Google now automatically converts Flash ads to HTML5 | Hacker News Reader

Google now automatically converts Flash ads to HTML5

venturebeat.com

207 pointscpeterso92 comments

Open articleView on HN

Shame that I automatically convert all ads to nonexistance.

It's too little, too late for rich media/content-heavy ads. Especially in the mobile sector:

* Ads are a performance burden on low-power devices

* They consume extra battery; power is wasted on superfluous cruft

* Bandwidth wasted; especially a concern given draconian data caps

* Tracking and privacy concerns

Not to mention the fact that ads are a primary attack vector (think of all the advertising CDNs that have been compromised and used to deliver malware). The real improvement here is moving away from a proprietary plugin.

Pointing out the problems of rich media ads isn't very useful here though, since most of us are probably aware already. My proposal would be simple, inline text, along the lines of "this page load sponsored by _". Doubt that will gain traction, though.

>They consume extra battery;

I find it fascinating that Google is always hurting their own mobile experience. Between all the issues android has with wake locks and memory leaks and the massive infrastructure of location and 'lifestyle' tracking for ad delivery, its a miracle Android devices get any battery life at all. Even without this html5 conversion, Android battery life often pales compared to iOS or Windows Phone.

The more I think about it, an OS from whats essentially an ad delivery company is probably a liability. Monetization will trump user experience. Get ready for some pretty invasive html5 ads.

Aren't most iOS apps ad-based, too?

Not most of them, in my experience. Most seem to rely on the initial purchase price only, and many of those that used to have ads have switched to in-app purchases.

While I could not tell you for sure why this is, several statistics indicate that iOS users have more money to spend and piracy is much less frequent. Ads also have a tendency to disrupt the "experience" of using an app which may be more discouraged on Apple platforms.

With that said, there are also many apps which make no attempts at monetization at all.

I would've thought on HN of all places that one wouldn't have to explain how silly it is to make conclusions based on incredibly anecdotal experience (particularly when more robust statistics are a quick Google away).

"According to Distimo App Analytics, the mobile apps with higher gross earnings in both the Apple iTunes App Store and Google Play Store use freemium monetization methods.In October 2014, nine of the top 10 apps in the iTunes App Store within the US were freemium games, with Clash of Clans, Candy Crush Saga, and Game of War: Fire Age leading the ranking (followed by Pandora Radio). You have to go all the way down to the 27th place on the list to find a paid app, in this case Minecraft."

http://thinkapps.com/blog/post-launch/paid-vs-freemium-app-m...

I've been keeping up with monetization statistics for apps, and last time I checked, ads+in-app purchase were 90-something % of revenue for iOS apps and growing.

It should be noted that your source focuses on revenue, and not the amount of ad-supported/non-ad supported apps in the App Store which was the original question.

Interestingly, these freemium games rely heavily on traditional television advertising.

That's probably why they're in size of a small tablet now.

Seems like this is the easiest way to fix battery issues. Google is in an ugly place not only with its relationship with advertisers but with its own OEMs, who modify Android with usually battery-killing results.

All that extra surface area means lots more battery. Its a classic 'throw hardware at the problem' approach and, unfortunately, that just seems where Android is going. As Lollipop has showed, Volta and ART are not battery savers. If anything, people are reporting worse battery life than under KitKat.

I'm not going to go as far as saying Android is bloaty and out of control; I just don't think you can have the level of QA and control you need for an ultra-efficient mobile OS when you're dependent on so many other chefs to deliver your product and monetization for you (OEM hardware and software, ad monetization, user tracking to better ad delivery, etc).

As Lollipop has showed, Volta and ART are not battery savers. If anything, people are reporting worse battery life than under KitKat.

Yet the GS5 shows dramatically better results. A number of handsets show dramatically better results. That you read someone's anecdote about their Nexus 5 is irrelevant.

Sorry, but I find your posts hilarious, and again I have to say that they're astroturfing. "I'm not going to go as far" -- sure.

Seriously, are people not aware that HN has been targeted?

