Live data from Hacker News

Ad Nauseam

hackerfactor.com

1–10 of 59 posts

Re: Ad Nauseam

#2
I admit it, I look at linkbait articles and sites... The problem is that it's crossed a line that there are so many intruisive ads, that the web doesn't work without blockers.

If I happen to click on an article from facebook on my phone, the resulting page shouldn't be something I can't even scroll/read because it's so riddled with ads.

Another part is an extension of what TFA says... they should be held responsible... current techniques are iframes, and when a timeout occurs or it bounces to another ad network, another layer of iframe and tracking scripts runs... if an average ad is 3 layers of iframes, and an average page has 5-8 ads, that'd 15-24 complete extra browser contexts just for ads...

Re: Ad Nauseam

#3
"But keep in mind: not everyone is Google, not ever web site has a huge amount of traffic. With online ads, payment is usually tied to the click-through rate (CTR). The CTR is typically around 1% (actual percentage varies by web site). So if 100 people visit your web page, then 1 person will probably click on the ad, generating a fraction of a cent. If each click pays $0.001, then you need 1000 clicks to earn $1. And if 1% of uses click, then that's about 100,000 visitors."

I'm not sure where these numbers come from, but unless you are in fact running a spam site, and likely even then, revenue per click is going to be higher than a fraction of a cent. As a random data point, it looks like the combined revenue per click from Adsense on our sites is around 30 cents per click at the moment.

Re: Ad Nauseam

#4
Just to throw out this idea I've been thinking about for a bit as a ad-golden-rule: The other end of an ad has to have an identifiable person attached (ideally a citizen or from a non-poor country): An ad-auditor.

So now the other end of the ad is not faceless/identiti-less. If the ad is found to serve malware, there's someone to ban/take action against (like banning from a good-paying ad-audit job for life). Ad-networks that require the golden rule can be white-listed by blockers, and become trusted. Networks that don't are considered malware haven.

Could this work? In the current ad-blocking war, the use of ad-blockers will only rise-and-rise, and something has got to give.

Re: Ad Nauseam

#5
"I think that we need to hold the web sites accountable for the content that they display. If browsers get infected by ads at Forbes or people buy knock-off watches from ads at Yahoo, then we need people to sue Forbes and Yahoo. Remember: these web sites authorized the placement of the ad on their web page."

So effectively what you're saying is that we should eliminate ad networks. There is no reasonable way to screen every ad before it is shown when using an ad network. So in order to be safe from lawsuits, publishers would have to go back to directly contracting with advertisers for every ad. Certainly there would be some benefits to that in terms of reduced low quality ads. The problem is, the added overhead of doing so would put many small publishers out of business. Dealing with individual advertisers is a huge job, with massive economies of scale; it just doesn't make sense for websites that are orders of magnitude smaller than Forbes and Yahoo.

Re: Ad Nauseam

#6

Just to throw out this idea I've been thinking about for a bit as a ad-golden-rule: The other end of an ad has to have an identifiable person attached (ideally a citizen or from a non-poor country): An ad-auditor. So now the other end of the ad is not faceless/identiti-less. If the ad is found to serve malware, there's someone to ban/take action against (like banning from a good-paying ad-audit job for life). Ad-netw…

The problem right now is that it can often be difficult for anyone to tell where an ad ultimately came from. A user will complain to a publisher about a bad ad. So the publisher gets a description of the ad, maybe a url that it points to. The problem is the url goes through various redirects, so the one that the user ended up on isn't the one the ad points to, so it can't be searched. But a bigger problem is where to even search. The unit housing that ad will probably be served by multiple networks, with an algorithm deciding which to serve from based on how many ads that user has seen, where they're located, and various other criteria. Then if that network doesn't have a sufficiently valuable ad to serve, they will either farm it off to another network (and so on), or pass it back to the publisher's ad server, which will find another network to serve an ad there.

This all happens in real time. So the point is, when you get a report of a bad ad on your page, it's almost impossible to even know what network it came from. The networks themselves don't know if they ultimately served you the ad, because maybe they got it from someplace else. And no one can search for it based on the url anyway.

Now, none of those things is unsolvable, although it would take significant new regulation. For instance, when an ad is served through a network, there should be a standardized way to add metadata to the ad to state that it was served via that network. In cases where it is passed through several networks, it would carry each of their metadata, in order from the original source through the various levels until the network that actually serves the ad to the publisher. That would at least allow savvy users to make an informed report to a publisher when they get a bad ad. Something else to look at might be requiring that either 1) the target url of an ad points directly to the eventual landing page, or 2) if a redirect is made, the original url be encoded either in the new url (as a fragment id perhaps) or at least as metadata in the page. There are probably plenty of caveats there. But if a user clicks on an ad and finds themselves at some page, there should be some way to figure out what ad took them there. That isn't currently the case.

Identifying the networks an ad has passed through would be the responsibility of those networks (with a standardized way of doing so). Avoiding or identifying redirects would be the responsibility of the advertisers, but networks would have to be required to periodically test ads for compliance.

Re: Ad Nauseam

#7
post #5

"I think that we need to hold the web sites accountable for the content that they display. If browsers get infected by ads at Forbes or people buy knock-off watches from ads at Yahoo, then we need people to sue Forbes and Yahoo. Remember: these web sites authorized the placement of the ad on their web page." So effectively what you're saying is that we should eliminate ad networks. There is no reasonable way to scree…

> So in order to be safe from lawsuits, publishers would have to go back to directly contracting with advertisers for every ad.

Or they would be forced to seek agreements with ad networks to cover such faults. Insurance, in some form.

Re: Ad Nauseam

#9
post #5

"I think that we need to hold the web sites accountable for the content that they display. If browsers get infected by ads at Forbes or people buy knock-off watches from ads at Yahoo, then we need people to sue Forbes and Yahoo. Remember: these web sites authorized the placement of the ad on their web page." So effectively what you're saying is that we should eliminate ad networks. There is no reasonable way to scree…

> So in order to be safe from lawsuits, publishers would have to go back to directly contracting with advertisers for every ad. Or they would be forced to seek agreements with ad networks to cover such faults. Insurance, in some form.

Right. I actually considered that when writing the post, then didn't end up mentioning it. I suppose that's possible, although I find it unlikely that large ad networks would indemnify small publishers in that way. Perhaps if they felt they had no alternative though.

Re: Ad Nauseam

#10
post #6

Just to throw out this idea I've been thinking about for a bit as a ad-golden-rule: The other end of an ad has to have an identifiable person attached (ideally a citizen or from a non-poor country): An ad-auditor. So now the other end of the ad is not faceless/identiti-less. If the ad is found to serve malware, there's someone to ban/take action against (like banning from a good-paying ad-audit job for life). Ad-netw…

The problem right now is that it can often be difficult for anyone to tell where an ad ultimately came from. A user will complain to a publisher about a bad ad. So the publisher gets a description of the ad, maybe a url that it points to. The problem is the url goes through various redirects, so the one that the user ended up on isn't the one the ad points to, so it can't be searched. But a bigger problem is where to…

With the ad-auditor idea, the model itself changes. The middlemen go away, and there's a person at the other end. Yes, the business of ads will change per this new model.

An analogy is financial-auditors -- a human has to be present and sign even if the report is for a company behind 10 shell companies.

Post reply on HN