Hacker News new | comments | show | ask | jobs | submit login
Response from Google Tech Lead, Re: “Google May Be Stealing Your Mobile Traffic” (alexkras.com)
342 points by akras14 340 days ago | hide | past | web | 195 comments | favorite

This is a good response and I appreciate Malte's effort to engage publishers, but it doesn't calm my nerves about the biggest publisher complaint about AMP:

Google owns the chrome (top navigation) on all AMP pages. This makes it easy to navigate back to Google (why would you ever want to leave?) and other publishers in the AMP ecosystem, but much harder to navigate to the publisher that created the page in the first place

In essence, this means that what was once a publisher-owned page is now shared property: between the Google and the publisher. By controlling the top navigation, Google more easily controls the content the visitor sees, keeps visitors on Google longer, provides greater opportunity to track visitors, and perhaps most importantly has the opportunity to earn more ad revenue.

Now imagine if this was a requirement for ALL pages served in Google search results. You publish a page and it appears in Google, but when the user clicks on it Google has pasted a new navigation on the top of your page. This is exactly what is happening with AMP.

This is especially troubling in light of all the anti-trust controversies Google is finding itself in, both in the US and abroad. A recent study showed that 49% of all Google clicks go to Google properties of one kind or another (Maps, YouTube, Ads, etc) http://www.slideshare.net/randfish/intro-to-mozcon-2016/24-L...

Does AMP count as another Google property that will push more than 49% of clicks their way? Hard to say, but it's a disturbing trend for a monopoly and a hard pill for publishers to swallow.

It's one more step along Google's path to eating the internet. Google-served content discovered through Google search rendered in Google's browser running on Google's operating system on a Google-manufactured device connected through Google's ISP and Google's network protocol (QUIC). Vertical integration.

You know the cool thing about this integration ? Any part can be replaced. Some instantly, some take more time because they involve physical stuff. People use the Google integration because it's there by default, not because they are locked-in.

At the moment yes you are right, but if we want it to stay that way we need to do something about it now. I agree that people mainly use Google because it's the default, but if we aren't careful there won't be an alternative left that can stand up against them. The only hope we'll have then is that Facebook, Apple, Amazon or Microsoft decide to tackle Google head on, but in most cases where that has happened in the past Google has won.

I'd already say that in terms of publishing online that Google is the only search engine publishers/marketers care about.

With search, we already approaching that point.

No other search engine comes close to the quality of Google, at least for certain topics.

Outside of Russia, China, Japan and some other countries Google is the dominant search with a big lead.

And that is troubling indeed! Competition is good. And necessary for a healthy market.

It also means that Google has a huge amount of control over what information can be discovered.

What if MS decides Bing isn't worth their while anymore in 5 years? Yahoo will have dwindled away by then, leaving only Google. In Europe, Google has +90% everywhere (except Russia).

And that's the point: quality. Altavista was a good search engine at it's time and now it's just memories. Because a "better" search engine appeared, giving results that were more useful.

If we continue down that road as you worry, at some point regulatory action might be appropriate. Google already seems to be breaking itself up via Alphabet, maybe they should just be cleaved entirely to avoid collusion and encourage competitors. On the other hand, the gov't avoided breaking up Microsoft back in the day, and this seems to have worked out OK as competitors sprung up into the gap the gov't wedged open anyway.

Bing is still using Google as its backend.

www.yandex.com is a really good alternative, I found it even offers better search results for most topic I look at.

Bing doesn't use Google as its back end ... I don't believe it ever did.

Yahoo did switch to using Bing as its back end, some years ago.

Not that I agree with your parent comment that bing still uses google results on the backend, they did at least used to: https://googleblog.blogspot.com/2011/02/microsofts-bing-uses...

Between Facebook, Apple, Amazon, Microsoft and Google, who do you want to be the overlords of the web? Each one of them is trying to win total dominance and eventually one will. Google seem like they would be the least evil of the five.

The things is, this is exactly whagt their marketing department has been pushing us to believe since their inception.

The motto, the glibly pedantic color schema, the focus on simplicity on their home site, and cleanly (yet viciously) integrated ad revenue, even the language of the letter in the original link, all of these things point to a naivete that ignores how corrupt the ethos of what Google's central argument is: meritocracy above all else, in other words: down with the dumb, down with the lesser, all hail the perfectly manufactured data stream that contains what seems like childlike wonder, but is instead an amalgam of people's minds. They are sucking the intelligence out of us and selling it back at a markup so high it begs disbelief.

The fact that there is Bing is good, adn Bing's search isn't all that far below Google's in terms of getting good data from the web, but the point is: aren't we forgetting that there was a loooong time when we derived a lot of good data without their services? I find the aggressiveness of Google's ignorance to be quite frightening. I find that they are so far removed from the effects of their casual relationship with owning what happens to the data they are "curating" that it seems anathema to good governance, and governance it is because we are subject to the whims of their algorithm's ability to provide us with data that goes beyond what the mind can handle and handle sanely.

There is a monumental schism between what the web is comprised of and the first page of Google's query results, and that is what I think is at issue. What about all of the rest of the data that we never see? How do we know it's not interesting? How could we ever spend the time to sift through it to make certain we aren't missing data that is far easier to comprehend?

Long post, but a quick example: I program a fair amount. I like C++ for various reasons, but the C++ reference page is dry as the Red Sea after Moses played the downbeat of the 40 year cadence the slaves danced to after raising one too many pyramids.

I would wager that there might be a great C++ reference page out there somewhere that has easy on the eyes detail, a fairly conversational tone and nice examples that aren't at the very end of the page. But, because Google's algorithm (and Bing's, Yahoo's, Duck Duck Go, etc) all get tons of traffic from the cppreference.com page, it or some Stack response is the go to answer for essentially every question for which I need an answer. How is that well crafted page ever going to get the right recognition if everyone is being diverted to the cppreference page? It's like the old trope, we do it that way because that's the way we've always done it. Sometimes this means we drive the speed limit because that's the safest way to drive according the engineers who calculated the road conditions when they were built, but it's a little different in this instance because it's the _quality_ of the road that matters, not merely the velocity we drive that dictates the "goodness" of the experience.

