The Myth of Microtargeting

Publishers moving to a contextual-only ad model are experiencing better sales and revenue. Will eliminating third-party cookies make the internet experience better?

Dr. Augustine Fou returns to the podcast to explore recent events surrounding third-party cookies.  These cookies enable the ad-tech industry to programmatically pinpoint an audience by collecting massive amounts of customer data.

But how effective is microtargeting based on this data?

Early this year, Dutch publisher Nederlandse Publieke Omroep stopped using third-party cookies and went to contextual-only advertising on their network of websites.  In doing so, they saw a 62% increase in ad revenue compared to the prior year! Even with their network not dominating their industries, the switch to contextual has created more revenue, more sales, and better clickthrough rates on the ads.

Is the narrative of highly targeted advertising a bunch of overpromised wishes? It seems so.  Eliminating the “mystery meat” in the middle and working closely with advertisers, publishers (and their advertisers) have seen much better results. Dr. Fou walks us through this trend and how returning to this model can save the ad tech industry.


Dr. Augustine Fou: [00:00:00] Bots can deliberately visit a bunch of sites like medical journals, healthcare-related sites and whatever pharmaceutical company sites and in, so doing make themselves appear to be doctors because we have this thing called behavioral targeting in ad tech, right? So most of the users are not logged into the sites that they’re visiting.

Bumper Intro-Outro: [00:00:27] Welcome to endless coffee cup, a regular discussion of marketing news, culture and media for our complex digital lifestyle. Join Matt Bailey. As he engages in conversation to find insights beyond the latest headlines and deeper understanding for those involved in marketing. Grab a cup of coffee, have a seat. And thanks for joining.

Matt Bailey: [00:00:51] Well, hello everyone. And welcome again to another edition of the endless coffee cup podcast. And I’ll tell you what we’ve got a returning guests today, and I’m really excited about that. Dr. Augustine Fou was with us before to talk about ad fraud, and I’ve asked him to come back again today to go a little more in depth on that. Dr. Fou, how you doing today?

Dr. Augustine Fou: [00:01:14] Hey, very good. Thanks for that. Thanks for having me again on the show.

Matt Bailey: [00:01:17] Oh, absolutely. I’ve got to tell you the last time you were on the show that very quickly turned into one of the more highly downloaded episodes. And I think you really struck a nerve with a lot of people, both in marketing and in the ad tech world, with what you are doing in bringing to light the problems with programmatic and with this bot traffic taking over.

Dr. Augustine Fou: [00:01:44] That’s great to hear. Well, we still see some of that going on, but,  I think as marketers just kind of look under the hood a little bit more, or maybe under the rug, if you will, a little bit more, and they’re able to see where the fraud is coming from. They can actually make their campaigns a lot better just by turning off some of those domains and apps, the mobile apps that are designed solely to rip off marketers, meaning just eat and lecture, ad budget as possible because that’s going into their pockets and not to showing ads.

Matt Bailey: [00:02:16] Absolutely. And it’s not without resistance from what I’ve seen.  Not everyone is excited about your message.

Dr. Augustine Fou: [00:02:23] Yeah. It’s kind of tough to hear because a lot of marketers, you know, when things were going smoothly and their agencies were cheering, you know, there’s all this stuff they could buy. You know, like I think a lot of marketers have approached digital, especially the display ads and video ads and programmatic, they’ve approached it with a TV buying mentality where a greater reach and frequency is what they wanted .

So in digital, greater reach and frequency means buying more ad impressions, right? Oh, well here’s 10 billion more you can buy over here and programmatic channels, regardless of how many humans there are on earth. Right? But nonetheless, there’s all these billions they can buy over here and they happened to be way cheaper too.

So, isn’t that awesome? So they ended up just pouring a lot of money into it because of the volumes, but unfortunately, the hard part of the message is that some of those ads were not shown to humans.  They’re shown to bots and it’s really hard or embarrassing to go back and admit that.

So in some cases I’ve seen the marketers just use what I would call rubber stamp reports from fraud detection companies that just say, “Oh yeah, we have fraud detection,” and, “Oh yeah, it says it’s 1%.” So they’re very happy seeing that the fraud is 1% and they would prefer not to look more closely to realize it’s more, right?

But I think some of the marketers, especially in this downturn, right? Cause we talked about a year ago and we didn’t face the pandemic as we are right now. So I think I’ve seen a huge pickup in activity where more marketers are really looking more carefully at their spending, right? And especially if they’re paying for fraud detection vendors that keep telling them fraud is 1% year over year, they’re starting to question that, right? Whether they’re actually measuring everything and catching all the fraud, right? And so we know that budgets are going to be tight, and remain tight into next year. So the marketers are saying, okay, we gotta, you know, before her bosses or CFOs come in and scrutinize us more, we better do this ourselves.

So they’re looking more closely, they’re doing more audits, they’re finding stuff that’s, you know, not a worthwhile digital expense. So if the fraud detection is not working for them, they’re going to cut it and start to look at the analytics themselves, so that they can actually do a better job cleaning their campaigns, themselves.

Matt Bailey: [00:04:43] Right. That’s great news that some of the upside of the pandemic has been that it’s forcing a closer eye, a closer measurement on this type of spending.

Dr. Augustine Fou: [00:04:53] Yeah, because before, you know, like when things were running smoothly, no one wants to rock the boat, right? No, one’s going to say, “Okay, well maybe we should look closer, even though everything seems fine. We should still look closer.” No, one’s going to do that, right? Everything’s going smoothly. Don’t rock the boat.  But now because the virus has already rocked the boat, they’re more willing to say, “Okay, well, we should probably take this opportunity to look closer.” And, you know, I’m hopeful that going into next year more and more marketers will do these audits and look at what’s working and what’s not, so they can figure out what to cut.

Matt Bailey: [00:05:26] Well, and it’s more of the audits. So what prompted me to contact you again about doing this is, again, because the pandemic has rocked the boat.  There is a story of a Dutch publishing group, NPO, and around January, they decided to remove all third party tracking on all of their websites.  They went to basically just beefing up their contextual advertising. And what is so surprising about this, is their revenues jumped dramatically. That they’ve had better revenue with contextual based advertising than with third party tracking and going through the exchange. To me, this challenges this narrative that has been so strong over the past, you know, I think it’s just been building over the past 10 years, of more data equals better targeting, which equals more revenue or more success. And this flies in the face of that.

Dr. Augustine Fou: [00:06:33] Yes, very clearly, and that’s one of several examples. But it boils down to something pretty basic. The people who are getting more revenues are not the publishers or the advertisers. It’s the middlemen selling you the tech to do it, right? So that should have been obvious, you know, early on, but it should be obvious.