Ugh. Blocking or not installing Flash used to be an easy way to avoid the most annoying ads. Now I need to figure out some mixture of request-blocking and CSS element-hiding to make them go away. And CSS hiding doesn't address the fact that most of the bandwidth I buy on a cell carries content I don't want: a 1000-word article is maybe 5k of text, but how many megabytes of ads and Javascript garbage? Nor does it prevent the tracking.

Simple, tasteful, first-party ads would make the world a much better place, but sadly, I don't think that's the direction we're heading.

If you can live without Flash, it's worth uninstalling. These days, the killer apps you'll experience from Flash are zero-day exploits.

Also, on my iPad, I find that many websites (news sites are the worst offenders) keep messing with my "tap to zoom on the main column of text" action. That is, my zoom setting is reset several seconds after the main content is displayed. I have a hunch that it is, in a large part, due to ad networks loading long after the main content has loaded and forcing a reflow of the page.

Like you, I don't mind ads per se. I even feel badly blocking them, knowing that I'm depriving a content creator of revenue. But the issues you point out are the very reason I can't have them in my browser.

I used a computer without AdBlock the other day, and it was like stepping back into 1999 with every page saturated with the 21st century equivalent of animated GIFs and auto-play MIDIs.

An alternative ad network isn't as far fetched as it might seem. Clearly delineated simple ads were a key differentiator in early iterations of Google Search.

Perhaps what we need is a "ethical ads" network, with strict requirements on what could and could not be done:

1. Plain text and link only, with clear separation of content and advertising.

2. No user tracking (Maybe with explicit opt-in? And even then with as-good-as-possible anonymization of the stored data.)

3. A deal with AdBlock and competitors to share a cut of the revenue, in exchange for allowing Ethical Ads through.

If advertisers and ad networks keep following the current trajectory, there will come a day when ads are simply too intrusive, and ad blocking will go (even more) completely mainstream.

Rich media doesn't do advertisers any good if no one ever sees their ads.

This already happens, and people were upset about it.

https://adblockplus.org/en/acceptable-ads#criteria https://news.ycombinator.com/item?id=5995140

Very interesting as well. I had no idea this existed.

I used to work in the ad industry and deeply analyzed this stuff. There's a lot of unspoken secrets in the industry:

1. The majority of clicks go to flashy, annoying ads. 2. People who click on these ads rarely convert, i.e. sign up for a newsletter or purchase a product. These are probably accidental clicks. 3. The ROI on advertising is often negative. 4. No one cares about any of this because click through rates are still the gold standard. Advertisers aren't savvy enough to know better, and ad agencies don't get paid either way.

It's a terribly inefficient industry, made worse by the fact that attribution is very difficult ("did this costumer buy this product because of an ad?"). This all has led to a fall in ad rates, which is just leading to more annoying ads. It's a vicious cycle that makes clean, honest text ads even less likely.

Fascinating. Inefficiency, to me, is opportunity (even if it can't always be leveraged.) A fundamentally different ad network is an idea I've been kicking around for a while. Perhaps I should get more serious. Do you know any resources, or any technical information about how people get started in this space?

@ROFISH: Absolutely a hustle business, but the more experienced I get the more I think that applies to 99.99% of businesses.

I think this is something that has already started playing out with designer and developer oriented networks like The Deck and Fusion Ads. Better design, standardized rules targeted towards specific groups.

The opportunity is more in the network than the tech. For something like this to work, you have to convince both site owners and advertisers to use your network. (Not that it's a bad idea, but it's like 97% hussle and 3% tech problem.)

I think the "easy" way to shove yourself into the market would be to tackle a niche, so you can specifically target both the advertisers and the publishers that have an audience that would engage with those advertisers. So if you decide to choose, say, online tackle shops on the advertiser side, you'd target fishing blogs on the publisher side. That way you're pretty focused on a niche and don't have to solve difficult ad matchup/inventory problems early on.

The bonus is that if you figure out how to make publishers and advertisers money, you're going to make a lot as well :)

The only real way I'd get on board with a network is if they had access to relevant advertisers in my particular industry that I didn't already have.

I've chosen to direct sell advertisements to companies in my site's industry which takes a lot more work than handing the reins over to a network or dropping in a widget, but the results are definitely worth it.

