5 ms·
I think it's more likely that ad networks will come up with a way for publishers to start serving the ads from their own domains and with randomized filenames e
by modeless 11y ago
I think it's more likely that ad networks will come up with a way for publishers to start serving the ads from their own domains and with randomized filenames etc. You can't block ads if you can't differentiate them from the content in an automated way.
- seiji 11y agoYou give them way too much credit. The simplest ad blocker is just a user stylesheet that hides "class='ad'". The complex ad blockers include variants of that as well as network-blocking bad domains, but publishers+advertisers are really really dumb in aggregate.
- modeless 11y agoThey're not "dumb", it's just not worth the effort for them to obfuscate their ads yet. That will change.
- seiji 11y agoPostulate: there's no automated ad-injection platform that also can't be hidden or blocked by automation. Google could transcode ads directly into YouTube videos, but they don't. They preroll or popover and both can be blocked. If YouTube did transcode pre-roll ads, we'd have an auto-skipper. If YouTube transcoded hard popovers on videos, well, that would probably upset more video authors than consumers.
- untog 11y agoEasy - they just get more pernicious. For an example, try this product placement in I, Robot, back in 2004: https://www.youtube.com/watch?v=3HpIZrOH4zc https://www.youtube.com/watch?v=3HpIZrOH4zc Even though it's set in the future they manage to sell them as being retro cool. It's almost impressive. I take the pessimistic viewpoint that ads are going to seep in no matter what. The harder you make it, they more they'll mask themselves. Personally, I'd prefer banner ads.
- seiji 11y agoproduct placement != invasive advertising. advertorials blur the line a bit, but it's not the same as plastering animated "they hate this ohio school teacher for knowing the secret of eternal youth!" ads all over my news. (free bonus: every 2,000th ad includes a free flash-delivered root kit!) plus, with a transparent AR setup with proper overlay ability, you can in-reality block photons bounced off or emanated from advertising units (billboards, bumper stickers, old spice commercials, etc) from even hitting your retinas.
- untog 11y agoAgain, IMO, but product placement is worse. I know invasive advertising is advertising. Mentally speaking it's quite easy to disregard. Product placement slowly blurs the lines to the point where you don't know if someone is honestly advocating for a great product or if they're being paid to do so.
- seiji 11y agoIf you are watching something produced by a production company (anything on TV or any movie) every brand you see is paid placement. It's like how before a movie comes out, all the actors in the movie go on all the talk shows. Think it just happens by magic coincidence? It's contractually part of their job to appear and promote the properties they appear in so it drives up demand for their productions (movies/books/tv shows/congresscritters).
- moonchrome 11y ago>If YouTube did transcode pre-roll ads, we'd have an auto-skipper. Don't respond to user seek commands until the add is done playing - at that point best they can do is block output and wait - which ends up a similar experience to just watching the ad for classic 2-10 min video youtube browsing or songs - for longer content buffering could work.
- wtallis 11y agoYeah, but is it really worth it for the server to keep track of when the ad should be done playing and refuse to deliver the rest of the stream until then? Because trying to do that in Javascript is a losing game.
- qu4z-2 11y agoFor me personally the change between "30s pre-feature ad" to "30s pre-feature darkness/silence" is more important than the change from there to "start playing straight away". I don't mind waiting, but I don't want things fighting over my attention.
- Encosia 11y agoI get the impression that today's CTR-driven advertisers would prefer not to show ads to folks who are actively trying to block them, because they would almost never have generated useful clicks anyway. Kind of like how online scammers purposely don't even try to sound legit, so that their leads are better qualified when the gullible do eventually respond. Effort invested in technical solutions to work around ad blockers might just lower CTRs and be worse for everyone involved. (edit: I don't mean to insinuate that online advertising is anything like a scam; just that there are parallels in terms of targeting the correct audience so that the effort is effective.)
- specialk 11y agoI know this probably sounds counter-intuitive but the company that provided the report mentioned twice in the article also have a blog post about how adblock users click more ads than non-adblocking users[1]. There's not much detail in the article to back up their claim but it's an interesting idea to play around with. [1] http://blog.pagefair.com/2015/adblock-users-click-more-ads/ http://blog.pagefair.com/2015/adblock-users-click-more-ads/
- Encosia 11y agoThat is definitely interesting. I wonder if it's because they're more likely to be contextually interested in ads on a site that they would go to the effort of whitelisting, but that wouldn't actually be the case overall if they were seeing every site's ads?
- philwelch 11y agoI think it's because people who use adblockers aren't used to tuning out ads.
- icanhackit 11y agoCould you create checksums/signatures of ad videos/banners then add the signatures to the distributed block-list? I guess the next step would be to automatically add noise/tiny amendments to the content to get around the checksum...but then you could use object/character recognition to identify content and block the ads based on that.
- seiji 11y agoYou're looking for perceptual hashing: http://www.phash.org http://www.phash.org
- chrischen 11y agoAt this point the CPU resources consumed hashing every image from a website will probably outweigh the cost of just displaying the ads (since they are downloaded anyways).
- icanhackit 11y agoWould hashing every image be much more resource intensive than rendering the images? Images would only need to be hashed once whereas images need to be rendered in various states at ~60Hz when a user scrolls through a website.
- joosters 11y agoI wish they did that today: no more 3rd party requests that will track you!
- eli 11y agoThat would be really really hard to pull off for most publishers. As a simple example: it breaks ad networks, which rely on you having a cookie that is shared across many domains to figure out what ads to show you. That doesn't really work if that 3rd party script is blockd.
- zmmmmm 11y agoIt just means the publisher has to proxy delivery of the ad for the ad network. This will not be hard. It's effectively what happens inside apps anyway - the ad network gives you an SDK with an API and you integrate it in to your web site.
- eli 11y agoNo, that won't do it. The ad network won't know who you are because it can only see what you've done on that site. With no way to carry an identifier between different domains, the value of the ad network is destroyed and ad techniques like retargeting (when you put something in your cart on one site and see a coupon for it on another) are impossible.
- zmmmmm 11y ago> because it can only see what you've done on that site I'm not sure what the problem is - all they need is to ask the publisher to set the cookie for them. Then it's a first party cookie that is relayed on the back end to the ad network, who gets all the same information from it that they used to get.
- eli 11y agoI go to Site A and get a cookie. Now I got to Site B and since it's on a new domain I have no cookies set and am basically a brand new visitor, so I get another cookie. Even if both those cookies are being relayed to some backend, how can it ever tell that I'm the same person on Site A as Site B?