google" and "Bing" for "Comcast" and "Verizon" in your example.

For that matter, at least in the U.S., it's a lot easier to switch out my search engine than my ISP if Comcast starts to muck with result content. FWIW, Verizon already redirects failures of domain name resolution to itself. I don't like it, and there's not much I can do about it if I want Internet around here.

> If Google takes advantage of their monopoly position and alters AMP content, nobody except the content providers will be the wiser.

That's a fair concern, but the day Google does that is the day a media firestorm blows AMP out of the water as a trustworthy tool.

> That's a fair concern, but the day Google does that is the day a media firestorm blows AMP out of the water as a trustworthy tool.

And in particular, it looks (at a quick check) like the google hosted files are in a very predictable location, so it should be easy to write something to check both that version and the hosted version against each other to monitor for discrepancies.

> I'd imagine that end users who don't trust Google aren't using Google.

Except Google use to just guide your to the content, you just had to trust them to give you the best possible answer. Now you have to trust they have the real one too.

> Users who care can see the URL and the domain for which the cert is signed. Most users do not care.

Most users do not care if their passwords are stored in plain text, or if a bridge they cross has been built by a proper engineer. Consumer ignorance is not a valid reason for shady practices.

> If Google edge caching really is critical, then I'd rather see an architecture in which the AMP content is signed by the originating site, and the signature is looked up and validated client-side.

AMP's purpose is to deliver content at minimal bandwidth and latency. Requiring the client to side-channel to the originating server (that can be God-knows-where relative to the requesting client) defeats the purpose of creating a system for low-bandwidth users to quickly fetch and display content on mobile devices.

Sure Google has no legal obligation to send the traffic to the site, but it is still sort of dishonest. Google convices people to install AMP under the premise that it will help them. Instead, it just allows Google to monetize other people's content, at the expense of the creator.

It's hypocrital to bill AMP as an improved mobile experience, while sticking a big button at the top. Also hypocritical to penalize sites for scraped, dublicate content, while doing the same themselves.

Well, it does help them -- it promotes their content ahead of those without AMP, and it still allows ads to be served from an approved ad network of the author's choosing [1].

As I wrote before, it's "more of a mutually-consented handholding with small amounts of arm-twisting" [2].

[1] https://www.ampproject.org/docs/reference/components/amp-ad#...

[2] https://news.ycombinator.com/item?id=12299230

The approved ad network is the interesting bit. Quite a position of leverage for Google.

It helps just as much as paying the mafia guys for "protection" helps the shopkeeper.

But the user is expecting to go to the site. Agree w/ OP's concern.

The user is expecting to see the content of the site, not "go to" the site.

Most users do not actually care what the URL bar says. We know this because of the number of people who get to Facebook by searching for "Facebook".

Google is not stealing traffic but it is stealing content by displaying it out of the original servers control. The real cost of a website is not hosting it but filling it with things people want to read. Content creators will want to get benefits such as ad revenue to compensate for their effort.

If Google is "stealing" content by displaying it out of the original servers control then we should stop all of the major search providers from caching content. Obviously this isn't going to happen because it's considerably more efficient to serve the cached page from Google's fast servers and network.

>The real cost of a website is not hosting it but filling it with things people want to read. Content creators will want to get benefits such as ad revenue to compensate for their effort.

Google doesn't create AMP pages for you - the content creator does. If the content creator wants to insert ads then they can do so via AMP.


    > Google is not stealing traffic but it is stealing
    > content
If I'm an applemonger, and I send you out on a bike to deliver my apples, you haven't stolen my apples.

AMP is a terrible format and makes 0 progress for the open web.

HTML is already fast, it's all the extra resources added that makes pages slow. For publishers, the reason sites are slow is a combination of ad revenue pressure, poor tech skills, lack of time and focus on other priorities like producing content in a saturated market. This is changing slowly so that UX is more important but creating an entirely new proprietary system that only takes time away from the main site (and just affects mobile) is not the right answer.

More interestingly, the #1 most used adserver on the planet is Google's own DoubleClick - which means they could singlehandedly make all sites (desktop + mobile) faster by implementing better tech in their own stack.

Hm. The x button he mentions is only visible if you got to the page from Google search. If you follow the direct link to the amp page then the header just mentions his original URL as the browser (mobile safari). I would say that such behaviour is quite consistent with mobile so... whatever.

can you imagine if every site that linked to yours had a bar at the top of your page that said "back to facebook" or "back to wikipedia"? mobile web users know how to use the back button but google is taking up your screen real-estate to drive traffic back to google.

this reminds me of sites in the early web that used to link out to sites and frame them with a banner at the top. websites countered this with frame-breaker scripts -- but publishers really shouldn't have to do that.

Hugely agree - this feels like a dark pattern callback to one of the early web's uglier constructs.

I mean - running a "we'll host your page for you in exchange for injecting a toolbar" service isn't necessarily an evil thing to do. But there's no reason to opt people into it just because they're tying to perform well on mobile.