All you need is a niche market and your first advertiser. Then start contacting the publishers in that space. You'll need somebody to help the publishers embed the ad codes. A decent developer could track everything or you could just use Google Analytics or some other commercial tool. Most publishers are willing to try something new, but consider the ad inventory is worth much more to Google (Adsense) because they have a ton more advertisers bidding on the inventory and they're tracking people's interests and so forth.

On the one hand, Google completely ignores 99% of its publishers. On the other hand, the bigger publishers receive advertising proposals all day long every day. So yes the advertiser can contact the publisher directly, they don't really need you, which (last I checked) is against the Adsense/Adwords TOS.

From what I know from people who buy AdWords ads, they manage to get a lower price-per-click and a bigger return by optimizing their campaigns

This has nothing to do with flashy ads, quite the contrary, so it's not only "inefficient ads that nobody clicks"

I've concluded that Gresham's Law is among the universals in economics. It joins a very few others, including the Law of Diminishing Returns and the Jevons Paradox.

What's the relevance of Gresham's Law here? Not criticizing, just genuinely curious.

The viscous cycle element.

Grisham's Law, generally, is a "bad foo drives out good foo", more specifically, when the costs of distinguishing good foo from bad foo are high, or when the market doesn't recognize a distinction between the two.

It was coined (so to speak) as an observation that "bad" (e.g., debased) currency drives out "good" (high-specie) currency. It shows up in Akerloff's "The Market for Lemons" and many other places. I've made something of a hobby of collecting references.

For more: https://www.reddit.com/r/dredmorbius/comments/2h63fp/is_gres...

>...when the costs of distinguishing good foo from bad foo are high...

Ah, there's the part I was missing. I didn't think of it beyond government regulation. Interesting insight.

It seems to arise out of several possible factors.

In the case of money, it's the statutory definition of value. Since specie has value independent of the nominal value (though usually below nominal, hence, seigniorage), where there's a positive, valuable metals are hoarded.

In most of the other instances noted, value determination is by other means, some set by market, some by difficulty in full value assessment.

But in all cases there's an informational asymmetry, where the party who's aware of the true value of a good (or service) may choose to act in ways which maximize benefits (or minimize losses) that are both 1) not fully captured in market value and 2) aren't accessible to the counterparty.

What about, I dunno, supply and demand, or "people respond to incentives"?

Don't feel bad about it. All the small advertisers are glad that you don't view their ads because they know you won't buy from them anyway and would be throwing out money without getting anything for it. Expecting your visitors to look at ads that you get paid for is just another type of fraud by the content-generator against the advertiser. :-)

I've always thought expecting direct revenue from web ads was looking at the problem upside-down.

Nobody purchases directly from mainstream TV ads. The purpose is brand awareness, and sidebar ads have always seemed infinitely more useful to build such awareness among the public.

Optimizing on CTR or impressions just seems like optimizing vanity metrics, because they're easy to measure.

>Nobody purchases directly from mainstream TV ads. //

I doubt this is true. A small minority, yes, sure, I can certainly accept that. Any company that does a product based advertising push with a product launch also seems to disagree with you. I don't watch mainstream TV (partly to avoid advertising) but the little I do see still appears to have some direct purchase focussed ads - particularly kids toys at Christmas (it happens to be that I saw kids FTA-TV just before Christmas).

CTR seems a pretty good metric for engagement (yes you can't some false clicks; also assuming no one is maliciously generating clicks).

Impressions seems prone to many problems that don't make it easy to measure accurately - just because an ad is requested doesn't mean it's rendered to the viewable page area.

Still, I run ads on my blog which make it self-financing. I also have donation buttons (Paypal) and modified my widgets to allow ads to be turned off per blog-post - so some posts don't have any ads at all. Never had a donation but have financed itself [ie server costs + domain names + perhaps a tiny bit more] on ads for ~8 years.

When I said "nobody purchases directly," what I meant was that nobody presses a button to buy Pepsi or Coca-Cola when their ads air. Infomercials would be a good counterpoint, as would direct sale telephone advertisements.

But the best and most expensive television advertisements, from primetime to the Superbowl, are bought on the assumption that they increase brand awareness, driving future sales.