Google (and the other big bots out there) are ignoring the quality in their algorithm(s) and merely settling for popularity as the guiding rubric, and I think we are all too aware of how that mentality can lead to unsalient choices.

They've already made it hard enough to make sure legit email ends up in the inbox of a gmail-recipient, rather than being silently flagged as spam -- that anyone hosting their own mail needs to consider if they need to special-case delivery for gmail recipients (I've considered routing gmail-destinations through authenticated smtp to gmail via a gmail-account - then realized that that was just silly and I feel better when gmail doesn't work -- but I don't depend on delivering my (private) email for revenue). Add to that the fact that google is one of few big email providers that are hard to grey-list (or rather automatically white-list once delivery has started) -- and they've already made a great effort of siloing off one of their arguably most open services.

If you think the megacorp that doesn't do customer support is going to be any easier to dislodge at other levels in the stack, once they manage to squeeze themselves in by undercutting the market, I'm afraid you're hopelessly naive.

Once they start to feel confident in their domination, there's nothing preventing them from going the Microsoft way, i.e.:

1. Introducing incompatibilities, forcing closed standards etc. which block any potential competition.

2. Stop spending as much on product development.

I want to Android as an example to prove otherwise.

While yes, you can install CyanogenMod (without Google Play Services), or others along those lines that remove the "Google" out of Android, it's true value (for most common users) is Google Play Services.

While you can say that you can get apps from apkmirror or F Droid, most apps still won't work without Play Services (case & point - Snapchat)

my entire point being - yes, it's very easy to have alternatives (and their importance can't be ignored), Google services (be it Play, or Gmail or others) are what people are looking for.

And people can just buy an iPhone instead. Windows phone is still chugging along.

Define chugging along. I was under the impression it was broken down at the side of the road and moment away from being abandoned.

Yes, alternatives inside the Android ecosystem are more technically doable than easily doable.

I was more thinking of other operating systems: we are fare from lacking functioning alternatives both on mobile and on desktop.

> You know the cool thing about this integration ? Any part can be replaced.

How do you replace search ? How do you get competitive against mail, youtube without the amount of data google uses to sell ad on them ?

All I hear is a challenge. I'm not a big data engineer, but I would be interested in seeing others take it as a challenge to develop a unique, effective, and open platform that rivals the Google monopoly.

AT & T and Standard Oil were also "challenges" and they weren't toppled by "unique, effective, and open platforms" :)

Yandex is pretty good, they have search and a better mail than gmail, just try it out, you will be surprised. I was.

There was NO ads! You can actually read your email - and NO ads! That is amazing!

The problem is that Google is also very good in lots of areas. This makes the challenge even more difficult.

There are existing alternatives, today, that are profitable. What makes you think it's not possible to use something else ?

There's a part that already can't be replaced while keeping all the others. Google Fi only works with Nexus devices. It could be a lack of OEM interest, but who knows what Google does that might keep it that way.

Google Fi is an MVNO, one that entered into an area with arguably insufficient competition. Regardless of how you might rate the level of competition before Google Fi entered the market, it has definitely added ambitious customer-oriented features.

Its device-dependence exists because it's a unique dual-provider MVNO (Sprint and T-Mobile). This allows them to broker the best deal for the end consumer. But the bar is high -- devices must support dual-SIM, which is rare. Devices without explicit support aren't barred from Google Fi but may not get the full features and you can't get help with those devices if you run into issues.

It actually is now Sprint, T-Mobile, and US Cellular.

or any device that accepts a nano-sim. I know several people who use a google-fi sim in iphones

Absolutely not.

Google is just sending a different type of traffic: "one page traffic", but let's not forget that it is Google that is sending the traffic to AMP pages in the first place, there should be no entitlement on the part of webmasters.

I'm only half sure you're trolling so I will point out for others that lost what I hope was your irony and not naivete: Google only exist because content creators allow their bots to index and display part of the often copyrighted content for Google's own monetary gain, so Google should feel no entitlement on owning anything on the Internet.

They don't feel entitled, they just don't care. Just like they don't care with all illegal content on Youtube. They didn't care when they started hotlinking images in Image Search.

Just like Facebook doesn't care about ripped video's on their platform. They know that publishers won't bite the hand that feeds them. They know that they can use their power as the defacto gateway to the internet as there is no competition and noone powerful enough to stand up to landgrabs like this.

Maybe it's even worse. They don't stop and think about the power and influence they have, because over the years it has become entirely natural to them to just take whatever they want, repackage it and call it their own.

What would Google be able to show it's user if it weren't for this greedy webmasters who poured over their sites to generate content?

What comes first? Hen or egg? Google isn't the internet, the web was there long before Google made things findable. You had to know how and where to look. And yes Google made things back then way better.

Non the less without content on the web a search engine is just blank. Unnecessary imho.

So Google making my content a shared property is something of concern to me. Therefore I am not using AMP.

Well, the hen and the egg came first. The egg (the content) was there and the hen (Yahoo) was already making it all findable.

This matches my own experience using AMP via google search. As a _user_ the experience is excellent but it pushes me toward using content in a very shallow way. If I were a content _publisher_ I would be thinking very carefully about the cost this has on retention.

Ill jump on the top post for now, and recommend everyone to use Yandex Search. Yes that is right. Not duckduckgo, because that one is anyway using Google in its backend.

Yandex is surprisingly good, take some time to try it out, whenever you think "oh shit results" and want to switch back to Google search, remind yourself how often you automatically rephrase your search when you dont find anything on Google either. Rephrase search, try again, you will find better results than what you had with Google.