I suspect the toolbar is there to clearly identify the origin URL. As you can see, safari gives you no hint that this page is not actually google: https://imgur.com/a/2BtMX

As for the X button, I suppose google wishes people to use google search result like a news reader. I agree that it is redundant with the back button but the interaction does seem faster (I suspect that the page is actually loaded in an iframe on top of the results page)

They could just use a better domain name such as googlecached.com

> can you imagine if every site that linked to yours had a bar at the top of your page that said "back to facebook" or "back to wikipedia"?

I saw a site do this yesterday. It was so noticeable because I only permit same-origin JavaScript and the whole site broke (most actually work).

Then I realised there is an intermediary providing this service:


It is the worst of all things.

Linkis is twitter cancer


You can also look at the source from an linkis infected link to find the original url

you can detect if your page is running inside iframe and if so, make the browser to load only your page, to get rid of services such as linkis.com

but does this work with AMP pages that restrict the javascript?

AMP is "HTML with some restrictions".

Speaks nothing of HTTP (otherwise they could've done the right thing and added a "text/html;amp=1" content-type).

In theory this should still work: https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/X-...

But I doubt that Google's AMP cache will respect existing headers (the very nature of the cache stuff means that they won't, but what they will do to non-cache headers on an item is undefined AFAIK and so they will probably just strip them).

a quick look at an amp page (from google's cache) shows it's being served with x-frame-options:SAMEORIGIN, so looks like the linkis thing won't work with them.

parent comment mentioned service such as http://linkis.com/ what i was reacting to. i doub't that you can use linkis.com with AMP

That sounds a lot like the trick that Digg used to pull... People didn't like it one bit. It might well have been the final nail in their coffin.

The fact that you can not go to the original page from the AMP page bothers me so much. I go as far as avoid AMP pages altogether just because of that reason.

Aren’t you supposed to put links to your other articles at the bottom of that AMP page? I mean, that seems like the best way to get users to visit your full site.

A tech lead from Google wrote his response, if anybody is interested:


I assume Google ads that are served through AMP will still count. In that case, you can think of AMP as a high-speed static host where you still get paid even though you didn't pay for hosting. But I agree it's very surprising behavior to the actual owner of the website.

The usability downside of providing a link to the actual webpage should be obvious - Google is trying to pretend this is the actual webpage. Why would they want a redundant link just to confuse the users?

That said, ever since Android separated google results view from chrome and added that strange "x->back" button thing I keep getting tripped up from a usability perspective. That's more the Android team being silly again, not just AMP, I think.

> Explore the site further OR hit the back button to go back to Google search results

I set Google to open links in a new tab anyway so the X the OP is complaining about is actually exactly my browsing style to click X and go back to search results.

Why doesn't the author just measure how much money they are making? They seem to think some things are bad (header bar, etc) but as far as I can tell they could find out very easily if using AMP is helping or hurting.

I work for several large newspaper publishers. We have been getting pushed to get all of our content optimized for the AMP experience as to not get left behind. As stated by others though this is good for the end consumer, not so much for the newspaper industry itself.

One of the biggest gripes is getting our paywall model onto the AMP sites, as we have very little input as to how that is done. It also takes ages to hear back about requests/suggestions with little feedback as to why they think its a bad idea.

The web is not your personal money generator. It's a place I go for information. I'm sorry that your fellow developers have made it such a user-hostile place that projects like AMP that defend the user experience make it harder for you to make money. But maybe if you're only interested in publishing for the money, I'm not particularly interested in courting your continued participation.

Posting my comment from the article here:

Hey, this is Malte and I am the tech lead of the AMP Project for Google. While I work on the AMP open source project, I did check back with the Google Search team that is more directly responsible for most of the points mentioned in the post. I personally find it very important to respond, because “stealing traffic” is literally the opposite of what AMP is for. The original idea behind AMP was to allow content to be distributed to platforms (such as Google, Twitter and Pinterest) in a way that retains branding and monetization control for the publisher. AMP traffic is the publisher’s traffic. Period.

I also realize that “just turning on the WordPress plugin” doesn’t get you there. Especially if a WordPress installation is heavily customized, one will need to invest similar effort to get the AMP pages to the same quality. While this may be a lot of work, this is by design: We recommend to really optimize AMP pages and fine tune them to your needs. AMP is not a templated format for that reason. While neither the AMP project, nor Google are directly responsible for the WordPress plugin, the AMP open source project working closely with the authors of the plugin(s) to improve the quality and scope. AMP is very flexible and should be capable of providing most features of a typical WordPress site, but this flexibility also requires respective work to make custom plugins and development show up in the AMP version.