Now in hindsight, of course, they’re going to sell you that because they’re going to make more money off of you. So it, you know, it’s very simple. The more targeting parameters you buy, the more the middlemen make, right? So, you know, it costs more to have five targeting parameters. It costs more to have 10 targeting parameters, so on and so forth.

So they’re obviously trying to sell you more is better. Now for the publishers, what that’s caused them to do over the years, is that they’re now adding more and more third-party trackers and stuff to the page. And that’s what’s created the current privacy problem, right? Because when a human goes to New York Times, they know that they’re interacting with that publisher, New York Times. What they don’t know, is the 80 other trackers that are loaded on the page from a third-party ad tech companies that they’ve never heard of.

Right? So there’s a privacy kind of angle woven into this as well. But you’re right. It’s like the ad tech companies have been convincing marketers that, “Oh, if you get, just get more data, that’s going to mean more targeting and that’s going to be more relevant ads.” Now to a certain extent, that is true.

And to put it bluntly, it’s like if you have two or three targeting parameters, your ads are going to be more relevant, right? So just take a simple example. If you’re marketing a beard trimming product, right, you really only need to market that to males, right? You don’t need to market that to women.

Matt Bailey: [00:08:20] Right.

Dr. Augustine Fou: [00:08:21] So if you just took one targeting parameter, men versus female, you’re going to make that more relevant, right? And then you take another parameter, like maybe age range or something. That’s going to make it a little bit more relevant. But when you start getting beyond 3, or 5, or 10 targeting parameters, the cost of diminishing returns kicks in, right?

So yeah, you’re not going to get much more relevant ads when you’re using 15, or 30, or 50 targeting parameters. You’re just paying more for that.  So I think some marketers, you know, got caught up in more is better and more targeting, must be more relevant ads. And then, you know, that’s led to a whole bunch of other, you know, bad stuff like, uh, bots could behave in a certain way to make themselves appear to be a particular target audience.

Right. So I think I’ve used this example before bots can deliberately visit a bunch of sites like medical journals, healthcare related sites, and whatever pharmaceutical company sites and in, so doing make themselves appear to be doctors because we have this thing called behavioral targeting in ad tech, right?

So most of the user are not logged into the sites that they’re visiting, but when they visit the page, a platform like DoubleClick can set a cookie. Right. So if this user that’s anonymous and not logged in visits, a bunch of medical journals sites and healthcare related sites, they’ll start to appear to be like a doctor because they behaved like one.

[00:10:00] And so when we have this behavioral targeting, the bots can do the exact things that they need to become that audience, because they know if they’re a doctor, they’re going to earn a higher CPM from the pharmaceutical advertisers who are trying to target them. Right. So in those cases, you know, they pretend to be that by visiting those sites, they then go to cash out sites where if they cause an ad to load on those third party sites, those sites make the money.

And because it’s, uh, the pharmaceutical companies think that they’re retargeting a physician they’re paying higher CPMs. So that’d be for every ad loaded on those cash sites, they’re making more money for that site. So in that sense, there’s not only the bad data problem, because these are not humans.

These are not doctors. These are bots pretending to be doctors and you have the problem of the pharma companies paying even higher CPMs and those CPMs going to these fake cash out sites. Right. So do you see how all of these different problems are interwoven? Right on the one hand they’re not even targeting humans.

Right. And on the other hand, on those sites of the humans do visit, uh, their privacy is being violated because there’s all these trackers they don’t know about.

Matt Bailey: [00:11:03] Oh, absolutely.

Dr. Augustine Fou: [00:11:04] It’s just become this big mess.

Matt Bailey: [00:11:07] Yeah. And it seems you have spawned so many different levels of either number one, acquiring more user data. You know, you don’t have to go far to see how apps are gathering much more data on people than what they say. Someone posted the other day about buying a new pair of Bluetooth headphones for fitness. And they’re asking them their medical history as they set up these Bluetooth headphones. And they said it just got to the point where I had to stop because I’m disclosing way too much info. These are headphones. So it’s like every connected device is wanting more data. Every app we use is using more data and, and, and frankly, it was interesting, especially on the dating apps. How much data they are pooling on people retaining and then using that to sell into the ad market.

Dr. Augustine Fou: [00:12:07] Yep. All of those. Right. And, and, and for most consumers, they’re starting to be sensitized to that. Right. So they’re starting to pay attention. Well, why is this ear buds asking me for my health information? Right. And so, you know, prior to this, they were not as sensitive, but I think in 2016, because of the Facebook Cambridge Analytica scandal, you know, more consumers are finally aware of these potential problems.

Right. So back then, it was just that people assume they were interacting with Facebook. They didn’t know that Facebook allowed third parties to come in and harvest their data, or, you know, Facebook was paid by third parties to get access to the data. Right. So consumers literally didn’t know that. And that came as a surprise. But because of that, it’s actually precipitated, you know, more privacy regulations and also more consumer awareness of these issues. But you’ve heard the term surveillance marketing, right, or surveillance economy. All that’s based on this kind of data collection. Right. I call it wanting a data collection because the marketers just think more is better.

And unfortunately, you know, like we said before, you know, if they didn’t even take into account the, the bot activity more is not better. Right? And a lot of times now with the new privacy regulations, if they didn’t get consent of the consumer ahead of time, it’s now starting to be a compliance risk for the marketers and the publishers, because some ad tech company collected the data without permission.

So then that becomes a compliance risk for the marketer that uses that data for ad targeting and the publisher on whose site, uh, the data collection was done with or without the consumer’s knowledge. So I think it really gets back to your NPO example, right? That a Dutch publisher, where, uh, when they ripped out the ad tech, they ended up making more money.

So I’ve said this before, you know, for marketers and publishers, um, they’re going to actually end up making more money or saving more money by using less ad tech. And, you know, like we said earlier, you know, if you’re paying the middleman, right, uh, then they’re making the money and your dollar is not even going towards showing the ads and the middlemen have every incentive to convince you to buy more from them, because that just means more of the dollar going into their pockets, as opposed to you doing better marketing.

So I think marketers just need to be more aware and more sensitive to that trade off that, you know, pass a certain point. Um, just buying more data parameters for targeting is not going to yield much more incremental benefit.

Matt Bailey: [00:14:51] That’s the question. I mean, as I was preparing for this and reading through some of the articles, the big question that popped in my head is have we been sold you know, so to speak a bill of goods when it comes to micro-targeting?

Dr. Augustine Fou: [00:15:04] Yeah.

Matt Bailey: [00:15:04] I, it just seems like, like I said, the narrative over the years has been so focused that, and also there’s another, I would say another false narrative that people are willing to trade ads for free content. Uh, you know, I don’t remember, as I told someone, I don’t remember signing that agreement. Uh, but that goes right along with it.