There is less spam/ads, and you will discover whole new websites, and network of websites, some even foreign sites but in English.

It feels like using Google from 2004.

www.yandex.com and no Im not payed by them.

I don't think DuckDuckGo is using Google.


In fact, DuckDuckGo gets its results from over four hundred sources. These include hundreds of vertical sources delivering niche Instant Answers, DuckDuckBot (our crawler) and crowd-sourced sites (like Wikipedia, stored in our answer indexes). We also of course have more traditional links in the search results, which we source from Bing, Yahoo, and Yandex.

Oh that is nice, was not the case last time I checked duck.

Still, duckduckgo is an US company, better stay away from those given the history of last decade.

So better go with Yandex, the Russian company?

Potentially, unless you happen to be in Russia... clearly if you live in say, Belgium (or Singapore, or New Zealand, etc.), it's unlikely that any Russian agency that has access to your searches will share that data outside their borders. Of course it's generally your service provider that you have to worry about.

Yes, if there’s multiple international companies competing, that’s better than one (even if it might be morally good) company having a monopoly.

As I understand it, they've never used Google's index.

> It feels like using Google from 2004.

Did they bring back the + operator? I still miss it even if you can use quotes instead...

To be fair, the user coming from a search engine is most likely to want to navigate back to Google than to stay on the page. It might not be benevolent but it is an improvement to the user experience of most people.

That's why we have back buttons, no?

Most (if not all) default browsers on mobile, auto scroll/hide the navigation bar off the display after a few seconds. For most non-techie users this means it's 'gone'. Ergo, they have no back button anymore. I've watched many people just open/shut/open/shut their mobile browser just to get back to previous pages - in many cases even re-searching the same query over and over again.

This is modus operandi for many users, hence why Google inserting a big 'go back to Google' banner at the top, is a bad[1] thing.


[1] Or from the standard user point of view, a good thing.

But this is just an iOS thing, right? Android has either a hardware back button, or a software back button that's always shown (except for fullscreen apps, such as Youtube and games).

Software buttons hide when using full screen apps and can be made visible again by a swipe from the bottom.

On Samsung Android the back button (↩) is always on the bottom right of the device (below the screen)

For the browser that is back, and if you opened the page in a new window then back means close it and reveal the previous page.

Maybe not everybody realizes that and uses it. I didn't at first but now it's a happy habit.

Isn't Google in charge of Android? Seems like they could fix the button-hiding there instead of adding an HTML workaround for every AMQP website.

Why? The whole purpose of a search engine is to bring you to the result site, not to search forever.

Because not everyone finds a useful page given their query from the first result they check. I often check between 2-4 results for some queries before I find what I want.

Also, given that it's hiding during scrolling now (or soon, depending on your platform) and you'll supposedly soon have the ability to define your own AMP menu, I think it mostly should calm the GGP's nerves.

Then this is a half-baked solution to the problems caused by making tabbed browsing hard.

On the desktop, I will open a bunch of tabs and close the ones that don't pan out. Not just after a Google search, but after anything resembling a search.

Mobile browsers allow this (though slightly awkwardly), but most mobile apps which display web content don't allow anything like this.

Do you think that it how most people search? (I'm not being sarcastic, I doubt they do but I obviously don't know)

I think it is the way most people who care about trying out multiple hits do it. For everyone else, the back button is good enough.

Given that landscape, what is added by a fat page wrapper that makes it easier to go back to the search engine?

But how much of that is in response to slow load times, and optimizing viewing the results without waiting? If AMP is fast, then whatever portion of the reason to do this was caused by slow loading is significantly reduced.

Also, in regards to your up-thread comment, on mobile browsers I will sometimes use that technique because (but only because it can be slow to wait for multiple to load), but doing so is generally more cumbersome. Hold on link, select to open in new window. If I could just quickly visit each as needed, returning if it wasn't what I was looking for, I would much prefer that (as opposed to on desktop, which while I would like quicker loading, I do still occasionally want to open multiple tabs for results as you suggest). That is, I don't think this is a half-baked solution to a problem caused by making tabbed browsing hard, I think it's a solution that addresses some of the reasons we use tabs in the first place (specifically, to parallelize loading times for possibly slow sites).

The purpose of a modern search engine is to find and deliver the requested information with the best experience possible. That does not always mean handing the user off to someone else.

They read that post and if you're lucky they stop and skim through another one, but that's not the one they came for so it doesn't interests them as much. Same thing when you go to a post through HN. Do you stay on that site or come back to HN?

> This makes it ... much harder to navigate to the publisher that created the page in the first place

What makes it hard? People click on links to navigate around the publisher's web site, and you can still have links in the AMP landing page.

All these navigation elements are secondary, and further down the page. You have a point that if the user wants to, they are free to click any link they want. The fact remains that the top-most, primary navigation element on an AMP page is an "X" that closes you out of the publisher's page to send you back to Google. After that, the second most prominent navigation element invites you to browse other publishers. By design, the publisher's own links must come after these two elements.

Yeah - that's really, really shocking to me. I paid attention to the AMP project early on, felt like I had absorbed its value proposition, and made a mental note to check on it from time to time.

I wasn't expecting the examples of its usage to involve altering the UI on AMP pages!

Is it just part of the blessed AMP WordPress plugin? Or is google actually injecting their own UI into AMP pages by default?

If the latter, that would seem like a power grab. Sure, it's "just a toolbar." And search results and adwords ads are "just links."

> Or is google actually injecting their own UI into AMP pages by default?

Yes. Google doesn’t actually link to AMP articles, but caches them on their own servers, modifies the markup, or removes non-compliant ad networks, and serves that to the actual user.

Oh. Wow.

The web is supposed to be a platform, and AMP is supposed to be an open standard ... but the end result of this is that Google Search will modify your web application with their own navigation UI?

