Episode Transcript
[00:00:00] Speaker A: So how can you identify those on your own? It's not easy anymore because of AI. This is 165 billion of theft. So you try to put it in perspective. So when people say, ah, it's probably not happening on my campaign, it's happening on your campaign, you just don't know what extent, but it's happening. You can eliminate fraud, you can grow your roas very quickly. So one of the things that we tell people is if you're spending any money in digital marketing, you're already spending money for fraud.
[00:00:25] Speaker B: Welcome to The Entrepreneur's Outlook Logbook I'm your host, Zach Bernard. You can find me on social at Zack B. In each episode, I bring on experts from various industries for you to learn about their strategies and insights driving extra business growth. Today we're joined by Rich Khan, CEO and co founder of Underwriter IO, an ad fraud detection platform that helps businesses identify and stop the bots, malware and fake clicks quietly draining your marketing budget. Rich actually wrote his first video game at just 9 years old and had it published at 12. A somewhat like early glimpse of the entrepreneurial journey and drive that would define the next three decade, decades of his career. After years of watching ad fraud quietly erode the performance of his own advertising business, he decided to launch Interrupt to give companies a real way to actually see what's happening with their traffic. And then today, Interrupt Technology has helped uncover the billions in ad fraud. Rich, it's great to have you on the show. Welcome aboard.
[00:01:24] Speaker A: Thanks for having me today.
Cool.
[00:01:26] Speaker B: Well, we've had a lot of people like marketing focus, but one of the areas that I find always very interesting because you think about fraud, you're thinking financial fraud, you're not really thinking about marketing fraud or anything like that. So it's not an area that we've ever talked about. So I'm kind of excited for this conversation here. We'll obviously dive into all of that here. But one of the first things, one of the things I always like to ask anyone that comes on the show, you've obviously been around the block, to put it very simply, built multiple companies built video games too. If you had to like restart maybe your company or your entrepreneurial journey from today, what's maybe the one thing that you feel you would do differently?
[00:02:08] Speaker A: This is a tough question, right? Because I've done a lot and I've always done it bootstrapping.
So I've bootstrapped 4 multi million dollar web properties just by self funding it. Let's call it bootstrapping. Just watching the dollar, a couple thousand bucks, nothing crazy.
And I just saw a story this morning that kind of reminded me of, you know, I, I know a lot of there are, I want to preferences. There are a lot of successful companies that bootstrap their way to, you know, hundred million dollar companies and up and they're out there. But there are a lot more stories of people that have built these big businesses by, by taking on investment. And it's not just the investment you know, you get typically when somebody puts a bunch of money into your company, they want to sure it's going to succeed. So they put a lot of resources in. So you maybe, you know, maybe if you're lacking on marketing expertise or sales expertise, they're going to find that expertise for you or they already have it in house, they can bring it in.
So I've never built a company doing that. So I mean it would be something I would have liked to have tried and maybe, maybe sometime in my lifetime I will try that. But I think one thing would be if you have a good idea and you get it off the ground and you get proof of concept, getting an early stage of funding to really fuel the growth. And the story I saw today was a company that's doing something that we're in the same spot. I won't mention names just to leave it out, but they just sold their company today to, for $2 billion. I mean it's astronomically like, it's just amazing. I'm excited for them because it's, it's great for the, for the whole industry to see a sale like that go through. But at the end of the day, like I said, I've always done bootstrapping. So it's always, you can only grow at the speed of cash.
It's nice when you have the, the cash and the extra experience that you may not have as, as a leader to bring in and I think that would have sped things up. So if I had to do things over or maybe my next, next go round I'll try to raise capital and try it that way.
[00:04:11] Speaker B: Yeah, that's kind of interesting because I've heard like the opposite from a lot of people. They're like, yeah, like I wish I, I did not raise capital. I kept more of like the equity or anything like that. But there, there's a pros and cons as you said. If you have someone that's investing in your company, they're the incent, their incentive in line to make sure that you succeed. They're going to give you the Capital, the advice, expertise, everything to make sure that it happens. So yeah, interesting answer. I've never heard that one. So always love to hear a different perspective on I guess just like diving into like the, the ad fraud side of things. For someone who's like never dealt with it or never really understood what it actually was, I'd love to understand like if you could walk me through what happens when a business brings you in. They go to you to like, like how do I find out if this is a real customer and, or if it's like just a bot just clicking my links and making, making me spend more dollars here?