Dr. Augustine Fou: [00:15:24] Yeah. You, you get at kind of the fundamental, uh, kind of the origins of the original internet, which is, it was a, a balance between three parties, the advertiser, the publisher, and the consumer, right. The consumer was going to get free content from the publisher because the publisher was going to rely on ad revenue.

So we call that the ad supported internet and things like that. But in the original form, it was a three legged stool that was balanced right between the publisher advertiser and the consumer. But when the fourth leg of the stool came in, it threw everything out of whack. And the fourth leg is bad tech.

Matt Bailey: [00:16:03] Wow.

Dr. Augustine Fou: [00:16:03] Right. And I said bad tech specifically, because you know, this fourth leg is not there to help any of the other three legs. In fact, it’s trying to extract as much revenue as possible for itself because they’re, um, you know, trying to make their VCs happy. Right. These are all the ad tech companies funded by Silicon Valley VCs, and they’re on their way to go in public or to private equity or whatever.

So, yeah. Their first priority is to make as much money as possible. So therefore you have these narratives like, Oh yeah, hyper targeting or micro targeting is better than whatever, whatever. So by selling that bill of goods, they’re convincing marketers to buy more of this shiny object or snake oil or whatever you want to call it.

Right. And for a time it was better. You know, they, they, they have some case studies, but again, if you look at some of those case studies, they were funded by Google.

Matt Bailey: [00:16:57] Yeah.

Dr. Augustine Fou: [00:16:57] Right. Or they were funded by the ad tech companies.

Matt Bailey: [00:16:59] Right.

Dr. Augustine Fou: [00:16:59] So what do you think they’re going to tell you, they’re going to tell you more targeting parameters is better, but what they don’t tell you is that beyond three to five targeting parameters, it’s point of diminishing returns, not even, uh, you know, flattening out.

So the marketers have to do that themselves. And too often, you know, kind of going back to the beginning of this podcast, the marketers are still approaching it with a kind of a reach and frequency mentality. So on the one hand, they’re buying mass quantities. On the other hand, they’re, they’re thinking that they’re buying all these targeting parameters for these mass quantities, but there’s also a thought exercise I bring clients through to say, okay, well, let’s think about this for a second. If you start off with a pool of targetable, Humans, they say it’s a hundred percent of the, the addressable audience. Okay. And you just choose you take one targeting parameter, gender, male, or female, and you just choose male and not female.

You’ve kind of cut that pool by half. Right? So you went from a hundred percent addressable to 50% of the original audience. And if you take one more targeting parameter like age range and assume that there’s five, just very simple calculation, right? You take the 50%, you break it into five and you choose one age range.

Now you’re down to 10% of the original targetable audience. And if you take just one more, say geographic region, and again, let’s just break the country into five parts and you just pick one. I only want to advertise on the West coast, for example. Um, now you’re down to 2%. Of the original a hundred percent addressable audience.

So with three targeting parameters, your addressable audience that fit those criteria is already 2% of the original, right, way, way smaller. What happens if you have 10 parameters? What happens if you have 50 parameters, right? Let alone 300 or 500 parameters that these ad tech companies are trying to sell you.

So in those cases, the addressable audience gets so small that it’s not relevant anymore. So then the ad tech companies went on and conveniently made up the concept of lookalike audiences. Right to go, Oh, well, you know, if you don’t have enough over here, let me go find you some other users that look alike because they happen to visit a similar set of websites.

And that’s where the bot problem comes in. Right? So that’s where the bots will deliberately visit medical sites to appear to be doctors, or they will deliberately look at backpacks during August and September timeframe to appear to be back to school intenders. Right or they’ll look at certain, um, electronics items or whatever. You remember the story about, you know, the, uh, auto intenders there’s 300 million auto intenders in the U S there’s only about 350 million people.

[00:20:00] In the U S including babies and grandparents. Right? So in some of those cases, uh, if you just thought about it, took a step back and thought about it for a second, you’ll realize the numbers are just don’t make any sense. So I think over time, you know, more and more marketers should look more carefully at, you know, these myths or these, uh, bill of goods that have been sold by the ad tech companies.

Matt Bailey: [00:20:03] Absolutely. It’s, it’s absolutely mind boggling. Yes. When you see some of the numbers and it, you know, I think it’s being sold as well. It’s repeat impressions. It’s it’s, you know, all types of things, but like you said, it’s just that moment of wait a minute, let’s, let’s look at that number and evaluate it. And unfortunately it seems like just that, that split second critical evaluation just goes out the window because people love big numbers. And yeah, if I could put a big number in front of my boss, he’s happy, you know, it’s and it’s funny because this goes back to, you know, when I first started, I was building websites back in the nineties and I used an early version of web trends to really find out what’s happening here. And it taught me very quickly that big numbers lie.

Dr. Augustine Fou: [00:20:53] Real humans behave in a very like it’s small numbers. And if you’re uncomfortable with small numbers, you really, it’s hard for you to do digital marketing.

Matt Bailey: [00:21:01] Absolutely.

Dr. Augustine Fou: [00:21:03] You know, think about, you know, I usually ask this of my students, right? When was the last time you deliberately clicked a banner ad? And it’s like, uh, I can’t even remember the last time I deliberately clicked a banner ad. Right. Let alone which ad it was. Right. So do you actually remember what ad it was?

So humans are very good at avoiding ads and some of them even use ad blockers, so they don’t even see the ad to begin with. Right. So there are those kinds of issues where branding type ads, display ads, and video ads have limited effect if obviously it’s shown to not a human, you know, it’s not going to have even the branding effect.

And so for a long time, I do talk about search ads, where that is the moment when someone types in a search, that means they’re looking for something and they even told you what they’re looking for. So wouldn’t that be the best time to put an ad in front of them? Right. So a lot of the brand marketers say, Oh no, no, no.

We have to, you know, kind of, uh, get our brand in front of them. Otherwise they won’t even know what to look for you and all that kind of stuff. Right? So there’s always these different schools of thought between branding versus kind of harvesting that demand when someone comes online to search for something.

Matt Bailey: [00:22:11] The numbers are very clear that when you do that type of response advertising, or, you know, I, it’s almost like a digital direct response that if I type in, you know, running shoes into Google, I get ads for running shoes. It’s based on what I need, but then also three targeting ad.

Well it’s because I was at the site, I did the shopping and because of my tracked behavior from the ad to the site to leaving. And, and, but the response rate of retargeting ads is significantly higher.

Dr. Augustine Fou: [00:22:45] Higher. Yeah. So, okay. There’s some, some good things about that. Right? So in theory, it’s like if you visited the site and then you ended up in, you know, not buying something, it would make sense for you to follow up and show them an ad. And so there is some legitimate, legitimacy to that, but let’s take it a step further and let me weave in a fraud angle that I don’t think most listeners realize.