Yikes. I mean, injecting your stuff into other people's web pages is bad when a cable provider uses it for ads. Is it not bad for the web platform if it's google that does it?

Or is it linked to your usage of their CDN, and not something that necessarily comes with being an AMP site? (Ie, google will cache it for you, but first you have to agree to let them inject your page). That wouldn't be so bad, because it'd mean that you could play in the AMP playground without agreeing to Google having their fingers in your users' navigation.

But if AMP means "Google-injected UI" right now, then that just seems bad.

You can have AMP without Google-injected UI, but only if you rename it, or don’t allow Google to index your page at all.

If Google indexes a page with AMP, and returns is at as a search result, it NEVER actually returns the AMP page, but always the modified one from the cache.

Google owns all the Chrome on Image Search too. And that didn't stop them from hotlinking to the images on your server. There's no way Google is backtracking on this because of the content owner's concerns.

Maybe they'll make an exception for Germany and France, where the old image search is still live.

FYI: It's long dead in France.

The extra chrome appears to be displayed in browsers but not in the Google app. On Android it's more natural to search from the app; I don't know what the situation is like on iOS.

I still don't understand why it's mandatory to use the Google CDN, but I probably don't want to know.

According to AMP documentation anybody can build an amp cache. But probably nobody did so yet. Also google will most probably link only their cache from the results. But it seems possible that you could build an AMP CMS that would use the advantages only for you, although I do not see much interest over well written HTML.

I agree. You'd have fast-loading pages - but you can get that by just writing decent HTML. What you likely won't get is Google's preferred handling in search results.

The one upside I see is that Google-moderated AMP pages might push ad networks and library developers to develop less bloaty solutions to stay compliant. Those solutions might make it easier to develop regular non-bloaty pages as well.

I think that it is easier to develop a non bloated page even today. But the problem is that the competition got extremely fierce and the ad networks absurdly data obsessed (without much proof that the targeting works).

I wonder whether we can get to a sane level now. Consumers do not seem to be prioritising sites with less bloat (otherwise there would already be a trend for leaner pages). Those who care probably use ad blockers and aren't the target demographic anyway.

There were some search engines that displayed their results in an iframe rather than by going to the link directly. This was a very bad move and angered users tremendously. It's interesting to see how Google is trying to revive this concept through the backdoor of providing the user with a faster loading page. What with mobile being set to be a larger portion of visitors to websites at some point in the future this makes Google effectively the only real website on their view of the web and the rest of us chickens as information providers to Google, to be terminated at will.

Digg tried the same thing, back when Digg was a thing. It didn't go over well, but they didn't have nearly the same power that Google does.

If you aren't comfortable with it, just don't use it, right? I'm probably missing something, but I don't understand how something can be such a hard pill to swallow when it is opt-in.

AMP content shows up before other organic search results. So it's not really optional unless you want to lose all your search-driven traffic.

I think Google had to do something, publishers overdid it with their crappy websites with 50 tracking links and banner ads. The mobile web was slowly dying and everything was moving to apps. AMP is their effort to reverse that trend somewhat. I haven't heard any better ideas (Facebook Instant Articles is worse imo).

do you realize Google bought and now owns 49 of those 50 tracking systems you mention?

the main goal of amp is to monopolize it further and make that last one they couldn't buy stop working.

Google can rank fast pages better. It's perfectly possible to build fast pages with html, but Google wants to own the Internet.

It's like spying on everyone so we can make world safer.

The guy said AMP doesn't influence ranking. Who do we have to believe ?

it may not be a direct ranking signal but i think this is a little disingenuous. by putting this "lighting" icon next to some results, the CTR (click through rate) on the SERPS (search engine results pages), will likely be higher for AMP pages, and that higher CTR _is_ a direct ranking signal which will eventually lead to better rankings.

that is for organic results, but the AMP news carousel _does_ appear at the top of the search results pages.

It doesn't influence ranking. It shows up at the top in the carousel.

Is there an option to avoid it from an end user's perspective? It's made google news so frustrating for me, it will go back a page in history or slide to the next story for no reason whatsoever. I have a Nexus 6P, use Chrome so there's no reason for it to act like that other than just being shitty.

I kinda just stopped reading Google news rather than try to fix it (:

You could try using the Firefox browser. I haven't seen any AMP pages since I moved to it.

Because those that do not opt in will be implicitly penalized by users trained to prefer AMP pages, whether that preference is warranted or not.

The top navigation, from a search on android, is almost the same between an amp page and a non amp page. The only difference is the text says "from ..." rather than just the site.

it exists for normal web pages and is actually named Chrome.

I wonder if AMP isn't just an open source/mobile friendly version of StumbleUpon. Not sure if anyone remembers those SU banners at the top of most blogs, trying to get more in-site clicks from deep linked articles.

As of now it's not a requirement, so I'm not sure how your hypothetical is relevant.

It is a requirement for certain placement; such as the high ranking top-level news carousel.

As a user I've learned to avoid AMP pages because the UX is horrible:

* back button is broken 1/2 the time

* the bar wastes 1/3 of my screen

* I can no longer see what site I am on in the url

* it's hard to navigate to the / of the site

* I can't forward the link

* being on a good network in US it solves no problem that I have

What I'd really like to see is a way to opt-out of seeing AMP'ed pages in my search results. Or at least a way to navigate from AMP page to its native version.

Further, I noticed that AMP is a signal for low quality content. I am guessing sophisticated publishers are conservative enough to wait and see. And individuals haven't bothered dealing with it. So you get low-end publishers in between.

Interesting. As a user, I've learned to only click on AMP results. They load quickly and reliably (no blank white site) even on LTE, and they never have intrusive adds. If they added an option for "only show me AMP unless I click a button to ask for the other junk", I would opt in day 1.

