Ad-tech intermediary lobbyists use Google and FB to massively overstate benefits of tracking, news at 11. Publishers waking up to this after years of being on the losing end of the Lumascape.
This chart gets "old news" reaction from smarties, or a "how can this go on" reaction from others.
But there's another angle: @acfou estimates fraud at 20%, It could be 40%. Google takes 60% of ad spend, $80B in USA last year. 0.2 x 0.6 x $80B = $9.6B; 0.4 x 0.6 x $80B = $19.2B.
But there's another angle: @acfou estimates fraud at 20%, It could be 40%. Google takes 60% of ad spend, $80B in USA last year. 0.2 x 0.6 x $80B = $9.6B; 0.4 x 0.6 x $80B = $19.2B.
The @guardian bought out its own inventory the other year and got paid 30 pence on the pound. They just did a similar test and found 72% ad fraud (domain spoofers) for video ads: mediapost.com/publications/a…. Ad fraud is huge and all exchanges (Google runs the biggest) profit from it.
Refresher: ad fraud = fraud sites/users faking ad delivery/view/action, fooling antifraud scripts on page, getting paid (even 30 days or more later).
Flip side is malvertising: exploit kit loader hiding in ad, criminals pay for low-priced ad slots, make big bucks via ransomware.
Flip side is malvertising: exploit kit loader hiding in ad, criminals pay for low-priced ad slots, make big bucks via ransomware.
Internet & Web standards by default do not guarantee integrity of domain names, or of DOM + other JS APIs (unless your script runs first and prepares the environment with excruciating care). Even https: certificates can be spoofed. Certificate Transparency helps but risks remain.
Who pays for ad fraud? The ad (space) buyers, the CMOs of the brands + agencies they use to market goods & services. Who does not get paid? The legit publishers at the far end of the abuse-train that is the evolved-not-designed ad-tech "Lumascape". An unstable & corrupt system.
Why does the "buy side" (CMOs et al.) fund such an inherently fraud-prone & increasingly duopoly-controlled system, which in spite of all its problems is growing to pass TV. 1/ You never get fired as CMO for spending too little. 2/ Agent-principal conflict => fraud helps agents.
3/ Internet eating TV (Netflix &c). 4/ Ads still pay enough to keep.
No publisher I know of can afford to cut all programmatic ads. The ads.txt file has zero integrity (typos, DNS spoofing, listed vendors reselling inventory under table). Heavy JS "nanovisor" approach is heavy.
No publisher I know of can afford to cut all programmatic ads. The ads.txt file has zero integrity (typos, DNS spoofing, listed vendors reselling inventory under table). Heavy JS "nanovisor" approach is heavy.
The centuries-old (from print media, city paper to department store) "direct sales" model is great if publisher has scale & margin to cover its costs (sales team, extra IT budget, attribution/confirmation tracking partner). Inherently risky programmatic ads adoption still rising.
Long story short: internet ads evolved based on accidental innovations such as the HTML image element, cookie, and JavaScript - especially third party scripts. No easy tech fix can be bolted on to ensure end-to-end properties such as integrity against fraud, for brand safety, &c.
This inefficient fraud+arbitrage market collapses, mostly by 1st + 2nd biggest superpowers taking it over. Many small fry in the Lumascape are dead fish swimming. Simplification by monopoly or duopoly has its own obvious issues, ofc. Users must choose better ways (e.g., @Brave).
In case not obvious: who gets paid when ad fraud is committed? Fraudsters, of course, but also every intermediary along the path from marketer to fake-publisher who takes a fee from the gross spend it receives before passing the net along. Perverse agent vs. principal conflict!
Correction: 1/ You never get fired as CMO for spending too much. Or if that is too extreme, try "you will get fired as CMO for spending too little" ;-).
Ad spend may drive sales but marketing is not science, lots of correlation isn't causation trouble. Bad sales? Blame weather!
Ad spend may drive sales but marketing is not science, lots of correlation isn't causation trouble. Bad sales? Blame weather!
