January 15, 2026 • 26 mins
Have a question? Send us a text!
In this episode, Tim Rowe sits down with Dr. Augustine Fou, Creator of FouAnalytics, to pull back the curtain on the multi-billion dollar ad fraud industry. They discuss why fraud isn't a tech problem but an incentive problem, how to spot the red flags of spoofed CTV inventory, and why the most powerful tool in a marketer's kit might be the pause button.
Key Takeaways
Ad Fraud is an Incentive Problem Fraud persists because the ecosystem is designed to reward volume. Dr. Fou explains that middlemen, exchanges, agencies, and tech platforms, make more money when more traffic flows through their pipes, leaving them with little financial motivation to filter out the bots.
- 0:00 – Why throwing more tech at fraud won't solve an incentive issue.
- 2:42 – The red flag of 100% click-through rates and how bot mechanics work.
- 5:20 – Transitioning from manual McKinsey-style audits to the FouAnalytics platform.
The CTV Conundrum & The CPM Trap In Connected TV, fraud is binary, it’s either 0% or 100%. Buying direct from premium publishers is safe, but chasing efficient CPMs on programmatic exchanges often means buying spoofed bid requests that never reach a real television.
- 6:23 – How fraudsters pretend to be Disney+ or ESPN to hijack programmatic budgets.
- 9:21 – Why low CPMs are actually driving up your total waste.
- 16:38 – How independent audits prove if your CTV ad actually ran on a TV.
Correlation vs. Incrementally Marketers often mistake concurrent sales for successful advertising. Dr. Fou breaks down how attribution models over claim credit for sales that would have happened anyway and why turnoff tests are the only way to find the truth.
- 11:42 – Why View-Through conversions are often used to hide fraudulent traffic.
- 14:12 – The Uber Case: Cutting $100M in spend with zero impact on app installs.
- 15:30 – Adopting the Small Business Mindset to focus on real business outcomes.
Connect with Dr. Augustine Fou on LinkedIn here or visit FouAnalytics.com.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Tim Rowe (00:00):
Ad fraud isn't so much a tech problem as it is
fundamentally an incentiveproblem.
You see, middlemen make moremoney when more volume flows
through their systems.
So they have little financialmotivation to cut off fraudulent
traffic, and fraudsters aresophisticated, continuously
evolving to bypass detection.
So today, we're learning abouthow to combat ad fraud, how it
(00:24):
actually happens, and simpleframeworks for avoiding it
altogether.
We'll learn about the red flagof low CPMs, why most reporting
models overclaim credit fortheir contribution, and we'll
hear about what happened whenUber turned off $100 million in
ad spend.
Welcome back to this day ofstreaming podcast.
I'm your host, Tim Rowe, andtoday I'm joined by Dr.
(00:45):
Augustine Foo, creator of FooAnalytics, an analytics pixel
for monitoring programmaticmedia investment for the purpose
really of identifyingfraudulent traffic inventory and
optimizing investment.
So if you want to understand adfraud once and for all, how it
shows up in streaming TV, andmost importantly, how to avoid
(01:06):
it, then this is theconversation for you.
Let's get into it.
Dr.
Fu, why is ad fraud such a bigproblem?
Dr. Fou (01:21):
It's gotten bigger and bigger over the years.
And the way I like to describeit is that ad fraud is not a
tech problem.
It's actually an incentivesproblem.
So it's uh it's kind of easy tounderstand if if you think
about the middlemen, uh, youknow, like the ad exchanges,
even the agencies or other adtech companies, uh, when there's
more volume that flows throughtheir pipes, they make more
(01:44):
money.
So they're not in a hurry tocut down the fraud because
that's gonna cut down theirrevenue.
So in that case, that's that'swhy I explained it as more of an
incentives problem.
And because it's an incentivesproblem, throwing more tech at
it is not gonna solve it, right?
So the bots and the fraudstersare very clever.
They know how to get around thedetection and avoid getting
(02:06):
caught, right?
Bad guys have a lot of practicedoing that.
So that's why fraud has gone upuh over the years and not down.
Tim Rowe (02:13):
That is that is a great point.
They've gotten better at it,and there is a strong incentive
related to uh to figuring outhow to be good at being good at
bad stuff.
Being bad.
Yeah, getting good at beingbad.
You've developed a series oftechnologies and frameworks and
methodologies for thinkingthrough and identifying and
(02:33):
ultimately resolving theseissues.
We'll talk about that, but I'dlove to talk about some of the
really great examples that youshared through some of your
published work.
There's a piece that youshared, 100% click-throughs on
pharmaceutical ads.
Can you tell us about thisstory?
How did how did someone get ahundred percent click-through
rate?
Yeah.
And what does that mean?
Dr. Fou (02:54):
Um that's how I actually got into the fraud
detection side of things becauseI was uh working at an agency
in Omnicom, and we were servinga lot of pharmaceutical clients
and med device clients.
So they were primarily doingsearch campaigns.
And then when you actuallylooked at some of the data, it's
like, what the heck is going onhere?
If you just take the rolled upaverage, the campaigns were
(03:16):
performing like you know, 9%,15%, 25% click-through rates on
average.
But then when you actually pullthe place report by domain,
you'll say, this domain has 100%click-through rate.
This other domain has 100%click-through rate, sometimes
greater than 100%, 100%click-through rate.
So, you know, if you're justlooking at the rolled-up
averages, you're gonna miss it,right?
(03:37):
It's gonna seem like, oh, it'sgreat, you know, this campaign
is performing so well.
But then when you look at theline items, the details, you'll
see stuff that just doesn't makeany sense.
So back then, this was you knowalmost 15 years ago, you know,
no one could really explain whatit was.
They were all just cheering, ohyeah, the campaign's working
really well.
It's like that doesn't make anysense because humans don't
(03:57):
click on ads that much,including search ads, right?
1% would be a good rule ofthumb, maybe 2%, but definitely
not 100 or 200%.
So, long story short, we nowknow it's the bots that are
clicking on the ads, and they dothat because it's a cost per
click model.
So they have to click it inorder to get the revenue.
(04:19):
And so for a lot of theadvertisers, you know, when when
they run a search campaign, alot of the ads do run on
Google.com, right, where humansGoogle things.
But there's also somethingcalled the search partner
network, where there's all theseother sites that run code on
the page to run search ads.