Only content owner should decide what ad will be shown. Offending this rule means breaking deals with advertisers, it's reputation losses.

Visiting non-AMP sites on mobile is masochistic (e.g. try searching news results). You get basically all the same problems, but also pop-unders, late loading ads, greedy javascript, autoplaying videos, and other bad behavior.

The AMP is really quite solid for the news use case, in my experience.

Using Firefox for Android with uBlock Origin pretty much solved that problem completely for me.

I use Firefox on Android + uBlock Origin as a default, but even loading the Firefox browser takes longer than loading, reading the first few lines of an AMP article, then going back because it wasn't interesting. And I have a high-end phone.

>And I have a high-end phone.

And bigger issues. My Galaxy S3 takes 4 secs to open Firefox without cache compared to 3 secs to open Chrome.

AMP is bad on mobile Safari. The URL bar does not hide and reveal properly, the reader button doesn't work, and the scrolling momentum feels very wrong. That's why I try to avoid AMP links.

Non-AMP sites are frequently improved by the reader button. That's how I prefer to read news sites.

Absolutely. How could I forget disabled reader!

Use uMatrix plugin to block unwanted content.

Agree it's a signal for low-quality bordering-on-clickbait content, so subconsciously also find myself avoiding them. Unrelated: formatting lists like in your comment makes them really hard to read on mobile.

Thanks, formatting fixed.

Yeah, I was quite surprised how positive a reaction the original complaint got, considering it was long-winded, repetitive, and seemed to go out of its way to misunderstand AMP.

The only valid criticism appears to be how google displays search results using its own URL and this toolbar – it seems to break rather fundamental assumption about http and has the potential to break all sorts of tools that rely on the established structure of the web and open standards, as has already happened with the refer(r)er as mentioned in this response.

I wonder if there's a way to get the same result without rehosting content on their own URL. Couldn't they allow publishers to achieve the same result with a CNAME, possibly for amp.<hostname>.<tld>? And do the google servers add anything beyond being distributed caches? Because if not, it seems this level of indirection is redundant for websites already hosted on CDNs.

Regarding the toolbar: yeah, that's a terrible idea. I have no sympathy for publishers who object to it because if it reduces your retention rates there's probably more wrong with the content than the presentation. But as a user, it's the sort of "assisted browsing" that feels intrusive, like resizing the window or a "you need flash" popup (I don't).

Considering their market share in browsers isn't far behind the in search, I wonder why that function isn't just a chrome feature. Funny thing is: it's a feature that exists in Safari ("Search results snapback").

If you look at the AMP docs (https://www.ampproject.org/docs/guides/discovery), publishers use <link> tags to point to the AMP page. For the original article, the AMP version lives at https://www.alexkras.com/google-may-be-stealing-your-mobile-... but the URL could follow whatever scheme the publisher wants.

Since Google search is most likely discovering the AMP page via the <link rel="amphtml"> tag, they really should just use that URL. The Google servers do add a couple minor things beyond being a distributed cache though. First, there's swipe navigation to other articles if the link was located in the top carousel in search results. Second, if you are on a laptop, the Google servers will redirect you to the full version rather than use the AMP cache.

Google has their version of the internet. Facebook has their version of the internet. As time slides forward just think of how many users won't know the real internet from these rubber doll versions.

Congrats humanity. Your best invention ever and it only 20 years or so to completely fuck it up. This is why we can't have nice things.

People need to own the means of production. In this case, it's the search engine and social network. We can't rely on companies, because they are greedy, nor on government, because it is corrupt. This is what happens when we don't own the tools we use - we are subject to exploitation, others making fortunes on our backs and abusing us by incessant tracking.

I'm amazed that there is no federated (not owned by a single entity), anonymous search engine to this day and we're still pouring our brains into Google. Google knows more about a person than his spouse, doctor and lawyer put together!

Correction: There is no federated, anonymous search engine that gives good results.

There cant be.

People/powers-to-be would fuck it up, there would be "copyrighted material" and eventually it would be declared an illegal activity.

Only companies such as Google and Facebook prevail not due to technical reasons, but due to political power. They are selected by the likes of CIA, invested in and nourished to their position, and do their bidding. And alternatives are extinguished.

Google should not have existed had copyright laws been enforced as they are towards another search engine, thepiratebay.org

Your response sounds more like the religious groups howling "world will end" whenever a new scientific progress is made. Internet survived because of its obvious commercial value. Google and Facebook are doing nothing new but making their ends meet.

Neither of those points are much of a defense. By that same token, one could defend e.g. Microsoft's aggressive misuse of their market dominance in the 90's: 'Windows survived because of its obvious value; MS is just making ends meet.'

It can still be a good or a bad thing for the web as a platform.

Hopefully adoption is slow enough that the default behavior of any "blessed" libraries like this WP plugin can be encouraged to be less ... well, not evil, of course ... but really, why does that toolbar and 'X' need to be there? Browsers have a back button.

My take is simply that this is not a good look for Google, and hopefully they realize that. The article featuring the WP plugin, and these discussions on HN that give a lot of attention to the carousel, and how it competes with non-AMP results, don't build goodwill for the project.

>many users won't know the real internet

Many users have no need for 'real internet'.

>completely fuck it up

No. 'Real internet' is still there, it's not getting absolutely smaller, only relatively to 'rubber doll' one

He avoids really hard addressing the only bit that matters

> Guess what happens when the "close" button is clicked inside the AMP view?

And the amount of disingenuous on this just rubs me the wrong way:

> If you are not comfortable with traffic on your AMP pages, please do not publish AMP pages.

Happy to address the close button:

You are on a search page, you go on layer deep by clicking. The close button goes back. I don't see what else it should do. Mind you that I do understand that a link to the origin server would be a good improvement (and I've stated that in the post).