In the online space, our ad networks seem to be stuck in a permanent parade of crosses between infomercials and billboards, attempting to push product right now. To me, that seems to be leaving a lot of opportunity on the table.

I think that eventually what's happening with TV and radio will happen with other types of media. I almost never watch TV ads anymore. I get more and more content through iTunes, Netflix, HBO, etc. where there are no ads. I TiVo anything I want to watch from the networks (even live events) and skip the ads. For radio, I use Podcasts, iTunes Radio and Pandora, and I pay to not have ads. I use AdBlock on the web, but mostly use sites that either don't have ads, or I pay to get rid of them when I'm able.

I'm sure free content with ads will exist forever, but I suspect people will find more and more paid options and start using them as ads get more and more intrusive.

You're forgetting about the market feedback effect. In a world where most eyeballs are in the paid options, advertisers will find ways to co-opt the paid options. The companies that accept those contracts will have more revenue with which to deliver the same quality at a lower price, and will attract viewership proportionally, etc., etc., etc.

> Perhaps what we need is a "ethical ads" network, with strict requirements on what could and could not be done

Personally I don't mind unintrusive (small) static image ads from quality advertisers tailored to the type of audience of the site. A List Apart I can remember having a worthwhile ad network whose ads I've actually been interested in several times in the past (which is very rare). Problem with most ads for me, and the reason I choose to block the majority, is the lack of high quality products being advertised and the fact that the click-through rate is so low the ads rely on other tactics to grab attention, fueling the desire to block them.

Exactly. Given the vast series of networks and data miners tracking and analyzing my online interests, why can't a site manage to show me ads that I actually find interesting?

there will come a day when ads are simply too intrusive, and ad blocking will go (even more) completely mainstream.

That is, if the war on general-purpose computing doesn't eventually make it nearly impossible to perform any form of adblocking...

Currently, the most popular form of adblock is probably a browser extension. Browser vendors are seeing extensions as a possible vector for malware (which is true) and the trend seems to be for enforced signing. This gives them control over what extensions they'll accept, which also gives them the power to prevent adblocking extensions. It isn't happening now but I wouldn't be surprised if it does in the future. On the desktop there are ways around this but they are far more involved than they used to be. The situation on mobile devices is worse.

https://news.ycombinator.com/item?id=7237725

https://news.ycombinator.com/item?id=9061592

Another form is a filtering proxy. Admuncher, AdSubtract, Proxomitron, Proximodo, Privoxy, etc. These were quite popular around the turn of the millenium... The downside is that setting one up to filter SSL correctly is a pain (for the ones that do support it) and now, thanks to Superfish, the general public probably has a fear of all things MITM, user-controlled or otherwise. On a mobile device it's rather difficult to install and manage one, so hosting it remotely is better, but probably most reliable way to block ads there.

A HOSTS file is another way. It works for blocking ad-CDNs quite effectively, but otherwise is very coarse-grained (easily defeated by ads hosted on same domains as non-ads.) Easy to use for desktop; will likely need to be done through DHCP/DNS server for mobile.

As much as I advocate adblocking, I only do it to the ads that are truly giving me a hard time. Tasteful static images and text-only ads don't really bother me at all. It's the annoyingly distracting animated, page-obscuring, misleading, screaming, user-tracking ones that get killed.

Ad blocking on mobile is also very difficult. Most mobile browsers do no support extensions at all, so it has to go through some ugly proxy trick that requires root.

In the mobile apps world, some ad networks (ej. mopub) have a new type of ad named "Native Ad".

With this type of ads you have control on how you display them, and the intended use case is to make the ad look like a part of your application.

This is an explanatory PDF: http://www.mopub.com/wp-content/uploads/2014/09/Native_Ads_O...

I believe that something similar is what the web needs.

> make the ad look like a part of your application.

That sounds terrible and deceptive for the end-user

so you would rather have a floating banner or full-screen ads like low quality apps do?

The problem with ads is that they suffer from the tragedy of commons. Each individual advertiser wants to be more prominent. All together virtually force you to block ads. In special cases, when a third party reviews all ads, as on PennyArcade, the ads tend to be much less annoying.