Um, there’s a concept in remarketing where we call organic stealing. And what I mean by that is. Okay. So let me just define a few things. Retargeting means someone visited a site and they are re-targeted with ads, uh, later, uh, whether for the site or for the product itself. Okay. So sometimes we’ve seen that, like, we looked at a particular dress on Amazon and that exact dress shows up in an ad, you know, right on the next website.

Matt Bailey: [00:23:37] And this is done through first party cookies. So it’s legitimate. I went to that site and they’re advertising to me.

Dr. Augustine Fou: [00:23:45] Yep. Now some of this is still creepy in the sense that they didn’t know that I already bought the dress right. Or bought the shoes or bought the swing set or whatever, whatever. And they still show me the ad. So that starts to get creepy and annoying.

Matt Bailey: [00:24:00] Right.

Dr. Augustine Fou: [00:24:01] So that’s retargeting. Now remarketing is similar in the sense that you’ve bought from a particular merchant before, like I’ve already bought from Macy’s. Right. So what they’re saying is, oh, well maybe if I advertise to this user, they might buy more from me.

So again, there’s some legitimacy to that theory, but if the person already bought from an online merchant, like Macy’s or J crew or whatever, they’re probably going to buy from them again.

Matt Bailey: [00:24:26] Right.

Dr. Augustine Fou: [00:24:26] So the problem comes in with some of these remarketing companies where we call it organic stealing. Right? So they’re taking credit for sales that would have already, that would have happened anyway. Right? The user was going to go back to Macy’s and buy something anyway, it wasn’t because of some magical secret sauce from the remarketing that did anything, the person was going to buy it anyway. And let me use one other kind of related example to illustrate this. This actually happened. So Uber, um, is now suing a hundred mobile exchanges for basically fabricating and falsifying the data to claim credit for app installs.

Matt Bailey: [00:25:07] Wow.

Dr. Augustine Fou: [00:25:07] Because Uber was paying a lot of money to get more installs of the Uber app. So there’s what we call organic installs where the human, uh, was going to install the Uber app anyway, because they wanted to, right. Not because we saw an ad and clicked on it. Right? Whereas these mobile exchanges were falsifying the data and claiming credit by doing what they call, um, click flooding or click injection to make an organic install, appear to be one that was caused by them so that they can actually earn a cost per install fee, right? So they typically are in between $5 and $10 CPI cost per install. So in that case, Uber was paying out these $10 CPIs for app installs that would have happened anyway.

Matt Bailey: [00:25:56] Wow.

Dr. Augustine Fou: [00:25:57] So in a parallel thread, some of these e-commerce merchants are paying out to these remarketing services for sales that would have happened anyway, right? What we call organic sales. So the key there is, you know, there’s, there’s the point that’s the marketer’s missing is incrementality. Right. Did the remarketing drive new sales that wouldn’t have happened anyway, he’s made the difference. It’s kind of subtle. And a lot of people just miss that there’s oh, we got more sales.

Right. And literally they go show their boss bigger numbers. Right? Here’s a, we got more sales. But what they didn’t realize is, uh, they didn’t actually have to pay for that or pay the remarketing company for that, because those sales would have happened anyway, did the green marketing company actually drive incremental sales above what would have happened anyway.

And that’s actually a lot harder question that most of them can’t answer.

Matt Bailey: [00:26:48] Yeah.

Dr. Augustine Fou: [00:26:49] Right. That’s where they’ve been sold this bill of goods. The numbers look awesome. Right? You said retargeting numbers, click through rates and sales, all that. And remarketing numbers look way, way better. Of course, they look way, way better because those sales would have happened anyway.

You’re just claiming credit for it and make it appear that the remarketing tactics actually drove those. And that’s actually not real. So I think marketers need to be much more careful, much more strict. When they, when they go into these channels and do retargeting and remarketing, and really look for incrementality as opposed to just big numbers.

Matt Bailey: [00:27:24] Boy, this is something I see a lot. And I think it’s because of number one, the focus on digital. And a misunderstanding of attribution that because we have digital analytics of Google, of Adobe and Salesforce, and everyone is in this digital analytic mentality, therefore, everything can be traced to an attribution.

And when you think in that context, all of a sudden now what happens is everything outside of the digital context, the magazine ad, the branding, the, you know, the mindset of the consumer, or even I hesitate to say a direct mail piece could drive that sale, but yet, because the sale happened digitally, you’ve got groups claiming attribution credit.

Dr. Augustine Fou: [00:28:20] Yes. Exactly. So it kind of boils down to just having more data doesn’t actually mean it’s better to have more data and sometimes the data is even wrong, right? Not only is it misleading, it’s, it’s outright wrong. So the problem of the digital attribution stuff is that yes, it appears that we have a lot more data digital than, than ever before. Right. It doesn’t mean that the TV ad didn’t work doesn’t mean that the magazine ad didn’t work.

And it also doesn’t mean that the digital thing did work because we have all this attribution.

Matt Bailey: [00:28:52] Right.

Dr. Augustine Fou: [00:28:53] Um, and it’s because you do need to take into account the possibility that the attribution systems are being gamed or you know, where the fraudsters are tricking it.

So, you know, I’ve written about the kind of cybersecurity implications of attribution. And the reason I wrote about that is that some of these attribution URLs are literally in the clear with plain text in it. Right. So you can literally see the item that was bought the user ID, or maybe the email address of the user that bought it and the price they paid.

All of this is in the clear as parameters in the URL.

Matt Bailey: [00:29:30] Oh wow.

Dr. Augustine Fou: [00:29:31] So when you can see all that. You know, a bad guy can just attack that and copy it off and replay it. So in some of the cases, you know, you probably won’t even believe this happens, but it does where the fraudsters are tricking the attribution platforms to make it appear that a sale happened when no sale happened.

[00:30:00] And by the time they settle up and figure out, oh, well actually my e-commerce system didn’t show the sale where the attribution systems show the sale. It’s already a month too late. Right, because if there’s a 30 day offset between when they have to pay out the revenue share for the sales that appeared in the attribution system, and then they don’t know that the sale didn’t occur in their own e-commerce systems until 30 days later they’ve already paid out in the back, has already made off of the money. So there’s this point where, you know, in digital, you really have to be careful because lot of those metrics and the attribution can be tricked.

And did you know that Google analytics can also be tricked? Right. So you hear about these fraudsters or traffic sellers selling you traffic. That’s supposed to come to your website.

Matt Bailey: [00:30:43] Right.

Dr. Augustine Fou: [00:30:43] Some of them don’t even have to bother building bot nets to actually have bots load your web pages because all they have to do is trick your Google analytics to appear to have traffic when no pages were ever loaded because basically there’s a vulnerability or a capability or feature, whatever you want to call it, where you can literally all the Google analytics API pass it a correct UA code, which identifies the Google analytics account. And you can send a fake traffic and it all registers faithfully.