So when there's a search adthat shows up and someone clicks
(04:42):
on it, the site gets some ofthe revenue and Google gets some
of the revenue, right?
It's kind of shared.
So then you can see how thesites have the motive, right,
the incentive to commit clickfraud using bots because they
can just click on ads all daylong.
So then they can make moremoney for their site.
So that's kind of what we'refighting, right?
And when we looked at the dataand you know, it didn't quite
(05:04):
make sense.
That's when I started digginginto it and I built some
technology tools to help me dothe audits, right?
To kind of say what the heck isgoing on here.
Long story short, is uh it'skind of evolved into a platform,
right?
I didn't set out to build fluanalytics, uh, it was more of
some tools to help me auditcampaigns.
And I was originally deliveringit through PowerPoints, right?
(05:25):
Kind of like McKinseyconsulting projects, but that
doesn't scale.
So by 2020, I just decided I'mgonna open up the platform and
let others use it.
So just like people learned howto use Google Analytics for
their websites, they can nowlearn how to use Foo Analytics
for their digital ads.
So that's uh I just gave it aname, Foo Analytics, and then
(05:46):
added user accounts.
And now majority of ourcustomers are the big
advertisers because theirmarketing folks and their
analytics folks are now equippedwith an analytics platform that
they can log into, look at thedashboard, and then figure out
where the bad stuff is, right?
Uh which are the bad sites,which are bad uh mobile apps,
(06:08):
and then just add those to ablock list and clean up their
campaigns.
Tim Rowe (06:11):
Huge.
Uh how does that change or howdoes that look different for
television, for CTV, streaming,for video?
How they're not not clickablenecessarily.
How does fraud show up insideof video?
Dr. Fou (06:26):
Well, it's you know, the click is like the outcome of
it, right?
So you show the ad and thensome portion of the users might
click on it.
But what's more important isthat a portion of the ads are
not even shown to humans, right?
And those a portion of thoseads are going to fake sites.
So that's what we're trying tostop, right?
Even before the click happens.
So if your ad is running on afake site, humans are definitely
(06:47):
not seeing it, right?
And so that's that's theproblem where we're measuring
the ads where they went andtrying to block those before the
ads go there and before themoney goes there.
Because once the money's there,it's really hard to get back,
right?
So a lot of people say, ohyeah, we can get refunds.
Once the money is in the badguys' pockets, you're never
getting it back from them.
(07:08):
So ultimately, it's going to besome middleman like maybe the
exchange that you bought fromthat's going to basically eat
the cost and give it back toyou.
So it's always better to notlet the money go to the bad guys
in the first place rather thantry to get a refund afterwards.
Tim Rowe (07:24):
So step one, prevent the money from going there in
the first place by havinganalytics.
Dr. Fou (07:30):
Yeah, having good analytics.
Yeah, because first you have tosee that you have a problem.
And the reason I'm using theword analytics as opposed to
fraud detection is that I thinka lot of people are familiar
with the legacy fraudverification companies, which I
won't name.
Basically, they just give you aspreadsheet and a dashboard
that says 1% IVT or invalidtraffic.
(07:50):
When you see that and nothingelse, there's really nothing you
can do with that number.
Right.
So the reason I'm calling myplatform analytics is that it
has the supporting details sothat you can go in there and
understand why something wasmarked a bot or why something
was marked, you know, there'sother forms of fraud other than
bots.
So if you say, okay, Iunderstand that, and these are
(08:11):
bad guys, then let me add themto a block list.
Right.
So these are kind of a processthat you go through maybe once a
month, uh, maybe once a week ifif you want to.
You don't have to do itcontinuously, but it gives you a
way to monitor where your adsare going and then continuously
kind of improve and optimizeyour campaigns as you go along.
Tim Rowe (08:31):
Cool.
So we're building a block listas we go along of our of our
known bad actors.
How is inside of CTV, how howare, I guess, how is the fraud
taking place?
Thinking about, let's thinkabout like our local seller
who's trying to explain theadvantages of CTV to uh a small
(08:53):
business owner, but they'relike, hey, I saw an article that
said fraud's gonna take mymoney, so I'm gonna put it on a
billboard instead.
How would you explain that tosmall business owners?
Dr. Fou (09:03):
So in CTV, um, there's actually a conundrum here.
So let me let me kind of putthis out there and let's see how
you how you react to it.
So in CTV, on the one hand, Ican say there's no fraud in CTV
because it's impossible.
And then on the other hand, Ican say there's 100% fraud in
CTV.
Now, how can both of thoseexist and be true at the same
(09:28):
time?
Tim Rowe (09:29):
I don't know, you got me stumped there.
How could that be true?
Dr. Fou (09:33):
So basically, if you're buying from a real seller like
Home and Garden TV, FoodNetwork, ESPN, right?
Obviously, these are onespeople have heard of and
actually subscribe to.
Disney Plus, whatever.
There's no fraud in CTV becauseit's impossible.
Because Disney Plus is nottrying to rip you off.
If you buy ads direct fromthem, the bad guys can't get the
(09:57):
fake impressions in there andcan't get money out.
Therefore, they're not doingfraud in CTV.
However, if you try to buy CTVon programmatic exchanges, and
especially if you're trying tobuy super large quantities at
super low prices, it is not CTV.
It is 100% fraud.
(10:17):
That's because any bad guy canjust upload a bid request and
say it's Disney Plus, say it'sHGTV, say it's a food network,
say whatever they want becausethey want to pretend to be the
legitimate apps so that they canget bids.
Right?
If they put a no-name app inthere, nobody's gonna bid on it.
So it's always better for thefraudster to pretend to be a
(10:39):
well-recognized streaming applike ESPN or Disney Plus in the
bid requests, but they stillhave their own seller ID in
there because they want to getpaid.
So that money is actuallydiverted away from ESPN, Disney
Plus, uh, you know, AmazonPrime, whatever.
Those sellers, the thelegitimate sellers don't even
see that.
And there's a bunch of longtail Roku apps, but that's
(11:04):
actually tiny compared to themainstream fraud.
Okay, so in that case, there'sno fraud in CTV because it's
impossible.
Applies to the case whereyou're buying from legitimate
sellers, ones that you've heardof before.
And let me note that legitimatesellers, including YouTube, has
unsold inventory.
How do I know?