I'm not sure what you find disingenuous. AMP are just web pages saying "HEY I A AMP, please use treat me as AMP". You can publish the same bytes minus the "HEY I A AMP" and get the old behavior.

>The close button goes back.

No. Just no. The back button goes back. Period.

The close button closes modal windows overlaying the original page and originating from the same site.

I can already hear you countering "But.. since the content is served from google.com due to caching, this pattern is technically correct".

Which is tantamount to saying "We serve the content. So we can add whatever the heck we want"

Very frigging clever Google. Well done. /s

The UI logically and visually belongs to the opening app, indeed. If the page itself opens a modal it opens on top of the bar and it's close button closes that modal.

but it's not an "opening app" ... it's a search engine. with links to other sites.

... or is it?

Philosophical question...

If I hit the search bar on my Android device, and it opens Google Search in an app interface, is it an app, is it a search engine, or is it both? What about with their new strategy of integrating Google Search more closely into apps? How about then?

The line is so blurry as to be irrelevant. The wording on the responses seem to be on a different side of that philosophical argument, but I'm not sure it is that black and white.

"You can publish the same bytes minus the "HEY I A AMP" and get the old behavior".

Yes but then you won't get the AMP icon next to your page when it comes up in the search results. Over time users will become conditioned to prefer clicking on AMP enabled links. With the icon you are basically categorizing the search results into "really fast sites" and "regular speed or slow sites", even though it is perfectly possible to achieve the same AMP speeds without using AMP.

Would you agree that the completely standards compliant, well-understood-by-users back button already replicates that behavior perfectly? Why does the close button have to have a different icon for the exact same purpose, if not to deceive?

Please address this question @cramforce.

Why are you introducing a new and totally unnecessary display element when the back button is built right in to the browser and everyone already knows how it works?

No matter how hard I try, I can't think of any good reason for this banner.

But hey, to the right of that X there sure is a lot of room for placing an adsense ad! What a strange co-incidence.

The banner is needed to show the logical origin, given that the browser is still on Google.com and the article is shown in a iframe (to enable pre-rendering). Our native apps do not have this constraint and thus don't have the extra bar.

I can't give you a 100% reason for the X. I know multiple UI versions were tested and this did well.

>to show the logical origin, given that the browser is still on Google.com and the article is shown in a iframe

What an amazing coincidence that this great new tech requires users to remain on google.com "for technical reasons".

If you have an alternative implementation strategy that enables pre-rendering and swiping while showing the source article origin I am all ears.

We can do it in our native apps, and we do.

I'm not familiar yet with the pre-rendering functionality so I can't comment except to ask - why are you making this contentious feature a requirement of using AMP? Why not allow publishers to opt out of the Google CDN + pre-rendering but still retain all the other benefits of AMP? This type of progressive enhancement would likely benefit more sites/users instead of an all-or-nothing approach.

Any time I see a website with a bar that has an "X" on it (think of all those "this site uses cookies..." bars) I expect clicking "X" to dismiss the bar, allowing me to see just the content minus the bar. If it took me back to the previous page, I would find that behavior surprising.

I second this. Google Translate does this, I don't see the reason for such inconsistency.

Why can't I use AMP but opt out of the AMP Cache? Seems strange that I have to choose between AMP and actually receiving my own traffic.

I guess I could use AMP but intentionally break my page's validation so it wouldn't be cached, but then I wouldn't get the special flag on SERPs, which effectively penalizes non-AMP sites.

There are things that can be done with AMP, that cannot be done with regular fast results (and you can very much build pages that load faster than AMP not only as fast!): - pre-rendering - result swiping

The internet is about hyperlinks, not hierarchical views - this move by Google feels very disingenuous.

So basically this guy says all of the issues are by design and you can choose not to use amp if you don't like all of it. Google isn't going to penalize non amp pages in search ranking. They are just going to not show your page in the carousel.

I think it's a pretty fair position to take. If you don't like amp don't use it. We'll see if amp catches on over a period of time.

Yes, accept our platform, design choices, and navigation framing or we won't put you in a competitive position with other search results. While AMP isn't a ranking factor (Google has stated this) AMP results are displayed more prominently. Arm twisting at it's finest.

Even if AMP isn't a ranking factor, clickthrough rates may be. If users click more on AMP links, then by improving CTR it could also indirectly improve ranking.

"Gosh, I mean, you have a right to sell oil, sure, but maybe if you play by our rules, you'll have an easier time shipping it on the rail networks we own." It's plainly Standard Oil-style monopoly behavior from Google here.

It seems more like UPS saying "hey, we'll ship whatever you want (within reason) but if you can fit it in one of our standard box sizes it'll get there faster and cheaper". What is the ulterior motive or side benefit Google gets from people adopting AMP? Honest question.

The problem is that if you don't play by their "standard box" rules, you may not end up in the truck at all.

The users browsing the AMP results may never feel compelled to scroll down to see non AMP results.

So a publisher's choice isn't: "standard size for faster delivery", it's: "use AMP or receive less traffic than you're currently receiving"

AMP articles keeps users browsing within Google's properties, with Google's tracking and ads. Non-AMP news pages are an exit out of Google's sphere, where users will get non-Google tracking and ads.

That said, I find myself clicking AMP articles 5x more than non-AMP news links because I know the AMP articles will be fast and clean.

It's using the same libraries and third party pixels as publishers were before.

Them becoming a "platform" where don't just link to content but can actively host, curate and present it.

> monopoly

I'm still surprised so many people here are using google and not duckduckgo. I use the duck all day long. I used to revert to google when I couldn't find when I want, but I've found in the last six months, if it's not on ddg, it's not going to be on google, either!

We don't have to support google. Let's just stop giving them everything for free. People sneer at using Yahoo for mail, but what is the problem, really? It's almost all aesthetic.

Why give away our power and beg google to throw away profit and treat us right?

Persuade your friends to use other email and search providers. Design your websites so they have Links sections your visitors can navigate instead of going back to google. That's what I do and it works.

Amp isn't ideal, but neither is Instant Articles on Facebook.

I'm hoping that Amp is really just Google training publishers how to make web pages that aren't terrible for users.

And then you have companies like Buzzfeed that have said "Ok, the deal has changed. We're ok with the new terms." and go hog wild changing their business model to adapt to the new reality.

It is truly fascinating to watch.

Instant articles is even worse, IMO. But Facebook is not pretending it's the next version of the Web. Google is, so they set up an expectation of Web like behaviors.

So what would you have done if you were in charge at Google?

At the time Amp was announced, Facebook's Instant Articles was really taking off. It was by far the best place to read news online and arguably still is.

To get a similar experience outside of Facebook (before Amp) you had to use an ad blocker. Right around that time, Apple added support for ad blocking in iOS and suddenly ad supported content outside of Facebook was looking unsustainable.

> So what would you have done if you were in charge at Google?

File an antitrust complaint against Facebook, sue Facebook, and get a warrant to get Instant Articles instantly shut down?

> suddenly ad supported content outside of Facebook was looking unsustainable

And just pivot regarding that?


well, AMP is open source. Instant Articles is not.

AMP works on many platforms, Bing just announced support of AMP.

So, what's the big deal?

Instant Articles is just RSS, isn't it? How is it less open than Amp?

I think one important point is hidden within this sentence:

"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 [...] control for the publisher."

(emphasis mine)

So Google sees itself as a platform in the tradition of Twitter and Pinterest - i.e. a controlled space on which content (with more or less control by the authors) is published. That's a significant difference to the gateway to the open web that they primarily still are at the moment.

I agree, this is the key insight. Google sees itself as moving toward what I see as a more reddit-like model. People might occasionally click away on links to other sites, but the idea is to keep them on Google's "platform" by default.

That's why clicking "X" means "go back to Google", because in this model, you never left. You just temporarily viewed some content in a pane which you are now closing.

Yeah, that line really stood out to me too. Google is attempting to shift away from being a search tool which funnels users to a destination site, and instead become the destination itself, where sites submit their content in search of users.

As positive as the response sounds, it is an empty promise.

> We’re looking at ways to make the source link more discoverable and will update once that is done.

If Google was actually going to fix the issue, they would have said "we will make the close button direct users to the original site and will update once that is done" OR "we are changing the x (close) button to a ← (back) button".

"x" means "close" and "←" means "back". This is confusing UX at the least and arguably a dark pattern.

> "we will make the close button direct users to the original site and will update once that is done" OR "we are changing the x (close) button to a ← (back) button".

As a user, I hope Google do not implement either of your suggestions.

When I click the 'X' I don't expect to get a copy of the page that loads a bit slower, and I'd be surprised less than 95% of Google's users agreed with me (but by all means do a test).

And I don't want the close button changed to a back button. When I click on a search result I might then click one or two links on the publisher's web site before wanting to check out the next result on the Google results page. I want a single button that takes me quickly back to that search page - I don't want to have to click a 'back' multiple times (and maybe overshoot or undershoot my desired history position).

Am I in the minority in that I found the function of the "x" to be very clear? I don't know if it is fair to call it "nothing less than a dark pattern" without a bit more data.

That is where we currently are. In our user testing it tests really well, but there appears to be confusion in this community which would be good to address.

My current feature request is to link the "publisher domain" to the origin article, so that one can always click that to get there and use it to copy the link.

Yup, everytime I've seen this X button, clicking it has done exactly what I expected.

I've also sometimes been on a non AMP version, and Chrome has suggested "a version optimized for mobile". Clicking that has opened the AMP version and after reading the article, clicking the X took me back to the non AMP version.

Yeah, I toned down my response, since it's not intentionally a dark pattern.

I don't understand why AMP even has that close button. But taking the user to the original site would just appear to reload the page, only slower.

Am I being uncharitable in saying the TLDR is:

This is working as expected and you need to do more work to make AMP work for you?

Why can't it work with templates, why does content that was created need to be created again for AMP? This doesn't seem very scalable.

Also, the attribution for ads and analytics basically means you need to reimplement your entire tracking code within the schema and spec of AMP's analytics attributes which only supports a subset of existing providers rather than allowing an abstract interface.

Also to the point that AMP doesn't affect search position, is this true if someone serves a shitty AMP page? Or is it only true that it won't boost position?

Well, it certainly works with templates (like the WordPress AMP plugin), but that only gets you that far. It is definitely work to adopt AMP, but most people who have done it are pretty happy with the workflow. We are always listening for feedback, though, and are looking for ways to make it easier.

Am I the only one here who has never seen an AMP site in the wild? Seems people in the comments are super familiar with the service.

I have never seen one either, but I don't use Google search anymore.

Perhaps you don't use it either?

I've never seen too, and I use only Google search (though not on Android).

I think they only appear if you're using Google Search from Mobile.

I use google search and chrome on Android and haven't seen one either. I don't see the AMP icon when I search for 'git tips'. Maybe it's US only for now.

"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."

So far I haven't thought of Google as a platform like Twitter or Pinterest. I thought there was a free web with Google and Bing and good old Altavista searching and indexing it. Is making a website now similar to posting something on Facebook? Or why would they have to assure you that your traffic is your traffic, "period"?

If that's the case we need an urgent change in direction!

It's nice to see Google investing in direct responses to the community. If you build a developer product, monitoring HN for criticism and responding is very high ROI. Few companies actually do this.

If hacker news had a speed rating next to the links I would probably click on them more instead of just reading the comments to decide if it's worth pulling up an article.

Note, It has been suggested that a speed rating on Google would be equivalent to the amp experience with less Google control.

Speed also depends on where you are geographically. Amazon.com is probably faster than Taobao.com in the States, but I'd have the opposite experience over here in China. But then Google is the slowest this side of the GFW, so that might be redundant.

True, but Google knows where you are and could test the site from that general region. It would cause additional overhead for Google but may be worth it.

Page render speed is just as important as server speed. A website bloated with ads is what I object to the most, or any type of pop over that obstructs the content.

I wonder whether they incorporate these things into their ranking metrics already, similar to how they do their Quality Scores for Adwords.

Good question, in 2010 Google said that they "... today we're including a new signal in our search ranking algorithms: site speed. Site speed reflects how quickly a website responds to web requests." [1]

I believe that page speed is also important, google provides PageSpeed Insights to help you improve your site. [2]

[1] https://webmasters.googleblog.com/2010/04/using-site-speed-i...

[2] PageSpeed Insights https://developers.google.com/speed/pagespeed/insights/

Doesn't convince me. Top bar is the biggest problem and they are "working on it"™ (we all know what does it mean) and "it will be landed in Chrome soon", when Chrome is not only browser in the world of mobile devices. Also from this response I see they want to make Chrome the new IE (if problem is solved in Chrome - it's solved), and it's frustrating.

End users, in generall, don't know and don't mind about these complaints against AMP.

What matters is if one gets the results he was searching for. And somehow those pages that have "amp" written nearby open faster, so let's click more on those.

Market pressure will drive decisions on whether it's better have amped pages (with those claimed drawbacks) or try to capture attention to the whole site with navigation and lighter pages.

And as a prisoner's dilemma, if sites that don't have amp are as fast as the amped, having "amp" near your link won't matter. If they are slower, the distinction will matter, independently of your specific optimization.

I don't see why Google needs to have its own cache for AMP pages. If a publisher has its own fast CDN, why not just let it serve the AMP pages from its own domain?

The pre-rendering is one feature that requires a trust relationship between the host (to only serve valid AMP) and the platform. Otherwise the page may e.g. use a lot of CPU while in the background.

So you’re saying AMP doesn’t actually speed up pages at all, and it’s only sped up because you pre-render?

You just admitted the purpose of AMP being moot.

There's also this little chestnut:


TL;DR Google AMP also hijacks app deep links

Previous HN comments on the original story: https://news.ycombinator.com/item?id=12722590

> “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

It doesn't scroll away. Just checked on Google news with both Chrome and Opera. Android 6. Am I missing something?

So if I built my own personal ad network, I could make a PR and have it be accepted?

Somebody is currently doing that:

An external contributor is currently building first class support for "My simple ad network that just wants to put an image with a link on the page" https://github.com/ampproject/amphtml/issues/5541

There are other ways to do it, though.

I've heard a lot of discussion in the ad community around how this limits competition of ad networks. Yes, it is awesome you have many on board for launch, but what of the ones who are not? Is there a publicly available process for getting included along with guidelines to set reasonable expectations for whether one will be included? If so, I'd love a link.

See https://github.com/ampproject/amphtml/blob/master/ads/README...

I think with the numbers of ad tech companies coming on board every week one can safely say that the process is working for them.

Does Google have any plans to let users opt out? Maybe an icon that takes you to the original? I tried opening the AMP link in a new tab and it took me to the site, so at least there's a workaround.

Since I normally use Firefox Mobile (has to try it with Chrome), I don't get the AMP icon and I'm taken to the requested site. I'm guessing this functionality is limited to Chrome and (maybe?) Safari on mobile.

AMP is attempting to solving a challenging problem, and although I don't personally agree with their solution, I've gotta recognize that opening the embedded AMP version of the page from the result of "git tips" was faster on Chrome than on Firefox. I'm hopeful that the lessons learned from this will be pushed upstream and help improve the web.

AMP works great in Firefox (modulo this bug right now https://github.com/ampproject/amphtml/issues/5479 :) and all other modernish browsers. Google Search currently chooses not to display it for FX, but that is not due to a limitation in AMP.

Instead, it’s because Google Search artificially penalizes Firefox users, and even removes search options for them.

There’s been discussion about adding Google Search to the list of sites for which Firefox Mobile always serves a faked User Agent because of this hostile and anticompetitive behaviour before.

One of the problems here is that when users click on a search result (marked AMP in Google's SERPs), they expect to visit a website. Instead, they visit a cached version of the page, on Google's site.

This is deceiving.

If Google's going to cache the page, then there should be a "cache" link so the user knows that they're clicking on a cached version of the page. Otherwise, deliver the user to the site that they think they're clicking on.

If it was Verizon or AT&T or some other ISP doing this there would be outrage. But I guess it's cool if it's Google.

so how do i view the original page...?

I'm interested in the intellectual property aspect of this.

Why exactly is it permissible for Google to "cache" my page and then serve it up to users?

That's literally them copying my page to their own servers, which is my content, my intellectual property, and then serving it up however they like.

I'm a little confused how that is legal.

>Hey, this is Malte and I am the tech lead of the AMP Project for Google.

Of all his 725-words response only 18 words directly address the problem:

>We’re looking at ways to make the source link more discoverable and will update once that is done.

That's 2.5%. Is he really a tech lead?

This seems to compare favorably with the original article. How much information density are you expecting?

Then you go on to imply that he's unqualified to be a TL because you don't like his polite and conversational tone in an article that criticized his project?

I just don't understand HN anymore.

I wrote the article and I agree. I brought up a lot of things without (as it's apparent to me now) clearly stating which were my fault (mostly everything) and what I blamed on Google.

>I didn't imply anything you've said. I just noticed that his response looks like a product of PR or marketing department rather then tech lead.

A appreciate the fact of quick reply, even if half of the credit for it should go to HN discussion.

Everyone has to do PR to some extent. One misplaced word will create 100 comments on HN complaining about it ;)

Applications are open for YC Winter 2018

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