[00:05:05] Speaker A: Well, I mean today with AI, it's getting more and more difficult to identify, but the back end data will always tell you what's going on. So I'll give you an example. Example. Let's say you're selling a product or service and the first conversion event is somebody fills out a form, says I want more information on your phone. Like they see an ad on Meta, it looks good. They, they click a link, they come to a, a VSL page, they read through it, they're like, wow, this looks interesting. They fill their information, hit submit and then you call this person.
No answer. And you call and call and call. You shoot a bunch of emails, no response.
More than likely that's probably.
And it happens where it's a real person who just got busy and decided at a spur of the moment they filled it out, was 2 in the morning and they forgot about and they don't want to, but it's a sign that it could be wrong. But a bigger sign is you get that person on the phone. So let's say Zach's information is filled out and your information is very readily available on the Internet. I fill out the form that's got Zach's information. I get you on the phone and you happen to answer the phone and I say, hey, you filled out information about growing a garden in your backyard or something where you're like, what are you talking about? I don't even have a backyard. Like why would I fill out this form?
That would be a clear sign that that was a fraudulent lead inserted. Now that would be the lead event. But remember, there might be 100 clicks that took place prior that you paid for, before that fake lead got generated.
So how can you identify those on your own? It's not easy anymore because of AI. So you need a third party solution that knows what they're doing to be able to identify that stuff. But what's interesting is a lot of people don't know this, but you know, some people, like, ah, it's fraud. Yeah, don't. I don't want to deal with it. Well, I just did the numbers for 2025 recently. I've published a paper on it.
Last year, as a global society, we spent over $750 billion on digital marketing.
And our conservative, our conservative estimate is that 165 billion was lost to fraud.
So, like, you know, when I hear that number, I look at the numbers all the time, I see the numbers and I'm like, that's 165 billion. But you know, how big is that? Well, I like watching Shark Tank. I don't know if you're an entrepreneur. That's probably a favorite of most of us, right?
[00:07:24] Speaker B: Yeah.
[00:07:25] Speaker A: But I love it when they get on stage and they're like, or on the, on the carpet, in the tank. And they say, they ask them questions like, well, how big is the market? What's the tam? Total addressable market? Like, how big is this? And like, well, you don't understand. Like the baby food industry is a $3 billion market or $30 billion market. And they're like, it's billion with a B. It's big.
This is 165 billion of theft.
So you try to put it in perspective.
I did read an article not too long ago that stated that the average fraudster earns 5 million a year.
So, I mean, there's a clear incentive. So when people say, ah, it's probably not happening on my campaign, it's happening on your campaign, you just don't know what, at what extent, but it's happening.
So what most people don't realize is that that's hurting your roas, which is if you're doing digital marketing, you're paying attention to roas. It's the easiest way, the fastest way to increase roas. I mean, there's a lot of things, you know, ad creatives and all that stuff. But the biggest quick win for that would be basically eliminating fraud from your account.
Yeah, if you can eliminate fraud, you can grow your ROAS very quickly.
[00:08:36] Speaker B: Yeah. Like, I know we've personally ran ads, definitely not the scale that I'm sure you've seen some accounts like, do it. But like, I still remember there was like Elite. It's like Peter Parkermail.com I was like, I mean, I know there's obviously funnel hackers and people like that that will just go in your funnel and try to reverse engineer and everything like that. But it is something that like a Lot of like, even like e commerce brand, for example, they're running ads and they see a spike in traffic and like what's going on here? Like the ads are not working properly or like what's happening oftentimes, Bots, ad fraud, all that fun stuff. And one other thing that I think people assume though is that like ad fraud is just like a, a bot clicking on like some links like overseas. But I'd kind of be curious to hear from your perspective, like what are marketers and business owners getting wrong about what fraud actually looks like today? Because I know we talked about like AI like really briefly like you said, like nowadays, like, hey, it's like very hard to detect. There's a lot going on there. So be curious to hear your perspective on that.