When I'm watching Sundayfootball, you'll see a little
(11:26):
screen that says enjoy the Zen,we'll be right back.
That's an unsold CTV ad slot.
Okay, so even the big guys haveunsold.
That's because the advertisersare chasing low-cost CTV
inventory that's actually notreal.
Okay, they're buying billionsof impressions, and it's it
(11:47):
didn't run on ESPN, it didn'trun on Disney Plus, it ran
somewhere else or didn't run atall.
Okay, and then the the flipside is there is fraud in CTV
and it's 100%, and that'sbecause it's everything outside.
Tim Rowe (12:01):
It seems then like there's a larger conversation
about what a CPM actually is,because if you're paying a
really low CPM, but you're notactually getting the thing that
you're paying for, what the heckwas it worth?
Can we talk about CPMs?
Dr. Fou (12:15):
Yeah, that's actually a very important topic because uh
what I hear all day long iswhen you know agencies and even
the advertisers say, oh, costefficiency, they're trying to
save cost.
Well, let me use a very simplemathematical calculation here,
just based on what I've seenevolve over the last 15 years.
In the early days of digital,you would be paying $35 CPMs to
(12:39):
a publisher like New York Times,you know, any any major
publisher.
So that's $35 per thousandimpressions.
Now you're paying $3 CPMs.
Okay, so people think, oh yeah,we had cost savings.
But actually that's not truebecause CPM is a price.
You still have to multiply bythe quantity.
So now you're buying $3 CPMprices, you're buying 10 times
(13:04):
the quantity, you're stillspending $30.
You see what I'm saying?
That's the cost.
So CPM is a price, it's not acost.
So there's this misuse of theterm cost efficiency.
Oh, yeah, lower CPMs, it's costefficient.
It's not.
So what ends up happening isnow, you know, when the
advertisers are chasing thereally low price like CPMs for
(13:26):
CTV, they end up buying thecrap, like the non-real stuff,
and they're buying lots and lotsof quantity of it.
So the simple advice is don'tbe afraid of paying higher CPM
prices because you're going tobe buying less quantity.
So you're still saving costsbecause fewer of your ads are
(13:46):
going to completely fake sitesand uh mobile apps.
Right.
So that's kind of how to thinkabout a CPM as a price and how
that impacts your cost, right?
So you can really save costs bybuying less of the fraud that's
out there.
Tim Rowe (14:00):
And that segues perfectly into reporting.
If I'm buying a bunch offraudulent inventory, how is it
that my reporting still reflectsthat it's working?
It feels like, okay, cool.
If I understand how fraud worksand I understand that I need to
take measures to prevent fraudfrom happening, and here are
(14:22):
corrective actions I can takeafter it happens to prevent it
from happening again, and Iunderstand the pricing.
How is it actually showing upin my reporting like it's
working?
Dr. Fou (14:32):
Yep.
So it's easily explained if youcan understand the difference
between correlation andcausation.
Tim Rowe (14:40):
Okay.
Let's let's explain.
Dr. Fou (14:42):
Correlation.
Yeah.
So correlation just means salesare happening over here while
you're doing digital marketingover there.
Yes.
They're unrelated to eachother.
So here's the thing (14:52):
like for a lot of the CPG companies, if
you're selling uh soda or ifyou're selling soup or paper
towels or whatever, that happensin the grocery store.
You're doing digital marketingover here.
It's not directly causing salesof paper towels, toilet paper,
soup and soda in grocery stores.
Okay, those are happening,whether or not you're doing
(15:16):
digital marketing.
And there's been someexperiments over the years, not
many, but remember when PGturned off 200 million of their
digital spend?
No change in their sales.
Okay, so those are data pointsthat I've seen that most people
don't want to hear about becausethey all want to think that the
digital marketing is working.
So now let me get to a littlebit more detail to why the
(15:39):
reports look so good.
Okay, so that's a matter ofattribution.
Some of it's fraudulent, someof it's just incorrect
attribution.
So in years past, there, youknow, when some of these ad tech
companies were trying to pushdisplay ads, they were saying,
oh, you know, it's so unfairthat search ads got all the
(16:00):
credit for the sales.
The reason search ads got allthe credit for sales is because
it's usually the last clickbefore the customer purchases
something.
So if you see something on TV,you're gonna say, oh, let me go
Google that, and then you see asearch ad and then you click on
it and then you buy something,right?
Because at that time you'rethinking about it, you're ready
to buy something.
(16:21):
So then that sale getsattributed to the last click,
which was driven by search.
So then all the people who aretrying to sell display ads say,
oh, those that's so unfair.
Display ads must have someimpact and must have helped to
cause that sale.
That is true, but it's not aone-to-one correlation.
It's not like one display ad,you know, a human sees one
(16:42):
display ad and then they go buysomething, right?
You have to see a lot ofadvertising, and then when
you're in, you know, when you'rein the right mood and the right
time, then you're gonna go buysomething.
But because of the display adsellers complaining, Google
basically developed somethingcalled view through conversions.
So that meant even if youdidn't click on the display ad,
(17:05):
it has some kind of benefit,right?
So they're gonna say when aperson is exposed to that
display ad, if there's a sale orpurchase that happens within
the next 30, 60, or 90 days,we're gonna attribute that sale
to the display ad.
That's a very, very rudimentaryway of doing attribution.
And it does work in certaincases, but it's also been now
(17:28):
abused.
All right.
So, for example, you can say,oh, you know, that individual
display ad didn't drive thatsale that happened 30 days
later.
It just happened 30 days later.
And you can see a scenariowhere the person was going to
buy the thing anyway, and theyjust happened to be exposed to
the ad.
It wasn't caused by the ad.
You see the difference?
(17:49):
So the problem is theseattribution or conversion models
over-attribute or overclaimcredit for the sales that would
have happened anyway, or thatwould have happened in the
absence of the advertising.
So, what a lot of advertisersare not yet doing is focusing on
incrementality, which means thesales that would not have
(18:10):
happened in the absence of theadvertising, right?
We actually want to say wespent these dollars in digital.
Did it actually drive moresales than those that would have
happened anyway?
Right?
That's really the key questionthey should be asking.
But kind of to close out yourquestion, it's they're seeing
sales happening, and they haveattribution models that say, oh,
(18:32):
those sales are caused by thead impressions, even though they
weren't.
So that's why a lot ofadvertisers believe that it's
working, even though there's awhole bunch of fraud, like it
could be 90% ads shown to bots,the sales still happened.
But then the model said, Oh,yeah, it's it's you know, those
sales are attributed to thesedigital campaigns that you have
running.
That's why you believe, andthere's a whole bunch of people
(18:55):
who wanted to believe thatdigital works really, really
well.
Okay, so that's how it coveredup the fraud, right?
Even if the campaign had 90%bots in it, impressions were
shown to bots, they'd still hadsales going on, right?
They still had purchases,conversions, all that kind of
stuff.
So they thought it was working.
But careful marketers andprobably more advanced marketers
(19:17):
are now kind of trying to teasethat apart and saying, okay,
well, we can run turn-offexperiments, like just turn off
the digital campaign in onestate and see if the sales in
that state continue at the samerate, right?
Those are simple experimentsthey can run to say, was there a
cause and effect?
And then now to kind of closeout the correlation versus
(19:38):
causation thing.
Correlation is just most of thesales and the digital
marketing, it's justcorrelation, right?
They happen at the same time.
It's not causation.
But I think more and moremarketers are kind of turning on
to that and getting smart andsay, okay, that's not good
enough.
We really have to figure outthe causation.
So there's differentmethodologies, there's you know,
media mix models.
(19:59):
And that kind of stuff.
But you have to be very carefulwhere errors are introduced
into those models and you'restill overclaiming credit for
sales.
So you've got to be careful toonly focus on the
incrementality.
Tim Rowe (20:11):
Someone that's listening today, they're maybe
either buying CTV, they'releading video investment for a
brand, or they're leading alocal sales team and trying to
navigate conversations like thisto drive adoption.
What's the one key takeawaythat you want a listener to
leave with?
Dr. Fou (20:29):
I think the best way to think about this is approach it
as if you were a small businessowner.
Okay, so small business ownershave finite budgets.
If they spend $1,000 and theydon't get any more sales, like
any more people walking intotheir grocery store or anyone,
any more people sitting in thechair at their barbershop, they
can't spend the next $1,000 ondigital ads.
(20:51):
So, you know, the bigmarketers, they have so much
budget.
And they have these media mixmodels and all that kind of
stuff that kind of attribute andtell them things are working,
and you have great ROAS on it,right?
Return on ad spend on it.
Those are the ones wherethey're spending way, way more
money than they should be.
So it's almost like think as ifyou were a small business owner
(21:12):
and you have to actually lookat real outcomes.
And then by the way, when youturn off the spend, the sales
should be going down, right?
Or should go away.
So if it's not, then you cantell the digital marketing or
the CTV ads that I'm buying didnot cause those sales, right?
Did the sales continue when youturn it off?
This was this goes back to, youknow, like an the Uber uh fraud
(21:35):
example from 2018.
They were, they thought theywere very clever, and they said
we would only pay when we getapp installs, right?
That's the ultimate conversion,right?
Get someone to install the Uberapp.
So what do you think the badguys did?
They faked the reporting.
Yeah, exactly.
Well, not even real appinstalls.
They just faked the reportingto make it look like there were
(21:55):
app installs.
There were some of these badguys that were so bold, they
didn't even run ads.
They didn't even have realdevices that installed the app.
So literally, it was all fake.
So what Uber did, and this wasan analytics person there, it
wasn't because of any specialtech or anything.
Kevin Frisch just said, let'slet me just pause the spending
(22:15):
for a week and see what happens.
He paused the spending for aweek, app installs continued.
He paused the app the spend foranother week, app installs
continued.
And he said, Okay, F this, I'mgonna just cut it all off, app
installs continued.
Because humans wanted toinstall the Uber app anyway.
It wasn't because all the spendin those mobile exchanges
(22:40):
caused the app installs.
So, based on that, I mean, mostadvertisers can do that
themselves.
They don't need any specialtech to detect the fraud.
They can just run these turnoffexperiments.
And you don't have to turneverything off, right?
Because you still have to domarketing, but just choose a
market, right?
New York or California orwhatever, turn it off and see if
(23:00):
the velocity of sales andconversions in that market
changed.
If it didn't, then that's adata point for you.
Okay.
Not that many sales andconversions were caused by the
digital marketing.
There may be other things likeactual billboards in the
physical world that you can buybecause people actually walk by
(23:21):
them or drive by them.
Right?
In digital, it's so easy tohide the fraud because
everything's just bits andbytes.
There's so many fake sites, somany fake apps.
No advertiser is going toreview an Excel spreadsheet that
has two million rows in it.
Right?
So they just ignore it and theyjust assume, okay, yeah, they
(23:42):
the vendors tell me it'sworking, so it's working.
So that's where we've gonewrong for many years.
But I think it's time to makedigital marketing better.
Tim Rowe (23:50):
I agree.
And you're working on solvingthat again in 2026.
Foo Analytics has been out fora decade and a half, but you are
launching audits.
Can you tell us about that?
Dr. Fou (24:01):
Yeah, it's it's very simple, but the word audit, I
think, is both scary, but it'salso something that people know
they need to do.
So we've basically put FooAnalytics tags into the ads,
say, for example, CTV ads.
The audit is basically taking alook for the advertiser to see
where the ads are going andwhether the ads even ran.
(24:23):
Okay, so if it didn't run,you're not getting what you paid
for.
And then if those ads went tocrappy websites and mobile apps,
those are not CTV, right?
You paid CTV prices and it wentto crappy mobile apps and
websites.
So those are the kinds ofthings that the audit will help
surface that I don't think theadvertisers currently know
because they're not getting anyof those insights from the
(24:46):
placement reports or even thelog level detail data that they
get.
So it's really taking anindependent look and seeing if
there's areas of improvement.
I'm almost certain that therewill be, especially in CTV,
because the fraud is so rampantand so easy to do.
Tim Rowe (25:02):
Dr.
Fu, this has been a greatinstallment, I think, in
reimagining the way that weapproach digital marketing.
I think that that's the greattakeaway from this is hey,
here's a permission slip to doit different.
The way that it's been donedoesn't necessarily mean that
it's the only way to do it.
And there's clearly a lot ofopportunity here for us to all
(25:23):
get better.
If folks want to learn moreabout foo analytics, learn more
about the audits, connect withyou, read some of the great
content, where should they go?
Dr. Fou (25:31):
Fooanalytics.com, so F-O-U analytics with an S.com,
or just Google my name,Augustine Foo, and I'm all over
the LinkedIn, and I havearticles uh with charts and
stuff like that, so you can seethe data.
Tim Rowe (25:44):
Definitely not hard to find.
We'll make sure that it's alleasy to find close by to this
episode.
Dr.
Foo, thank you.
Thank you, Tim.
And if you found thisconversation to be helpful,
please share with a colleague ora client.
Start a conversation today, andwe'll see you all next time.
Ad fraud isn't so much a tech problem as it is
fundamentally an incentiveproblem.
You see, middlemen make moremoney when more volume flows
through their systems.
So they have little financialmotivation to cut off fraudulent
traffic, and fraudsters aresophisticated, continuously
evolving to bypass detection.
So today, we're learning abouthow to combat ad fraud, how it
(00:24):
actually happens, and simpleframeworks for avoiding it
altogether.
We'll learn about the red flagof low CPMs, why most reporting
models overclaim credit fortheir contribution, and we'll
hear about what happened whenUber turned off $100 million in
ad spend.
Welcome back to this day ofstreaming podcast.
I'm your host, Tim Rowe, andtoday I'm joined by Dr.
(00:45):
Augustine Foo, creator of FooAnalytics, an analytics pixel
for monitoring programmaticmedia investment for the purpose
really of identifyingfraudulent traffic inventory and
optimizing investment.
So if you want to understand adfraud once and for all, how it
shows up in streaming TV, andmost importantly, how to avoid
(01:06):
it, then this is theconversation for you.
Let's get into it.
Dr.
Fu, why is ad fraud such a bigproblem?
Dr. Fou (01:21):
It's gotten bigger and bigger over the years.
And the way I like to describeit is that ad fraud is not a
tech problem.
It's actually an incentivesproblem.
So it's uh it's kind of easy tounderstand if if you think
about the middlemen, uh, youknow, like the ad exchanges,
even the agencies or other adtech companies, uh, when there's
more volume that flows throughtheir pipes, they make more
(01:44):
money.
So they're not in a hurry tocut down the fraud because
that's gonna cut down theirrevenue.
So in that case, that's that'swhy I explained it as more of an
incentives problem.
And because it's an incentivesproblem, throwing more tech at
it is not gonna solve it, right?
So the bots and the fraudstersare very clever.
They know how to get around thedetection and avoid getting
(02:06):
caught, right?
Bad guys have a lot of practicedoing that.
So that's why fraud has gone upuh over the years and not down.
Tim Rowe (02:13):
That is that is a great point.
They've gotten better at it,and there is a strong incentive
related to uh to figuring outhow to be good at being good at
bad stuff.
Being bad.
Yeah, getting good at beingbad.
You've developed a series oftechnologies and frameworks and
methodologies for thinkingthrough and identifying and
(02:33):
ultimately resolving theseissues.
We'll talk about that, but I'dlove to talk about some of the
really great examples that youshared through some of your
published work.
There's a piece that youshared, 100% click-throughs on
pharmaceutical ads.
Can you tell us about thisstory?
How did how did someone get ahundred percent click-through
rate?
Yeah.
And what does that mean?
Dr. Fou (02:54):
Um that's how I actually got into the fraud
detection side of things becauseI was uh working at an agency
in Omnicom, and we were servinga lot of pharmaceutical clients
and med device clients.
So they were primarily doingsearch campaigns.
And then when you actuallylooked at some of the data, it's
like, what the heck is going onhere?
If you just take the rolled upaverage, the campaigns were
(03:16):
performing like you know, 9%,15%, 25% click-through rates on
average.
But then when you actually pullthe place report by domain,
you'll say, this domain has 100%click-through rate.
This other domain has 100%click-through rate, sometimes
greater than 100%, 100%click-through rate.
So, you know, if you're justlooking at the rolled-up
averages, you're gonna miss it,right?
(03:37):
It's gonna seem like, oh, it'sgreat, you know, this campaign
is performing so well.
But then when you look at theline items, the details, you'll
see stuff that just doesn't makeany sense.
So back then, this was you knowalmost 15 years ago, you know,
no one could really explain whatit was.
They were all just cheering, ohyeah, the campaign's working
really well.
It's like that doesn't make anysense because humans don't
(03:57):
click on ads that much,including search ads, right?
1% would be a good rule ofthumb, maybe 2%, but definitely
not 100 or 200%.
So, long story short, we nowknow it's the bots that are
clicking on the ads, and they dothat because it's a cost per
click model.
So they have to click it inorder to get the revenue.
(04:19):
And so for a lot of theadvertisers, you know, when when
they run a search campaign, alot of the ads do run on
Google.com, right, where humansGoogle things.
But there's also somethingcalled the search partner
network, where there's all theseother sites that run code on
the page to run search ads.
So when there's a search adthat shows up and someone clicks
(04:42):
on it, the site gets some ofthe revenue and Google gets some
of the revenue, right?
It's kind of shared.
So then you can see how thesites have the motive, right,
the incentive to commit clickfraud using bots because they
can just click on ads all daylong.
So then they can make moremoney for their site.
So that's kind of what we'refighting, right?
And when we looked at the dataand you know, it didn't quite
(05:04):
make sense.
That's when I started digginginto it and I built some
technology tools to help me dothe audits, right?
To kind of say what the heck isgoing on here.
Long story short, is uh it'skind of evolved into a platform,
right?
I didn't set out to build fluanalytics, uh, it was more of
some tools to help me auditcampaigns.
And I was originally deliveringit through PowerPoints, right?
(05:25):
Kind of like McKinseyconsulting projects, but that
doesn't scale.
So by 2020, I just decided I'mgonna open up the platform and
let others use it.
So just like people learned howto use Google Analytics for
their websites, they can nowlearn how to use Foo Analytics
for their digital ads.
So that's uh I just gave it aname, Foo Analytics, and then
(05:46):
added user accounts.
And now majority of ourcustomers are the big
advertisers because theirmarketing folks and their
analytics folks are now equippedwith an analytics platform that
they can log into, look at thedashboard, and then figure out
where the bad stuff is, right?
Uh which are the bad sites,which are bad uh mobile apps,
(06:08):
and then just add those to ablock list and clean up their
campaigns.
Tim Rowe (06:11):
Huge.
Uh how does that change or howdoes that look different for
television, for CTV, streaming,for video?
How they're not not clickablenecessarily.
How does fraud show up insideof video?
Dr. Fou (06:26):
Well, it's you know, the click is like the outcome of
it, right?
So you show the ad and thensome portion of the users might
click on it.
But what's more important isthat a portion of the ads are
not even shown to humans, right?
And those a portion of thoseads are going to fake sites.
So that's what we're trying tostop, right?
Even before the click happens.
So if your ad is running on afake site, humans are definitely
(06:47):
not seeing it, right?
And so that's that's theproblem where we're measuring
the ads where they went andtrying to block those before the
ads go there and before themoney goes there.
Because once the money's there,it's really hard to get back,
right?
So a lot of people say, ohyeah, we can get refunds.
Once the money is in the badguys' pockets, you're never
getting it back from them.
(07:08):
So ultimately, it's going to besome middleman like maybe the
exchange that you bought fromthat's going to basically eat
the cost and give it back toyou.
So it's always better to notlet the money go to the bad guys
in the first place rather thantry to get a refund afterwards.
Tim Rowe (07:24):
So step one, prevent the money from going there in
the first place by havinganalytics.
Dr. Fou (07:30):
Yeah, having good analytics.
Yeah, because first you have tosee that you have a problem.
And the reason I'm using theword analytics as opposed to
fraud detection is that I thinka lot of people are familiar
with the legacy fraudverification companies, which I
won't name.
Basically, they just give you aspreadsheet and a dashboard
that says 1% IVT or invalidtraffic.
(07:50):
When you see that and nothingelse, there's really nothing you
can do with that number.
Right.
So the reason I'm calling myplatform analytics is that it
has the supporting details sothat you can go in there and
understand why something wasmarked a bot or why something
was marked, you know, there'sother forms of fraud other than
bots.
So if you say, okay, Iunderstand that, and these are
(08:11):
bad guys, then let me add themto a block list.
Right.
So these are kind of a processthat you go through maybe once a
month, uh, maybe once a week ifif you want to.
You don't have to do itcontinuously, but it gives you a
way to monitor where your adsare going and then continuously
kind of improve and optimizeyour campaigns as you go along.
Tim Rowe (08:31):
Cool.
So we're building a block listas we go along of our of our
known bad actors.
How is inside of CTV, how howare, I guess, how is the fraud
taking place?
Thinking about, let's thinkabout like our local seller
who's trying to explain theadvantages of CTV to uh a small
(08:53):
business owner, but they'relike, hey, I saw an article that
said fraud's gonna take mymoney, so I'm gonna put it on a
billboard instead.
How would you explain that tosmall business owners?
Dr. Fou (09:03):
So in CTV, um, there's actually a conundrum here.
So let me let me kind of putthis out there and let's see how
you how you react to it.
So in CTV, on the one hand, Ican say there's no fraud in CTV
because it's impossible.
And then on the other hand, Ican say there's 100% fraud in
CTV.
Now, how can both of thoseexist and be true at the same
(09:28):
time?
Tim Rowe (09:29):
I don't know, you got me stumped there.
How could that be true?
Dr. Fou (09:33):
So basically, if you're buying from a real seller like
Home and Garden TV, FoodNetwork, ESPN, right?
Obviously, these are onespeople have heard of and
actually subscribe to.
Disney Plus, whatever.
There's no fraud in CTV becauseit's impossible.
Because Disney Plus is nottrying to rip you off.
If you buy ads direct fromthem, the bad guys can't get the
(09:57):
fake impressions in there andcan't get money out.
Therefore, they're not doingfraud in CTV.
However, if you try to buy CTVon programmatic exchanges, and
especially if you're trying tobuy super large quantities at
super low prices, it is not CTV.
It is 100% fraud.
(10:17):
That's because any bad guy canjust upload a bid request and
say it's Disney Plus, say it'sHGTV, say it's a food network,
say whatever they want becausethey want to pretend to be the
legitimate apps so that they canget bids.
Right?
If they put a no-name app inthere, nobody's gonna bid on it.
So it's always better for thefraudster to pretend to be a
(10:39):
well-recognized streaming applike ESPN or Disney Plus in the
bid requests, but they stillhave their own seller ID in
there because they want to getpaid.
So that money is actuallydiverted away from ESPN, Disney
Plus, uh, you know, AmazonPrime, whatever.
Those sellers, the thelegitimate sellers don't even
see that.
And there's a bunch of longtail Roku apps, but that's
(11:04):
actually tiny compared to themainstream fraud.
Okay, so in that case, there'sno fraud in CTV because it's
impossible.
Applies to the case whereyou're buying from legitimate
sellers, ones that you've heardof before.
And let me note that legitimatesellers, including YouTube, has
unsold inventory.
How do I know?
When I'm watching Sundayfootball, you'll see a little
(11:26):
screen that says enjoy the Zen,we'll be right back.
That's an unsold CTV ad slot.
Okay, so even the big guys haveunsold.
That's because the advertisersare chasing low-cost CTV
inventory that's actually notreal.
Okay, they're buying billionsof impressions, and it's it
(11:47):
didn't run on ESPN, it didn'trun on Disney Plus, it ran
somewhere else or didn't run atall.
Okay, and then the the flipside is there is fraud in CTV
and it's 100%, and that'sbecause it's everything outside.
Tim Rowe (12:01):
It seems then like there's a larger conversation
about what a CPM actually is,because if you're paying a
really low CPM, but you're notactually getting the thing that
you're paying for, what the heckwas it worth?
Can we talk about CPMs?
Dr. Fou (12:15):
Yeah, that's actually a very important topic because uh
what I hear all day long iswhen you know agencies and even
the advertisers say, oh, costefficiency, they're trying to
save cost.
Well, let me use a very simplemathematical calculation here,
just based on what I've seenevolve over the last 15 years.
In the early days of digital,you would be paying $35 CPMs to
(12:39):
a publisher like New York Times,you know, any any major
publisher.
So that's $35 per thousandimpressions.
Now you're paying $3 CPMs.
Okay, so people think, oh yeah,we had cost savings.
But actually that's not truebecause CPM is a price.
You still have to multiply bythe quantity.
So now you're buying $3 CPMprices, you're buying 10 times
(13:04):
the quantity, you're stillspending $30.
You see what I'm saying?
That's the cost.
So CPM is a price, it's not acost.
So there's this misuse of theterm cost efficiency.
Oh, yeah, lower CPMs, it's costefficient.
It's not.
So what ends up happening isnow, you know, when the
advertisers are chasing thereally low price like CPMs for
(13:26):
CTV, they end up buying thecrap, like the non-real stuff,
and they're buying lots and lotsof quantity of it.
So the simple advice is don'tbe afraid of paying higher CPM
prices because you're going tobe buying less quantity.
So you're still saving costsbecause fewer of your ads are
(13:46):
going to completely fake sitesand uh mobile apps.
Right.
So that's kind of how to thinkabout a CPM as a price and how
that impacts your cost, right?
So you can really save costs bybuying less of the fraud that's
out there.
Tim Rowe (14:00):
And that segues perfectly into reporting.
If I'm buying a bunch offraudulent inventory, how is it
that my reporting still reflectsthat it's working?
It feels like, okay, cool.
If I understand how fraud worksand I understand that I need to
take measures to prevent fraudfrom happening, and here are
(14:22):
corrective actions I can takeafter it happens to prevent it
from happening again, and Iunderstand the pricing.
How is it actually showing upin my reporting like it's
working?
Dr. Fou (14:32):
Yep.
So it's easily explained if youcan understand the difference
between correlation andcausation.
Tim Rowe (14:40):
Okay.
Let's let's explain.
Dr. Fou (14:42):
Correlation.
Yeah.
So correlation just means salesare happening over here while
you're doing digital marketingover there.
Yes.
They're unrelated to eachother.
So here's the thing (14:52):
like for a lot of the CPG companies, if
you're selling uh soda or ifyou're selling soup or paper
towels or whatever, that happensin the grocery store.
You're doing digital marketingover here.
It's not directly causing salesof paper towels, toilet paper,
soup and soda in grocery stores.
Okay, those are happening,whether or not you're doing
(15:16):
digital marketing.
And there's been someexperiments over the years, not
many, but remember when PGturned off 200 million of their
digital spend?
No change in their sales.
Okay, so those are data pointsthat I've seen that most people
don't want to hear about becausethey all want to think that the
digital marketing is working.
So now let me get to a littlebit more detail to why the
(15:39):
reports look so good.
Okay, so that's a matter ofattribution.
Some of it's fraudulent, someof it's just incorrect
attribution.
So in years past, there, youknow, when some of these ad tech
companies were trying to pushdisplay ads, they were saying,
oh, you know, it's so unfairthat search ads got all the
(16:00):
credit for the sales.
The reason search ads got allthe credit for sales is because
it's usually the last clickbefore the customer purchases
something.
So if you see something on TV,you're gonna say, oh, let me go
Google that, and then you see asearch ad and then you click on
it and then you buy something,right?
Because at that time you'rethinking about it, you're ready
to buy something.
(16:21):
So then that sale getsattributed to the last click,
which was driven by search.
So then all the people who aretrying to sell display ads say,
oh, those that's so unfair.
Display ads must have someimpact and must have helped to
cause that sale.
That is true, but it's not aone-to-one correlation.
It's not like one display ad,you know, a human sees one
(16:42):
display ad and then they go buysomething, right?
You have to see a lot ofadvertising, and then when
you're in, you know, when you'rein the right mood and the right
time, then you're gonna go buysomething.
But because of the display adsellers complaining, Google
basically developed somethingcalled view through conversions.
So that meant even if youdidn't click on the display ad,
(17:05):
it has some kind of benefit,right?
So they're gonna say when aperson is exposed to that
display ad, if there's a sale orpurchase that happens within
the next 30, 60, or 90 days,we're gonna attribute that sale
to the display ad.
That's a very, very rudimentaryway of doing attribution.
And it does work in certaincases, but it's also been now
(17:28):
abused.
All right.
So, for example, you can say,oh, you know, that individual
display ad didn't drive thatsale that happened 30 days
later.
It just happened 30 days later.
And you can see a scenariowhere the person was going to
buy the thing anyway, and theyjust happened to be exposed to
the ad.
It wasn't caused by the ad.
You see the difference?
(17:49):
So the problem is theseattribution or conversion models
over-attribute or overclaimcredit for the sales that would
have happened anyway, or thatwould have happened in the
absence of the advertising.
So, what a lot of advertisersare not yet doing is focusing on
incrementality, which means thesales that would not have
(18:10):
happened in the absence of theadvertising, right?
We actually want to say wespent these dollars in digital.
Did it actually drive moresales than those that would have
happened anyway?
Right?
That's really the key questionthey should be asking.
But kind of to close out yourquestion, it's they're seeing
sales happening, and they haveattribution models that say, oh,
(18:32):
those sales are caused by thead impressions, even though they
weren't.
So that's why a lot ofadvertisers believe that it's
working, even though there's awhole bunch of fraud, like it
could be 90% ads shown to bots,the sales still happened.
But then the model said, Oh,yeah, it's it's you know, those
sales are attributed to thesedigital campaigns that you have
running.
That's why you believe, andthere's a whole bunch of people
(18:55):
who wanted to believe thatdigital works really, really
well.
Okay, so that's how it coveredup the fraud, right?
Even if the campaign had 90%bots in it, impressions were
shown to bots, they'd still hadsales going on, right?
They still had purchases,conversions, all that kind of
stuff.
So they thought it was working.
But careful marketers andprobably more advanced marketers
(19:17):
are now kind of trying to teasethat apart and saying, okay,
well, we can run turn-offexperiments, like just turn off
the digital campaign in onestate and see if the sales in
that state continue at the samerate, right?
Those are simple experimentsthey can run to say, was there a
cause and effect?
And then now to kind of closeout the correlation versus
(19:38):
causation thing.
Correlation is just most of thesales and the digital
marketing, it's justcorrelation, right?
They happen at the same time.
It's not causation.
But I think more and moremarketers are kind of turning on
to that and getting smart andsay, okay, that's not good
enough.
We really have to figure outthe causation.
So there's differentmethodologies, there's you know,
media mix models.
(19:59):
And that kind of stuff.
But you have to be very carefulwhere errors are introduced
into those models and you'restill overclaiming credit for
sales.
So you've got to be careful toonly focus on the
incrementality.
Tim Rowe (20:11):
Someone that's listening today, they're maybe
either buying CTV, they'releading video investment for a
brand, or they're leading alocal sales team and trying to
navigate conversations like thisto drive adoption.
What's the one key takeawaythat you want a listener to
leave with?
Dr. Fou (20:29):
I think the best way to think about this is approach it
as if you were a small businessowner.
Okay, so small business ownershave finite budgets.
If they spend $1,000 and theydon't get any more sales, like
any more people walking intotheir grocery store or anyone,
any more people sitting in thechair at their barbershop, they
can't spend the next $1,000 ondigital ads.
(20:51):
So, you know, the bigmarketers, they have so much
budget.
And they have these media mixmodels and all that kind of
stuff that kind of attribute andtell them things are working,
and you have great ROAS on it,right?
Return on ad spend on it.
Those are the ones wherethey're spending way, way more
money than they should be.
So it's almost like think as ifyou were a small business owner
(21:12):
and you have to actually lookat real outcomes.
And then by the way, when youturn off the spend, the sales
should be going down, right?
Or should go away.
So if it's not, then you cantell the digital marketing or
the CTV ads that I'm buying didnot cause those sales, right?
Did the sales continue when youturn it off?
This was this goes back to, youknow, like an the Uber uh fraud
(21:35):
example from 2018.
They were, they thought theywere very clever, and they said
we would only pay when we getapp installs, right?
That's the ultimate conversion,right?
Get someone to install the Uberapp.
So what do you think the badguys did?
They faked the reporting.
Yeah, exactly.
Well, not even real appinstalls.
They just faked the reportingto make it look like there were
(21:55):
app installs.
There were some of these badguys that were so bold, they
didn't even run ads.
They didn't even have realdevices that installed the app.
So literally, it was all fake.
So what Uber did, and this wasan analytics person there, it
wasn't because of any specialtech or anything.
Kevin Frisch just said, let'slet me just pause the spending
(22:15):
for a week and see what happens.
He paused the spending for aweek, app installs continued.
He paused the app the spend foranother week, app installs
continued.
And he said, Okay, F this, I'mgonna just cut it all off, app
installs continued.
Because humans wanted toinstall the Uber app anyway.
It wasn't because all the spendin those mobile exchanges
(22:40):
caused the app installs.
So, based on that, I mean, mostadvertisers can do that
themselves.
They don't need any specialtech to detect the fraud.
They can just run these turnoffexperiments.
And you don't have to turneverything off, right?
Because you still have to domarketing, but just choose a
market, right?
New York or California orwhatever, turn it off and see if
(23:00):
the velocity of sales andconversions in that market
changed.
If it didn't, then that's adata point for you.
Okay.
Not that many sales andconversions were caused by the
digital marketing.
There may be other things likeactual billboards in the
physical world that you can buybecause people actually walk by
(23:21):
them or drive by them.
Right?
In digital, it's so easy tohide the fraud because
everything's just bits andbytes.
There's so many fake sites, somany fake apps.
No advertiser is going toreview an Excel spreadsheet that
has two million rows in it.
Right?
So they just ignore it and theyjust assume, okay, yeah, they
(23:42):
the vendors tell me it'sworking, so it's working.
So that's where we've gonewrong for many years.
But I think it's time to makedigital marketing better.
Tim Rowe (23:50):
I agree.
And you're working on solvingthat again in 2026.
Foo Analytics has been out fora decade and a half, but you are
launching audits.
Can you tell us about that?
Dr. Fou (24:01):
Yeah, it's it's very simple, but the word audit, I
think, is both scary, but it'salso something that people know
they need to do.
So we've basically put FooAnalytics tags into the ads,
say, for example, CTV ads.
The audit is basically taking alook for the advertiser to see
where the ads are going andwhether the ads even ran.
(24:23):
Okay, so if it didn't run,you're not getting what you paid
for.
And then if those ads went tocrappy websites and mobile apps,
those are not CTV, right?
You paid CTV prices and it wentto crappy mobile apps and
websites.
So those are the kinds ofthings that the audit will help
surface that I don't think theadvertisers currently know
because they're not getting anyof those insights from the
(24:46):
placement reports or even thelog level detail data that they
get.
So it's really taking anindependent look and seeing if
there's areas of improvement.
I'm almost certain that therewill be, especially in CTV,
because the fraud is so rampantand so easy to do.
Tim Rowe (25:02):
Dr.
Fu, this has been a greatinstallment, I think, in
reimagining the way that weapproach digital marketing.
I think that that's the greattakeaway from this is hey,
here's a permission slip to doit different.
The way that it's been donedoesn't necessarily mean that
it's the only way to do it.
And there's clearly a lot ofopportunity here for us to all
(25:23):
get better.
If folks want to learn moreabout foo analytics, learn more
about the audits, connect withyou, read some of the great
content, where should they go?
Dr. Fou (25:31):
Fooanalytics.com, so F-O-U analytics with an S.com,
or just Google my name,Augustine Foo, and I'm all over
the LinkedIn, and I havearticles uh with charts and
stuff like that, so you can seethe data.
Tim Rowe (25:44):
Definitely not hard to find.
We'll make sure that it's alleasy to find close by to this
episode.
Dr.
Foo, thank you.
Thank you, Tim.
And if you found thisconversation to be helpful,
please share with a colleague ora client.
Start a conversation today, andwe'll see you all next time.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
Betrayal Season 5
Saskia Inwood woke up one morning, knowing her life would never be the same. The night before, she learned the unimaginable – that the husband she knew in the light of day was a different person after dark. This season unpacks Saskia’s discovery of her husband’s secret life and her fight to bring him to justice. Along the way, we expose a crime that is just coming to light. This is also a story about the myth of the “perfect victim:” who gets believed, who gets doubted, and why. We follow Saskia as she works to reclaim her body, her voice, and her life. If you would like to reach out to the Betrayal Team, email us at betrayalpod@gmail.com. Follow us on Instagram @betrayalpod and @glasspodcasts. Please join our Substack for additional exclusive content, curated book recommendations, and community discussions. Sign up FREE by clicking this link Beyond Betrayal Substack. Join our community dedicated to truth, resilience, and healing. Your voice matters! Be a part of our Betrayal journey on Substack.