And, you know, so in that case, the bad guys don’t even have to waste time and bandwidth on making real bots. They just trick Google analytics. So if you, if you realize that that can happen and you’re on the lookout for whether that happens to your other stuff, then you’re probably good, but you’re probably okay.

Right. But if you’re not even aware that tricking the attribution is a possibility , and then you just assume all of the attribution data is correct. Right. All the targeting data is correct. Uh, you’re the prime target for fraudsters.

Matt Bailey: [00:31:51] Wow.

Dr. Augustine Fou: [00:31:51] Right. They, they love those marketers. Right. Who don’t even know what to look for.

Matt Bailey: [00:31:56] Right.

Dr. Augustine Fou: [00:31:56] So they can be easily tricked. Oh, I’m getting such great engagement. I literally heard a marketer say I’m seeing, I’m getting such great engagement because they were citing the click-through rates. Bots can click on stuff. So of course, they’re going to show you better click through rates over here, and that’s how they trick you into allocating more money to their fake sites because you’re getting quote unquote, better engagement.

Matt Bailey: [00:32:18] Hmm.

Dr. Augustine Fou: [00:32:18] It’s not that that many more people liked your ads so much more that they clicked on it, right? It’s because of bot activity. So marketers should always have this sensitivity and this awareness that they can be tricked, right? Not that they are, but they just need to be on the lookout for it. Otherwise, they’re an easy mark.

Matt Bailey: [00:32:38] Okay. On that note, let’s take a quick break here because I’ve got to catch my breath after that, knowing there’s a little problem with attribution, but then, uh, you know, blowing a hole in it. I love it. I it’s it’s. This is just, uh, I love talking with you and learning about all these things, uh, listener, go ahead and refill your coffee and I’ll see you on the other side of the break.

Sponser: [00:33:04] If you are a regular listener of endless coffee cup, well, you know, I love coffee. If you are listening for the first time, then the name of the podcast, should be a big clue. I love conversation. And I think that we have a lot to learn from each other. If we only take the time to sit and talk over a cup of coffee, maybe we’ll come to a resolution, maybe not, but at least we’ve learned from each other.

It’s in this spirit that I’m happy to announce a regular sponsor of the endless coffee cup. A coffee sponsor In Care of Coffee is a coffee company that is unique in the market. They are directly tied to the communities that produce coffee and ensure that the proceeds go to the coffee farmers, the local producers, and their communities.

There are no middlemen or markups. The proceeds from the sales go directly to the communities that work so hard to bring us the enjoyment of a great cup of coffee. Go to, follow the link in the notes, and use the code “endless” to get 10% off your order. Again, that’s the discount code “endless” to get 10% off your order.

I’m enjoying the whole bean Weiwei Tenango coffee today. Please support our sponsor and know that your support goes directly to the farmers, producers, and communities that help power us. You’re at the Endless Coffee Cup Podcast.

Matt Bailey: [00:34:35] We’re back with Dr. Fou who has just, uh, hopefully blown your mind about attribution and Google analytics. And, you know, I want to talk a little bit further about, about measurement and metrics because before the break you talked about it, you know, and I hear this a lot. People love engagement metrics and, and, you know, we’ve talked before and I think, uh, you know, Bob Hoffman is, is, you know, someone who talks about this a lot is that when we, when we can’t measure revenue or direct sales, we move the goalposts and engagement has become one of those.

I like to joke with people that years ago, the word engagement meant commitment. Today, it means someone clicked on something and we’re celebrating it and how it’s in very misleading metric and it doesn’t represent behavior.

Dr. Augustine Fou: [00:35:32] Yeah. I mean it could, but you also have to understand the, the potential fraudulent side of it.

I mean, in the early days when the bots don’t have financial motive to, to trick these things then of course that’s representative, right? So if you have more clicks on something compared to something else that has less clicks, okay. You probably got a little bit more engagement with the former, right? So I mean, all of these things start off as plausible, legitimate and whatever, but sooner or later, if the bad guys have a financial motive to do it, they will find a way to trick the system.

So these days when marketers are so focused on the click through rate, what do you think the bots are going to do? They’re going to click on more stuff. Right. So, you know, we talk about these big numbers. I mean, we talk about these engagement metrics. I just call them vanity metrics or quantity metrics. If any of those things, right, whether it’s the number of impressions, the amount of traffic, the number of clicks, the click through rate, all of those quantity metrics can be easily faked by the bots. And so if those were the key metrics, I, they, uh, marketers call them KPIs, key performance indicators or something. The bots will trick the exact thing that you want to look at or that you are looking at.

Right. So that, uh, it makes it look better. So whether it’s the attribution stuff that we talked about before the break, or, you know, any other metric that you are focused on, they will trick the thing that you’re looking at. Right. And, and so it makes it look really, really good. And so you just keep buying, right?

And so I would say not all the middlemen are involved, right? The, the ad tech’s changes, whatever, they’re not necessarily the ones who are managing the botnets themselves, but they’re certainly very happy that someone else is managing the pot and  being very large volumes of ad impressions through their systems.

So their revenues depend on the volumes and the more volume, the more money they make. So in those cases, again, it kind of gets at the, the point about, um, you know, ad fraud is not a, a tech problem. It’s an incentives problem. Right? A lot of the middlemen are kind of standing by, they’re not saying anything because it’s making good money for them.

So they’re not in a hurry to, to squash the fraud.

Matt Bailey: [00:37:47] So I want to run this by because working with a large tech company a few years ago in part of their training, they were telling marketers to, yeah, go ahead and expect fraud, but it’s going to be hard to detect because your CPA is going to be better than most other channels, and so you may look at it as a victory that I’m spending less, I’m getting better leads, but there’s going to be fraud. And, and the, the good metrics will mask, you know, probably a good percentage of fraud. Is that still the case or, you know, they’re tying it to a revenue metric, uh, and just saying, you’re going to miss some of this.

Dr. Augustine Fou: [00:38:28] So there’s a couple of things kind of intertwined in that. So on the one side, yes, they’re tying it to revenue metric. But what they’re missing is that just kind of, like I said before, the organic sales that would have happened anyway, their revenues might’ve happened anyway, when you didn’t have to spend the money in digital to get it, right.

So that’s the part that they’re missing. And simply by looking at revenues and not incrementality, that’s where they’re making that mistake. Okay. The other thing is it’s a misconception where, I mean, I hear it as the fraud is priced in, right? So you kind of say you know to expect fraud, but, uh, I hear it as a, the fraud is priced in. What they mean by that is 10 years ago when they had to buy ads from big publishers, it’s probably say $30 CPMs for round numbers sake. Right? So $30 to buy ads from the New York times, whatever. Now through programmatic channels, they can buy ads for $3, like 1/10 of that. So there’s an, oh, wow, I’m getting so much cost savings. Um, but no, you’re not actually getting cost savings if you’re buying 10 times the quantity.

Matt Bailey: [00:39:37] Right.

Dr. Augustine Fou: [00:39:37] So even at $3 CPMs, because you’re buying 10 times the volume, you’re still ending up spending $30, but the difference is, 10 years ago, when you spent $30, you were buying direct from a good publisher that had real human audiences. Now, when you’re spending $30, you’re paying $3 CPMs to long-tail crappy sites on exchanges where most of their audience is not human it’s bots. Right?

[00:40:00] So in that sense, um, the “fraud is priced in” argument is flawed and it’s kind of an excuse that a lot of marketers still use to keep buying in programmatic.

Matt Bailey: [00:40:14] Right.

Dr. Augustine Fou: [00:40:16] If they just paid higher CPM, maybe not $30, but say they, instead of 3 they, they paid 10 and they bought less quantity. They would still be saving money overall.

So it kind of ties back to your CPA argument, right? Cost per acquisition.

Matt Bailey: [00:40:32] Right.

Dr. Augustine Fou: [00:40:32] Those are artificially low because you’re claiming credit for sales that would have happened anyway. If you actually took into account the incremental sales that you drove from your digital marketing, right, those sales that would not have happened anyway, then the CPA calculation looks a lot different and then you’ll realize, oh, well maybe I didn’t have to spend all this stuff over here in digital. And I’m still going to get that. Right. So there’s a lot of these things where, you know, I think I chalk it up to a shallow understanding of, of these things by the marketers.

Not, not to say, not to fault them. It’s just like, sometimes you got to just think through one or two more layers to realize, okay, well, I may not have had to spend that money in digital to get the outcomes. So I can’t actually count that as part of my, you know, uh, sales metric for, uh, for the digital programs.

Matt Bailey: [00:41:26] You bring up a good point there, and I, and I, yeah, I certainly don’t fault marketers. I mean, many times, you know, I met a well-known brand and I’m teaching marketers and invariably, I end up teaching marketers who are doing the buying and I’m teaching them how the ad tech system works. And invariably, very few, if any, understand it.

Dr. Augustine Fou: [00:41:49] Yep.

Matt Bailey: [00:41:49] Their, their minds are getting blown. And most of them at the end of it are saying, I’m not sure I like this after they start learning about it. But what I’m finding is the ad tech system has hit so fast. And it’s speeding up. Whereas you look at your, your marketers, who many are a marketer for less than five years.

Dr. Augustine Fou: [00:42:12] Yeah.

Matt Bailey: [00:42:13] And especially when they are in role with the buyer or working with an agency, they have even less experience. And so, they’re treading water, faking it till they make it and just signing off, doing things, whereas the speed of the, the tech and the knowledge is running so fast ahead, and they’re just trying to get a basic understanding of what I’m supposed to be doing today. Uh, that’s what I’m seeing on the ground.

Dr. Augustine Fou: [00:42:41] Yeah, I I’m seeing the same thing. They’re ending up just falling back to assumptions. Right? So the assumption that a lot of them make is that an ad is shown to a human when they visit a webpage. So I would say that is true 10 years ago, 20 years ago, right? When digital advertising first started. When a human came to a web page and the page loaded, the ads would load.

So that was what happened back then. But these days you can no longer just assume that, because of the amount of bot activity. And because of the ability of fake sites, um, you know, to be created and used for fraud. So I kind of chalk this up to our transition from the physical world into the digital world, right?

So in the physical world, there’s basic physics where you cannot stick more than so many billboards by the side of the highway, until you, you ran out of room, right? All right. You can’t have a hundred pages of ads in the magazine. You have a finite number of pages you can dedicate to ads. And similarly for every 30 minute TV sitcom, you can only stick so many minutes of ads.

I’ve seen the ad load go up, but nonetheless, it’s finite. It’s it’s limited. But when you move into digital, you remove these constraints of the physical world. So you can now create unlimited websites. And now you can create unlimited ad impressions. And the way you can see this happening at the macro level, macro economic level is, you know, fundamental economics will tell you if there’s higher demand and a constrained supply, prices should go up.

But what have we noticed, or what have we observed for the last 10 years? Prices have gone down. And that is despite this huge flow of dollars from traditional offline channels into digital, right. We are now up to about $150 billion per year in digital spending in the U S, but yet, prices have gone down and that’s because the supply grew even faster.

And so I have this chart that shows, uh, Pew internet data, right? They’ve been the research firm has tracked, uh, humans, usage of the internet, social and mobile for the last 20 years. And I have in this chart, the green line and yellow line that has basically plateaued in the year since 2010, roughly. Whereas the blue line, which reflects digital spending, has continued to shoot upward.

So the area, that divergence between that blue line of digital spending and the green and yellow lines, which represent humans’ usage of the internet, mobile and social, that divergence has continued to increase. So basically, okay, there’s so many more ads and so much more ad spending that cannot be explained by more humans’ usage of the internet, social and mobile.

So it’s gotta be something else, right? Who knows what that’s something else? Well, I kind of know what that something else is. It’s bot activity. So, you know, we can kind of see, see a lot of that kind of stuff, but like you said, some of these marketers are literally very young marketers. They don’t have the experience to know that a 30% click through rate is not real.

Right, they wouldn’t have lived through the good old days, you know, like in 1995, 2000 where, you know, humans clicking on banner ads is probably 0.1% or 1 in 1,000 , you’re not 30%. Right. Okay. So without those frames of references, they think more is better, right. They fall back to these assumptions and very often they get caught up in buying the shiny objects.

Right. And the agencies are usually no help, because they love selling shiny objects because oh, blockchain, let’s go sell that.

Matt Bailey: [00:46:26] Right.

Dr. Augustine Fou: [00:46:26] Oh, you know, programmatic, let’s go sell that. And all of these things seem to have much better engagement as in high click through rates. Let’s go sell that. So all of these it’s kind of, you have been this, uh, uh, deadly spiral, if you will. It’s spiraled out of control where the agencies are, are selling shiny objects. The marketers are too young to know the difference, and so they end up buying the shiny objects, and once they bought it and they they’re used to these large numbers, they now need further justification to prove to their bosses that they did something good.

Right? So now they buy fraud detection that tells them there was no fraud. And so they show all these things. So again, it just piles up on top of each other, but the good thing is, you know, the good that’s come out of this pandemic is for the first time, they’re kind of taking a pause and saying, okay, um, we probably better look into this a little bit more because some of this is getting to be so implausible that, you know, before our CFO and CEO finds out, we better look into this.

And so I’m glad that’s happening, right? So they’re doing more audits of the campaigns and looking more carefully to see, does this even make any sense? Right. And so that’s a good trend that’s going into next year.

Matt Bailey: [00:47:36] Well, it’s interesting that you bring this up, that, you know, there’s a new shiny object and the agency is more than willing to sell the shiny object, which being in the agency world, I can tell you that agencies are selling it without knowing what it is. Uh, you know, it’s the typical sales, get it sold. We’ll figure out what it is later.

Dr. Augustine Fou: [00:47:52] And the clients, they’re not asking hard questions anyway.

Matt Bailey: [00:47:54] Yeah, exactly. Well, and most of the time when the client is asking for the shiny object, that’s what they want. And there’s a fake, I always used to say, you know, and this was 10 years ago or more that, you know, when someone had that shiny object, they were under the assumption that more visits equals revenue, and when you’re under that assumption, you make bad decisions.

You talked to, uh, John Marshall, who I just had on a couple of weeks ago for his book free is bad. And he goes also into the news aspect, and that’s where I want to take this is, who is it that tells us there’s a shiny object? It’s it’s the media. And especially, I think our, our marketing media does a lot of disservice by, you know, the flavor of the month, the shiny object of the month, this is the new thing, this is going to revolutionize things and, they don’t report on things critically. And that’s our own marketing media that, you know, we see it now in Tik Tok, it’s the new darling of the media, and, and I think that’s more driven by, if we have a story about Tik Tok, people will click on it and read it, which increases our ad revenue.

Dr. Augustine Fou: [00:49:07] Yeah. It’s, it’s more the, you know, if the new sites are advertising dependents, they have an incentive to write more salacious stuff, because there’s been studies over the years that say bad news spreads faster than good news, or at least gets higher click through rates and stuff like that.

So in those cases, the bad news sells pages, page views, right? And therefore sells ads. So, unfortunately, it’s created this negative incentive for new sites, right? If, if they report on some normal news, some of it’s just boring and nobody clicks on it and they end up reporting on, you know, the more sensationalistic type stuff.

Right. And that’s why, you know, clickbait works, uh Taboola and Outbrain know that, you know, firsthand. They know how to take advantage of that to get people to click on stuff. Right? So in those cases, the news organizations have been [00:50:00] pulled in this direction where they are, you know, writing more and more stuff that is more sensationalistic rather than fair balanced, and rather than, you know, critically analyzed from both sides.

Right. And I see that happen. And the other thing, like you mentioned, John Marshall wrote in his book, um, free is bad, it’s because when it’s free, they’re not making money off of you, the consumer. That they’re making money some other way. And in all likelihood it’s by selling your data. Right. And by selling targeting.

So just, you know, this kind of ties it full circle to where we started this podcast. Um, You know, the consumer’s data is now being collected and bought and sold and traded by these, ad tech companies. The consumer doesn’t know about it, or has no recourse after it’s being done to them, right? They didn’t give consent to the data being collected and they have no recourse after it’s been collected and not to mention that the data could be totally outdated and wrong.

Right. So, you know, you’ve seen these cases where, okay, we searched for, you know, uh, some baby products as a gift for a friend and then Amazon and all the ad tech systems think we just had a baby or something. Right. So again, they’re making all these wrong assumptions, the data’s wrong, the data’s polluted, and so you’re still targeting based on that.

And then the other simple thing, and I’m going to bring this back to, you know, just normal marketing common sense is that the hyper targeting is leading marketers. So this shiny object called hyper targeting, right? And so more and more marketers are throwing more money at that. The shiny object is leading marketers into doing way, way too much of that micro targeting stuff. And so they’re trying to retarget the same user that has either visited their site or have bought from them before. And so say for example, that’s 1% of the population have, have already bought from you before.

What about the other 99% that hasn’t purchased anything from you before? Right. So if you’re focusing all your budgets on targeting that 1%, you know, it’s almost like you’re trying to squeeze more blood out of that turnip. You can’t, right. I always like to say in class, right, a family buys four quarts of milk or can drink four quarts of milk a week, no matter how much milk advertising they see, they can’t drink a fifth quart of milk in the same week. They literally cannot. Right?

So that’s what I mean by you can’t squeeze more blood out of that turnip. And so when you’re hyper targeting and you’re trying, and you’re hitting up the same people with way more ads, you’re not only not going to get them to buy more, you’re going to get them pissed at you.

Matt Bailey: [00:52:31] Right.

Dr. Augustine Fou: [00:52:32] Whereas you’re not spending any dollars or enough dollars on the other 99% that have never bought from you before. So wouldn’t it be good if you just advertise and did branding, right? This again, closes the loop with what Bob Hoffman has, has been saying. You’re missing out on basic branding to the people who haven’t bought from you, because they don’t know about you.

And if you just allocate some of your dollars back to that, instead of to the shiny object called hyper targeting, then you would actually drive way more sales because all of those people in the 99% that didn’t even know about you now know about you and can actually buy from you. So the incremental benefits, incremental sales that you get from that could far far outweigh the couple of incremental sales you can get from these people who already have bought from you and probably don’t even want to buy more from you.

Matt Bailey: [00:53:24] Yeah. That’s a great, great point because I think we’ve seen a lot of brands, you know, they try to build their brand by micro-targeting and in doing, yeah. And in doing so you’re missing, you know, how much of the population that, that could be interested. I, I think that was a couple of years ago, Google came out with some numbers that, uh, I think it was 40% of people that buy baby products don’t have babies.

I think it was 60% of people searching for mobile gaming aren’t males between 18 and 24. It just goes to, you know, common sense that just because I think this is a good demographic target doesn’t mean these are the only people that are interested in my new brand.

Dr. Augustine Fou: [00:54:09] Yeah. So I think it’s led marketers down that path, and so I think this, this brings it all together in a nice, uh, neat package, which is the hyper targeting was the shiny object that ad tech companies sold as a bill of goods to marketers who were too inexperienced too real hot.

He said, okay, if I’m, if I’m just hitting that 1% over the head over and over and over again, that’s why they’re installing ad blockers, because they don’t want to see your ad anymore, right? You can’t squeeze more blood out of that turnip. You’ve already completely squeezed it dry, they’re not going to buy much more from you. So really it’s about finding that balance, focusing on incremental sales and whether you could actually take some of those dollars.

And it’s almost like having the pendulum swing back the other way, a little bit, right? As in swinging towards digital for far too long, chasing shiny objects and vanity metrics for far too long. If we start swinging the pendulum back the other way and doing basic advertising, right? Do brand awareness for the people, the 99% who don’t even know about you yet.

Uh, you’re, you’re going to get actually way more incremental sales than beating your existing customers over the head, try to get them to buy more.

Matt Bailey: [00:55:19] That is a great, great point. It’s interesting too. I wanted to bring this up. So there was a, a wired article that I had looked at for, you know, prepping on the, on the show here uh, and the, and the headline is can killing cookies, save journalism? And, uh, it was a great, great article. I was notified by my browser that they were trackers, uh, and they were blocking them. But the article, you know, I’m going to put it in the show notes, but it’s a great article because it brings us out of that, that, you know, like John Marshall, the free is bad. If it’s free news, that means it’s ad supported, which means, you know, there’s some trust issue.

Dr. Augustine Fou: [00:55:58] They’re making money in some other ways. Yeah. Um, and then there’s conflicts of interests that are built right in. And you know, that cookie issue on the wired article kind of ties back to your NPO example, right, where the publishers started to make more money when they stripped out all the ad tech. And that’s because again, the third-party cookies have led the market. That was a shiny object for a long time, right? Because programmatic and extra targeting that they’re enabled by third-party cookies, led marketers down the path to say, oh, well, we can actually show ads on all these long tail sites, and because we know the cookie, even if a human showed up on these long tail sites, we can still get an ad in front of them. Uh, what that’s disguising or covering up is the fact that there’s not actually a lot of humans that show up on those long tail sites. Right?

So the, the thought exercise I have my students do all the time is okay, quickly name off 10 domains, 10 sites that you visit every day. Okay. And then the follow on would be name off 10 apps that you use every day.

Matt Bailey: [00:56:58] Yeah.

Dr. Augustine Fou: [00:56:59] Most of them can get up to 5, 6, or 7 relatively quickly, but they start slowing down when they try to name the 8th, or 9th, or 10th app that they use every single day.

Matt Bailey: [00:57:10] Right.

Dr. Augustine Fou: [00:57:10] And that’s because they don’t, they literally use 5 to 7 apps regularly. Like I use Chrome every day. I use Gmail every day. I use, you know, new Google maps every day, but after 10, I can’t even remember the next app. I might have a bunch of installed. I’m talking about what you use. So similarly on websites, there are people who visit long-tail sites.

But there are not enough people that visit enough long tail sites enough of the time to generate tens of billions of page views to then support all this stuff that’s sold in programmatic. Right? So because of that misconception of the long tail of sites, uh, it’s led marketers, uh, to then chase these cookies on the long tail sites and all that, when we do away with the cookies, like the wired article says, we’re actually going to get back to more realistic digital marketing.

And that’s because without the third-party cookies chasing you everywhere around the web, you end up advertising on the sites that humans do go to in large quantities, right? And these are the mainstream sites that you’ve heard of like New York Times, Wall Street Journal, USA Today, Washington Post, whatever.

And in those cases simply by showing your ads to more humans because these sites have human audiences and less ads to non-humans like bots on long tail sites, you’re already getting better marketing outcomes. And that’s the part that the ad tech companies don’t want you to hear, because that means when you’re buying direct from these good publishers, less of the dollars go into the pockets of the ad tech middlemen.

So they don’t want you to hear this, but you know, don’t even take my word for it. Don’t even take the wired, wired article’s word for it. Do your own experiments, right? When you turn off digital ad spend in programmatic channels, or when you turn off the use of third-party cookies, do you see any change in business outcomes?

I’m going to bet you, I’m gonna bet every listener on here, they’re not going to see much change in business outcomes, so they should run that experiment.

Matt Bailey: [00:59:12] I love it. No, that is a, a great challenge I think, for any marketer who is questioning their spend and questioning what’s happening, uh, and it doesn’t take much go out and look, Uber did the same thing, uh, and they found that what almost a hundred million was a really just doing nothing.

Dr. Augustine Fou: [00:59:30] Yeah. It was basically the mobile exchanges were claiming credit for it, even though the organic installs would have happened anyway. Right? So when they turned off the paid spend, uh, the organic, the rate of installs kept happening because it’s flipped. They were all organic installs, anyway. And you’re about to say.

Matt Bailey: [00:59:47] Yeah, they did the same thing they cut way back, saw no difference. And so that caused them to reevaluate everything that was going on as well.

Dr. Augustine Fou: [00:59:57] And then the third example from recent memory is Chase. They cut their programmatic reach from 400,000 websites showing their ads to 5,000 websites showing their ads, right? That’s a 99% decrease.

[01:00:00] And guess what? No change in business activity, no change in business outcomes. So the other, uh, 395,000 sites, like what the heck were their ads being shown to? Right? Cause it’s certainly didn’t drive a business outcome for them.

Matt Bailey: [01:00:25] Wow.

Dr. Augustine Fou: [01:00:25] So, you know, the, the key thing is, again, don’t take our word for it. Run your own experiments. And this year the pandemic has given you the excuse, the opportunity to actually run those turn off experiments, right? Turn off, you know, not every marketer can turn off $200 million of digital spend like P & G did. You can certainly turn off $200 or $200,000 and try it for yourself. If it doesn’t impact your business outcomes, cut the next 200,000 and the next 200 until you start seeing some impact. Because right now, if you’re just blowing that money on digital marketing and not getting any incremental outcomes anyway, you shouldn’t be spending that.

Matt Bailey: [01:01:06] Absolutely. Absolutely. That is a great, great challenge. And, uh, Dr. Fou thank you so much for making the time to come on the show today. Always a pleasure, always a, a learning experience and yeah, I’m some of these things, yeah, I’m going to have to dig into a little bit further because the, the, the Google analytics injection is, uh, you know, that.

Dr. Augustine Fou: [01:01:31] I have a link for you. We did a proof of concept from a long time ago, so I’ll, I’ll send you a YouTube video of how that happens.

Matt Bailey: [01:01:38] I would love to see that, I’ll put that on the show notes too, for any of the listeners that are curious.

Hey, thanks again, Dr. Fou, how could people find you if they would want to learn more about ad fraud and what you have to say about it?

Dr. Augustine Fou: [01:01:50] They can just Google my name. So Augustine Fou, August plus I N E. And last name is Fou and they’ll find my LinkedIn, my SlideShare, and, uh, my Forbes articles. And then from there, they can contact me for any more specific questions.

Matt Bailey: [01:02:05] Great. Thank you so much. Listener, I hope you enjoyed this episode of endless coffee cup and, uh, boy, can’t wait to talk to you again. And hope you spend some time recovering. Thank you again, Dr. Fou.

Dr. Augustine Fou: [01:02:19] Talk to you next time. Thanks Matt.

Featured Guest:

Dr. Augustine Fou, Ad Fraud Investigator

Marketing Science Consulting Group, Inc.



Listen to Dr. Augustine Fou on an earlier episode:

Endless Coffee Cup: “The Amazing World of Ad Fraud”

Support Our Sponsor!

In Care of Coffee

In Care of Coffee

The money from your coffee purchase goes directly to coffee farming communities in Guatemala. Instead of money going to brokers, resellers, or distributors, it goes to the people and their families who work the land to bring you your daily coffee.

Use the promo code: ENDLESS  for 10% off your order.