>I used a computer without AdBlock the other day, and it was like stepping back into 1999 with every page saturated with the 21st century equivalent of animated GIFs and auto-play MIDIs.

Really? I'm not sure that's entirely true. I don't have adblock on the temp laptop I'm using because I just forgot to install it and I have barely noticed any change in my browsing habits.

Another thing that really gets me, given responsive design, is the lack of responsive ads... ad sizes that can go from leaderboard, to tranditional, to micro (phone) sized... as you stretch/reduce your screen size...

As it is, my plan is to render whatever original size, and resizes just won't see/render ads (flux/react)... it's too much of a pain to do more than that, and don't want to show multiples of ads for a single view.

I don't think things need to be plain text.. but if limited to HTML+CSS only, that would be a pretty great start.

It's true that ads use bandwidth and battery but ad-blockers and web standards implementations built into browsers are not "free" either.

It takes CPU to find and strip out ads, and bandwidth to contact filtering servers or whatever else guides the behavior of blockers (and generally the more blocking you opt in to, the more of this there is). I have seen an astounding drop in Chrome performance when using HTML5 YouTube videos on my laptop, where attempts to scroll and click take seconds to even respond and I haven't yet figured out why.

I would still expect the cost to be less overall without Flash but it would be interesting to measure everything that is really going on and figure out if ads are really the worst offenders.

Could you share how specifically and on what levels are your blocking adds?

Off-Topic: Can you expand on why you feel data caps are "draconian"?

Almost 1500 days ago I made this prediction https://news.ycombinator.com/item?id=2300123

"(I can't wait for my quad core computer to grind to a halt with dozens of unblockable canvas advertisements wallpapering every website I go to from now on...sigh, the end is near"

The death of Flash doesn't necessarily mean a better world.

Here. Have a cookie...?

Question: not knowing anything about browser plugins, will it be possible to build a "Click to Canvas" plugin, similar to the Click to Flash plugins avialable? It seems like anything that can access the DOM can insert stuff into (override) or outright block elements, right?

I'm sure we'll see such a thing. The problem is that it'll just shift to somewhere else in the browser, and killing off all the codepaths that might be doing something to the non-working canvas might be subtly hard.

It'll stay a game of cat-and-mouse for a long time I'm afraid.

Good points. Thanks. I guess it's inevitable given the free (as in, subscription free) nature of web content. Heck, even paying a subscription to the NYT doesn't opt you out of seeing their ads, unless you take matters into your own hands with something like AdBlock.

https://github.com/kkapsner/CanvasBlocker

" when action script will easily be converted to html5, then all mobile apps will finally move to the cloud"

that thread is a gold mine.

Indeed, converting Flash ads into HTML means they can be far more subtly integrated into the page in a way that makes them much harder to separate out and block. With Flash, there was just one thing to block and most of the ads would be gone.

I have a feeling that the trend towards client-side rendering and single-page apps is partly driven by this too; when the JS for displaying ads is part of the JS for the app's core functionality, it gets more difficult to distinguish and remove.

perfect.this is very helpful.thank you

I browse with javascript disabled, at first I thought, most websites would be unusable, but I barely feel the difference on most sites.

Bonus: Pages load way faster.

Most websites (eg. youtube, github, reddit) are unusable without javascript. HN is a rare exception, but even here, the search doesn't work at all without javascript.

I use scriptblock

https://chrome.google.com/webstore/detail/scriptblock/hcdjkn...

I have youtube and google allowed for javascript.

Reddit browses very fast, ImgUr is kind of a headache because it doesnt load all images, but both are usable.

Bonus point, the extension doesnt work on incognito,so if you want to see certain website, launch it in an incognito window to browse it quickly.

The only major problem with reddit, is that I can not comment... but I can live with that, and when I feel the need to make a comment, I just allow the site temporaly.

Those are sites that most people know about and use frequently (maybe not Github), just like Facebook, but that's a very different thing than "most websites". There are tons of other smaller sites that don't need JS at all.

I have it off by default (whitelist), and most of the sites I come across just as search results don't need it to show the information I'm looking for.

I run the flash blocker extension in Chrome. So I expect, the next extension I'll be running is a <video> tag blocker with an option to enable on demand. (edit: known as "click to play" mode)

Or just use an adblocker and you don't have to worry what format the ads are in.

Adblockers don't catch everything, and also on some sites I am happy to support via Ads but I cannot stand being interrupted with auto-play video. I've really enjoyed the complete silence from ads in Flash format and am not looking forward to a repeat of the early 2000s.

+1, I starting using click-to-play _years_ ago as a quick way to make sure that YouTube didn't autoplay when I opened it in a new tab. I don't use ad blockers for ethical reasons, so a pleasant side effect turned out to be neutralizing the most egregious behavior that ads would display (by and large autoplay etc tend to be Flash ads).

Running an adblocker when using Chrome on Android is non-trivial.

fyi you dont need an extension, Chrome has native click to play support

So when are they gonna convert Google Finance to HTML5 instead of Flash?

There's no financial benefit there. But, for ads...

Also relevant: https://www.eff.org/deeplinks/2015/01/new-drm-boss-same-old-...

HN thread: https://news.ycombinator.com/item?id=9013211

We dove into this at work (where we make a lot of flash and HTML ads) and it took a 40k flash ad and made it into a 100k HTML add that also required the nearly 500k swiffly.js file to run.

So it's a start but it's not magic nor replacement for a competent human just yet.

Speaking of ads, what can block that obnoxious "THE BEST OF VB, DELIVERED" that pops up like a visual air horn just as you're starting to grasp the content of the first paragraph?

Since I didn't see such a thing, I'm betting it's NoScript.

HTML5 and Javascript are cancer..

This looks like an optional conversion step for advertisers that want to target mobile devices that don't have Flash. So not wholesale replacement of Flash ads.

I'm torn.

On one hand, this means that there will be less flash overall polluting the net. A good thing.

On the other hand, it also means that the polluters have less incentive to actually remove their flash-crap. Someone else will transform their guano into form that is more widely deliverable. A bad thing. Flash will persist even longer.

And of course, flash blocker will now kill less ads than before. Damn.

> And of course, flash blocker will now kill less ads than before. Damn.

That was my concern too, but I guess the only difference will be you'll have to block a bunch of scripts and images instead of blocking just one flash file.

Why does it matter that Flash persist if it's being converted to html5 on the fly anyway? By the way, the swiffy service Google offers has a extension to Flash (the authoring tool) so it can output html5 directly.

Hopefully they will finally fix the bug in Chrome where sites that require Flash pop up a "this site requires flash! - would you like to install it?" bar at the top of the screen, even though it isn't bundled with the app.

It is particularly annoying considering their own efforts to remove Flash from usage on YouTube, that is, with supporting HTML5 video.

This is really good as somebody who uses google ads from time to time, albeit rarely. It's time for flash to take a walk.

I've turned off JavaScript in the browser. And only allow it for sites where I really need it.

I though this day would never come.

On the plus side: death to Flash ads!

On the negative side: life to more ads!

What are these "ads" you speak of?

why swiffy? Why not use Shumway? Is this just a case of "Not invented here" syndrome?

Looks like Swiffy was started first.

http://en.m.wikipedia.org/wiki/Google_Swiffy

http://en.m.wikipedia.org/wiki/Shumway_(software)

And I think Spaceport (now defunct) was started even earlier than that. Shame Google didn’t acquire it.

Swiffy is an offline publishing tool to convert SWFs to SVG. Shumway is a client-side SWF interpreter.

On a related note: can someone explain the advantages of Swiffy over Shumway? Upon first glance, it looks like Shumway has the ability to translate Actionscript to Javascript and the Flash visual elements to HTML5 elements. Whereas with Swiffy, the SWF is just converted to what's basically an animated SVG. What would make Swiffy more attractive for Google to use over Shumway if Shumway seems to be a more complete solution?

Most Flash ads don't really have a great deal of ActionScript in them beyond the odd call to stop and play in the timeline.

Shumway is a very heavyweight dynamic recompiler and includes a complex canvas/webGL rendering engine. Shipping it just for an ad is a bit much.

Naturally, Swiffy's method is quite limited in what Flash features it can support, but that's OK for its purpose.

Uhh, hooray.

Reply on news.ycombinator.com