[00:09:36] Speaker A: So I agree a lot of fraud comes from overseas, but you won't see it.
So what happens is.
So I'm from Middletown, Delaware. Let's say somebody was using my information, right? They would basically go to a lead form, fill out my name, Rich Kahn, put my email address, my phone number, it all matches. They'd hit submit. Now if I'm the one who bought that lead and I see the IP address came from overseas, I would think twice about is this legit or not? Right. It'd be obvious giveaway.
But what these guys are doing is they're bouncing off of what's known as residential proxy IP addresses and so they can make that IP address appear from Middletown down, Delaware. So when you get the lead, it's got my information, everything lines up, everything checks out, it passes the test, you start calling that lead and you start creating all kinds of other issues. Yeah, so it's hard. I mean it's getting more and more difficult to identify fraudulent activity unless you're using a third party fraud solution like Anora. Our system is designed to help sniff that stuff out in real time so you can take advantage, take advantage of action on it right then and there. You don't have to wait until the end.
[00:10:44] Speaker B: Yeah, I mean I think like a good example, a good comparison, I mean I'm not sure if it's the best comparison, but if you look at like Stripe for example, like whenever you're trying to like process a payment, they have like the, I forgot the zip code. They're looking at a couple things like IP address and everything to make sure that everything looks good. There's no like fraud going on. That's in terms like financial transaction. But the same thing kind of happens when you have like lead forms and everything for like marketing, obviously you need to have a good solution that can detect all those things, go through all these tests and everything. But yeah, I mean I know that a lot of these people are going to go some very far length, just like trick some people by using, as you mentioned, like a residential proxy just to make it look like it's like the actual person.
I'm sure that in your time you've probably seen some pretty crazy operation like being ran by like some companies and everything like that try to go through your platform. Like, okay, that's pretty big, that's impacting like a lot of people here.
But I'd imagine, I guess one of the big hesitation that a company has in terms of bringing in a fraud detection company or a platform like an Earth, for example, is maybe the fear of losing the real customers and the process because you kind of have to dissect which one are the good customers, which one are not. And it's like I don't want to, for example, block someone who was never actually threat, maybe they were just traveling overseas and then they have a different ip, but they have a credit card like the US and what's going on there. I'd be kind of curious like how do you guys address that or prevent that in a way?
[00:12:12] Speaker A: Well, funny you should ask because I've been selling this product now for the last nine years. We've been building it for the last 20. Yeah, we built it for our own company because we needed it, but we've been selling it for the last nine years and I've done thousands of demos with real people, with companies have given me feedback and the biggest concern they have from using other solutions in the marketplace is false positives mismarking a real person as fraud. Because again, while fraud is binary, behind that click is either a, a real person or it's not. Right? That's the reality of things. But it's not like we have a camera in their face and we can, we can look at all this, you know, information and can and say it. So we've had to build it off of a digital footprint.
But what we found out is that the average fraud solution has a 15% false positive rate.
So that means, you know, there's going to be a number of real people that are trying to go through your funnel that are potentially good buyers for you that are going to get turned away based on using a fraud solution. So your knee jerk reaction of hey, I don't want to turn away any clients is because somebody has probably tested a solution in the past that's caused this issue.
I had a client, one of, one of our earlier clients, it was a great example because we were new in the space and we thought we had the better solution in the market. So we're like, all right, let's go up. And we had this client come up to us and say she's tried several different fraud solutions.
The false positive rate is too high for her to be able to rely on the data.
And she's like, I'm actually praying that there's a solution that actually works. But she basically said, you'll be the last one I test. Otherwise I'm just going to assume it doesn't work.
So she ran 18 million visitors through our network and spent three months manually identifying and dissecting that data, looking for a false positive, even just one time, where we mismarked a real person as fraud. And the nice thing is they were processing leads, so they have the backend data to back it up.
So if we said it was bad, they would go listen to the call recordings. Did that. Call recordings say that they didn't fill out a form. And if they, you know this. So they had human verified data. Just took months to do.
[00:14:28] Speaker B: Yeah.
[00:14:28] Speaker A: And after that study was over, they gave us a case study. They they, out of 18 million transactions, they couldn't find a single false positive in the group.
So it got us thinking. We run our shop. I don't know if your, your audience or yourself is familiar with a process called eos.
It's based on a book called Traction. We run it. We've been running our company that since 2019.
And one of the things that Traction will teach you is if you, if you do something different that nobody else in your marketplace does, can you offer a guarantee behind it?
So, again, in our space, offering a guarantee is tough, sure. But we finally rolled out a couple years ago a written guarantee for our clients that we are 99.999 accurate when identifying somebody as a fraudulent visitor.
If not, we have a whole process to give service credits to get the data so we can figure out what happens, so we can prevent it from happening again.
Look, no system is perfect, but that's pretty close. I mean, if you're wrong one in a hundred thousand times, yeah, it's not terrible.
[00:15:29] Speaker B: But you said 18 million.
That's an even lower percentage. I mean, no, technically zero, because you didn't have any.
[00:15:37] Speaker A: Right? The zero false positive, we aim for that, but we guarantee the 99.999.
And that's just what we do. That's really what makes us very different in the marketplace because we have people pushing billions of dollars a month through our network and asking us to decide what to do with that traff all the time. So I'll give you a perfect example of a false positive, how it negatively impacts a client. Sure. We had a writer that was working for us and she was working for us probably for about six months or eight months, something like that. And she went to go make a purchase at a fashion site she's done, she's used many times in the past. She got to the site, it was a new design so she recognized they made some changes, loaded up a couple hundred dollars of the stuff in her cart, went to check out and let her check out.
She recognized it, thought she was a bot or she was fraud. So she's like I, I work for a fraud company, I know what this is. Like let me just switch browsers.
She switched browsers, same thing. She says, you know what, let me switch devices. Yeah, switch devices. So she realized, okay, this solution is blocking my entire IP address for some reason, which is wrong, but they're blocking my entire IP address. So she went to a competitive site, similar company, loaded up the shopping cart with a couple hundred dollar stuff, made the purchase, no problems. And then what was most interesting was her response.
She said, you know what, moving forward I'm just going to use this new company because I didn't have any issues.
So the false positives can be very detrimental to a brand.
Whether you're an E commerce brand, a lead generation brand or even just generating clicks for volume. It can be very impactful in a negative way if you're blocking real people because that's that fine line of protecting your, protecting your brand from fraud while at the same time protecting your real clients from being able to, you know, look at it this way. We had the issues where you know, captcha was the big thing for a lot long time and there was a Harvard study that something like 35% of real people fail. CAPTCHAs select, you know, select the, the area with all the bicycles, select the area with all the buses and like you do it like six or seven times like and you give up. Right? But human fraud farms, right, which are out there in the marketplace, pass those because they're human, it's a real person. Now of course you got AI so that invalidates pretty much all that. But yeah, false positives, what you're talking about is detrimental. And like I said, I've tested a lot of solutions. We've had customers that come to us, go to other companies and then come back to us that accuracy means everything. And a lot of people are like, hung up on bells and whistles. Like, I want the platform to do this, I want the platform to do that. And I'm like, the crux of the whole system is two things, accuracy and thoroughness. And those are two distinct selling points that we offer over anybody else because we know how important it is.
But I get where you're coming from, where, hey, false positives are really bad.
So you really got to be accurate on your solution. And you really. I always tell people when you're looking for a fraud solution, and if you're spending money in digital marketing, you should have one.
In this day and age, you just can't do it yourself.
I had a lot of people with technical resources say, hey, we're going to build one in house. And I'm like, well, you've got to have a team of a couple of people at least that this is what they know. They're experts. In this space, you're going to have to have a data lake. You're going to have a processing center. You're going to have to have machine learning or AI that you build and design. You're going to have testing abilities. And I go, so you're going to spend a lot of money where you can spend a fraction of that and have access to a better solution that has access to not just your data, but everybody's data.
The more data you have, the more patterns you can find. So again, you know, in this day and age, I used to be able to give tips and tricks on, on how to minimize fraud. They don't work anymore. So I always tell everyone, you really need a fraud solution. So one of the things that we tell people, and I'm just going to drop it here as well, is if you're spending any money in digital marketing, you're already spending money for fraud. Right. So let's just say person spending $50,000 a month in digital advertising. And for some people, that sounds like a lot. For some people, it sounds like what they spend an hour. You know, it really depends on the client. But I'm just going to use that example.
On average, and I should say this for last year, on average, last year we were identifying 25% of the traffic to be fraudulent, which meant they were spending 12 and a half thousand dollars a month on fraud.
[00:20:07] Speaker B: That's crazy.
[00:20:08] Speaker A: Let's take it a funny step forward.
[00:20:09] Speaker B: Sure.
[00:20:10] Speaker A: I've identified that the majority of fraud that takes place is done by organized crime.
So if you think of it that way, you've donated twelve and a half thousand dollars a month to organized crime. Now that's just a funny way of looking at it, but you get where I'm coming. So you're already spending all this money on fraud. If you can solve that for a fraction of the cost and then take that 12 and a half thousand and move it to an audience that converts your roas goes to the roof.
[00:20:36] Speaker B: Yeah.
[00:20:38] Speaker A: It almost doubles your roas overnight. Now.
[00:20:40] Speaker B: Crazy.
[00:20:41] Speaker A: I said 25 fraud because this is a number I've been talking about for the last several years. I just watched fraud jump from 25 at the beginning of the year to over 40 last month.
AI is.
[00:20:52] Speaker B: Yeah, I was gonna say the market. AI is probably like a big component of the just adding like even like more frauds. That's.
[00:21:00] Speaker A: Let me ask you this question.
If it's, if it's at 40% right now and again this is just a few hundred customers that we have that we're tracking. Let's just assume that my, my sample is the average for the Marketplace and it's 40% fraud, what percentage of fraud would it take to shut down the whole thing?
[00:21:20] Speaker B: That's true.
[00:21:21] Speaker A: There's a number. I don't know what it is.
[00:21:23] Speaker B: Yeah.
[00:21:23] Speaker A: But if it, let's assume if everything you bought was 100% fraud, you would have no true back end conversion. You would stop spending money. Go, you go do something else. Right.
But between 40% and 100, there's probably a number where things start to drop. We had this happen back in 2000. I know I was selling a company in 2000. I lost the sale of a company two weeks prior because the dot com bubble burst.
[00:21:47] Speaker B: Yeah.
[00:21:47] Speaker A: It burst because everyone was sinking money into the dot com space and fraud was destroying it.
[00:21:53] Speaker B: Yeah.
[00:21:53] Speaker A: And they realized, wow, I'm spending all this money, I'm not getting any returns. Stop spending. And everybody kind of pulled out at the same time. It was April 2000 and that's when April 14, the dot com bubble burst.
[00:22:03] Speaker B: Yeah. I want to like ask you something though because you mentioned like. Yeah. What happens if we get to like that certain percentage. Obviously you don't know what that percentage is going to be. So maybe it's a weird question to ask, but what do you think will happen? Like if we get to a point where like in most transactions end up being like frauds. Because detecting platform can be great, they're going to help detect that. But in the Long run there, there will need to be some sort of like solution pushed out to ensure that we are still pushing out like the, the, the non false positive in like a way.
[00:22:35] Speaker A: So essentially what I see, I don't ever see it getting to the point where it shuts down because then the fraudsters lose out too. Right. These are smart individuals that realize that you know there's an old saying, pigs get fat and hogs get slaughtered. Right.
So yeah they're, they realize that if they, if they got too greedy that's gonna, they're gonna shut the whole system down too. But I think what's gonna happen is certain channels, so again not just looking at digital marketing as a whole, but certain channels. