Tweet

Dr. Augustine Fou

Follow @acfou

24 Feb, 6 tweets, 2 min read

this is so hilarious, it's not even funny

bad guys don't have to run any ads or use any bots to make lots of money

they've already optimized this supremely. How?

they are flooding the networks with faked bid requests

imagine 1 million QPS (queries per second), all computational (minimal bandwidth used)

even if only 1% gets through (99% caught by some anti-fraud mechanism), that's 3 x 10^8 ads they get paid for (assumes 1 in 100 bid win)

and the good guys are left trying to figure out where the money went and how to get refunds for it, if it is even detected/caught

remember when I said "marketers paid millions of dollars for excel spreadsheets"

a lot of the ads never even ran; there was NO ads, just excel spreadsheets memorializing the amount bought and paid for, and the clicks they got.

damn.

more details coming out
linkedin.com/pulse/stone-me…

“Defendants knew that a substantial portion of the mobile inventory they sold to Uber .. was NONEXISTENT, nonviewable and/or fraudulent, and that such inventory was not attributable to legitimate riders installing the Uber App,”

“Defendants actively concealed nonexistent, nonviewable, and/or fraudulent inventory and prevented Uber from uncovering the true facts, for example, by hardcoding misleading names into TUNE to deceive Uber into believing installs were driven by ads on approved sites.”

• • •

Missing some Tweet in this thread? You can try to force a refresh

This Thread may be Removed Anytime!

Twitter may remove this content at anytime! Save it as PDF for later use!

More from @acfou

Dr. Augustine Fou

@acfou

25 Feb

oh lovely, adtech works

in my 1 cent CPM experimental campaign, you can see the price paid to the publisher - around 0.8 cent CPMs

this 20% goes to 1 intermediary (the DSP); this is the shortest possible programmatic supply chain

data from #FouAnalytics

what do I get for 1 cent CPM ads?

oh goody, look at these devices/IP addresses loading my ad 100s of times, despite my having set freq cap of 1 (lifetime)

oh, and in the 28% of the impressions where "intersection" was measurable, the intersection was ZERO 65% of the time.

this means the ad did not intersection in any way with the viewport (the portion of the browser the user can see)

Read 5 tweets

Dr. Augustine Fou

@acfou

18 Feb

@RioLongacre

@RioLongacre @garicruze confirmed.

that is a great example -- basic "due diligence" is what is needed to see the fraud and do something about it

forbes.com/sites/augustin…

@RioLongacre

@RioLongacre @garicruze and confirmed, less than 50% of an advertiser's dollar goes to "working media" (showing digital ads) when they buy through the long, complex, and non-transparent programmatic supply chain

forbes.com/sites/augustin…

@RioLongacre

@RioLongacre @garicruze and that's not even taking into account how crappy targeting is (because it's all derived data)

forbes.com/sites/augustin…

Read 4 tweets

Dr. Augustine Fou

@acfou

16 Feb

good publishers don't have a big bot problem; they've been consistently clean all along

forbes.com/sites/augustin…

@Chronotope

+ @Chronotope @jarroddicker

as long as the publishers don't source traffic and use Taboola/Outbrain everywhere, they don't have a bot problem because fraud bots are practical - they go to sites that pay them for traffic

forbes.com/sites/augustin…

Read 4 tweets

Dr. Augustine Fou

@acfou

14 Feb

a short story, in 5 frames

1/5

2/5

3/5

Read 5 tweets

Dr. Augustine Fou

@acfou

12 Feb

I literally don't know what they are smoking to make this statement

“The trade in traffic hasn’t been eliminated, but it is substantially harder to buy fake traffic and it’s substantially more expensive to buy realistic-looking fake traffic that will actually get the buyer paid"