Getting more literal about “stealing traffic”: there are audience measurement platforms that attribute traffic to publishers. They might in theory wrongly attribute AMP traffic to the AMP Cache (not Google) rather than to a publisher because they primarily use referrer information. That is why we worked with them in worldwide outreach to get this corrected (where it was a problem), so that traffic is correctly attributed to the publisher. If this is still a problem anywhere, AMP treats it as a highest priority to get it resolved.

“Ask Google to give users an easy option to view the original post.”

Let us start by saying that we love URLs as much as everyone else, and we tried hard to make the AMP URL scheme as usable as possible given the technical constraints of web apps. We’re looking at ways to make the source link more discoverable and will update once that is done. AMP is super flexible in terms of how a publisher can direct traffic to their site. Typical ways to get to a publisher’s homepage (like clicking the logo) should just work and are in no way restricted. Also, make sure to check out amp-sidebar (https://ampbyexample.com/components/amp-sidebar/) for adding a menu to your AMP pages.

If you are not comfortable with traffic on your AMP pages, please do not publish AMP pages. Google Search has 2 types of AMP related features:

Normal search: AMP does not influence ranking. Your pages will appear in the same spot with or without AMP. AMP specific features (such as the “Top Stories Carousel”): For these features, we believe that AMP is the format that currently delivers the best possible user experience on the mobile web. That is because AMP allows for consistent speed, caching, pre-rendering, and enables swiping between full-length pages. This is a big deal for topics where there isn’t “that one best result” that a user might want to look at.

“Google takes away ad revenue on AMP pages”

AMP supports over 60 ad networks (2 of them are owned by Google) with 2-3 coming on board every week and makes absolutely no change to business terms whatsoever. There is no special revenue share for AMP.

“If Google cares so much about the mobile experience, why cover 15% of the small mobile screen with a fat bar at the top?”

The Android users might have already noticed that it is now scrolling out of the way and the same is coming soon for iOS (we’re just fighting a few jank issues in Safari). Similarly we’re spearheading a long term effort (https://github.com/bokand/NonDocumentRootScroller) to allow web apps to define how the address bar is hidden on scrolling. It looks like this will land in Chrome soon, providing even more space to web pages.

It's been known all along that Google AMP serves pages from its CDN and that Google adds a bar at the top. The author would have know as well if they had paid attention from the get go instead of bragging about their 5 minute "install". Nothing to see here, move along.

Yes there is something to see here: Google doesn't provide a way to reach the original site from the version they serve from their CDN.

You can click on any of the links in the article. It will get out of the Google AMP session, and open them in a browser. You can have your own site header and whatever series of links pointing inside your site, like you would on your regular non-amp landing page. This author is whining about a non-problem.

The new Google News interface has done this for a while and I absolutely hate it.

I don't know if it's because I'm using an iOS ad blocker or the sites are misconfigured, but many AMO sites simply fail to load on my iPhone. AMP has been an awful experience for me, as an end user.

"Cache-Control: private"

Google has to be respecting Cache-Control headers, right? Set your AMP pages to return that. Then they won't be allowed to cache them.

I've been wondering if AMP isn't an play to actively decreased app installs. You can't include a Smart App Banner in an AMP page.

My biggest problem with AMP is I can never go directly to the source. AMP has buggy scrolling and can be quite annoying sometimes.

long press on link and click open in new tab on search

thanks. but it doesn't work.

I think the next step for Google is to provide a free hosting service, where ad income is shared by Google and the content provider.

I agree, since that's pretty much what AMP is doing (free cdn/hosting.)

Seeing analytics built into AMP makes me laugh. Does that not go against producing the most enjoyable user experience possible?

No because you can have both at the same time? When was the last time you noticed there was analytics happening on the page you were on?

Would a robots file block google from caching your site? I guess that would block google entirely though..

Not the AMP version. The AMP version is there for Google only, so if you want to prevent them from caching, simply don't serve an AMP version.

I ran into this but didn't know what was going on. Seems kind of sleazy.

"may be"?


We've banned this account for repeatedly violating the guidelines. Profanity is not against the rules, but we do ask that commenters post civilly and substantively and this account has done neither.

Please try to refrain from using profanity. It doesn't add anything to your point but mere shock value and it gives a bad impression of Hacker News to visitors.

That'll show them. Cut your nose off to spite your face.


Try using www.yandex.com as your search engine, it is surprisingly good, and has less ads disguised as "search results" and results from blog-spam sites and such nonsense.

I can't wait 'till the author finds out that sometimes the browsers cache resources... These terrible browsers stealing his page impressions.

Are you implying this is the same thing?

So you installed a WP plugin without doing any research and hoped it would make your pages faster by using unicorns and magic, or to quote

> Most importantly, I was surprised to find out that instead of redirecting users to an optimized version hosted on my server, Google was actually serving a snapshot of the page from their own cache.

and now you're upset that they really only made it faster via caching it, not actual magic.

Were you perchance born yesterday or are you just very naive...?

Your comment is very similar to victim blaming.

Applications are open for YC Summer 2018

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact