6 ms·

Fraud is priced in on the industry. Advertisers assume a percentage of their money is going to be wasted on bots, networks (like Google) assume they will misid

by mpeg 8y ago

Fraud is priced in on the industry. Advertisers assume a percentage of their money is going to be wasted on bots, networks (like Google) assume they will misidentify a percentage of impressions and have to reconciliate the numbers, but also that they'll miss some. It's very easy to demonise Google and think they stand to benefit from this all, but in reality they are hurt by it as much as anyone else, and if there was guaranteed zero fraud in their network they could raise prices and increase profit.

ehsankia 8y ago
The fact that they put so much money and effort into unraveling this extremely complex fraud says all that you need to know.
- nindalf 8y ago
  Apparently not for some people, like the current top comment.
  - ehsankia 8y ago
    Right, my point was that no matter how hard you try, no matter what you do or what you say, there will always be people who assume malice from it all.
wu-ikkyu 8y ago
>It's very easy to demonise Google and think they stand to benefit from this all, but in reality they are hurt by it as much as anyone else False if bot traffic > estimated bot traffic For others like FB and Twitter, it's a fact their active user counts are inflated by bots and that they are disincentivized to accurately account for it.
- mpeg 8y ago
  This article and research is talking about third party advertising, where an ad network is used to serve ads on third party sites (greatly simplified, as in reality there could be lots of other players in the chain) What you're talking about is a completely different scenario, where the publisher is also the ad network (FB, Twitter) so inflated user activity could be beneficial for them. That's also arguable as they provide excellent metrics and different pricing models (CPM, CPC, CPA) so if their numbers were as bad as you think more advertisers would just drop FAN. In the case of 3ve, methbot and the like, the process for the fraudsters is different and has nothing to do with bots in social networks: 1. In online advertising, it's very difficult to know who is authorised to serve ads for a given domain. Some domains have hundreds of ad suppliers and they can change all the time. This is alleviated by ads.txt, an IAB initiative to list all your adtech partners in a machine-readable file on your web root. 2. Since it's hard to know what inventory a player in the middle of the chain really has access to. So you can make a deal with an ad player to serve ads on the new york times, but actually be delivering shit inventory. This happens a lot at a smaller scale with dodgy ad vendors that simply deliver lower quality inventory; and added to this, until recently there was no way of cryptographically signing bid requests on OpenRTB protocol (v3 adds this feature) 3. If you have access to lots of consumer IP addresses (3ve uses a large botnet) you can essentially create a hosts entry for publisher.com pointing to localhost, serve a site with fake content loaded with real ads, open a real browser to it and collect on the money from the partnerships and accounts you opened. The advertisers think they're buying real inventory and these users pass every check (browser, network...) PS. I do understand the apprehension though, I spent 10 years in adtech and at times the ecosystem is a total clusterfuck, but it is slowly getting better.
avita1 8y ago
It's impossible for Joe ads buyer to price in fraud, because he doesn't actually know how much fraud is happening. Only Google/the other big exchanges have information to get even close to knowing how much fraud is happening and they have a vested interest in letting fraud slip through and not detecting. And yes, Google cooperated in this specific case, but only after the problem got really bad. It's really hard to take a charitable view when Google, a large player in the ads space, and a company known for technical excellence, allows stuff like this to happen: > An agency or advertiser looking to place an ad would only see the name of TBS.com in the ad exchange and not know they were buying ad space on a counterfeit site. Then the bots would visit the fake site and view the ad to generate revenue. (This technique of masquerading as a legitimate publisher is called domain spoofing. It’s a long-running issue, though a new standard called ads.txt [1] is helping curb it.) This article (https://digiday.com/media/ft-warns-advertisers-discovering-high-levels-of-domain-spoofing/ https://digiday.com/media/ft-warns-advertisers-discovering-h...) is a year and a half old. Google obviously had the resources to discover this years ago and failed to do so. [1] https://digiday.com/marketing/wtf-ads-txt/ https://digiday.com/marketing/wtf-ads-txt/

by mpeg 8y ago

ehsankia 8y ago

The fact that they put so much money and effort into unraveling this extremely complex fraud says all that you need to know.

nindalf 8y ago
Apparently not for some people, like the current top comment.
- ehsankia 8y ago
  Right, my point was that no matter how hard you try, no matter what you do or what you say, there will always be people who assume malice from it all.

wu-ikkyu 8y ago

>It's very easy to demonise Google and think they stand to benefit from this all, but in reality they are hurt by it as much as anyone else False if bot traffic > estimated bot traffic For others like FB and Twitter, it's a fact their active user counts are inflated by bots and that they are disincentivized to accurately account for it.

mpeg 8y ago
This article and research is talking about third party advertising, where an ad network is used to serve ads on third party sites (greatly simplified, as in reality there could be lots of other players in the chain) What you're talking about is a completely different scenario, where the publisher is also the ad network (FB, Twitter) so inflated user activity could be beneficial for them. That's also arguable as they provide excellent metrics and different pricing models (CPM, CPC, CPA) so if their numbers were as bad as you think more advertisers would just drop FAN. In the case of 3ve, methbot and the like, the process for the fraudsters is different and has nothing to do with bots in social networks: 1. In online advertising, it's very difficult to know who is authorised to serve ads for a given domain. Some domains have hundreds of ad suppliers and they can change all the time. This is alleviated by ads.txt, an IAB initiative to list all your adtech partners in a machine-readable file on your web root. 2. Since it's hard to know what inventory a player in the middle of the chain really has access to. So you can make a deal with an ad player to serve ads on the new york times, but actually be delivering shit inventory. This happens a lot at a smaller scale with dodgy ad vendors that simply deliver lower quality inventory; and added to this, until recently there was no way of cryptographically signing bid requests on OpenRTB protocol (v3 adds this feature) 3. If you have access to lots of consumer IP addresses (3ve uses a large botnet) you can essentially create a hosts entry for publisher.com pointing to localhost, serve a site with fake content loaded with real ads, open a real browser to it and collect on the money from the partnerships and accounts you opened. The advertisers think they're buying real inventory and these users pass every check (browser, network...) PS. I do understand the apprehension though, I spent 10 years in adtech and at times the ecosystem is a total clusterfuck, but it is slowly getting better.

avita1 8y ago

It's impossible for Joe ads buyer to price in fraud, because he doesn't actually know how much fraud is happening. Only Google/the other big exchanges have information to get even close to knowing how much fraud is happening and they have a vested interest in letting fraud slip through and not detecting. And yes, Google cooperated in this specific case, but only after the problem got really bad. It's really hard to take a charitable view when Google, a large player in the ads space, and a company known for technical excellence, allows stuff like this to happen: > An agency or advertiser looking to place an ad would only see the name of TBS.com in the ad exchange and not know they were buying ad space on a counterfeit site. Then the bots would visit the fake site and view the ad to generate revenue. (This technique of masquerading as a legitimate publisher is called domain spoofing. It’s a long-running issue, though a new standard called ads.txt [1] is helping curb it.) This article (https://digiday.com/media/ft-warns-advertisers-discovering-high-levels-of-domain-spoofing/ https://digiday.com/media/ft-warns-advertisers-discovering-h...) is a year and a half old. Google obviously had the resources to discover this years ago and failed to do so. [1] https://digiday.com/marketing/wtf-ads-txt/ https://digiday.com/marketing/wtf-ads-txt/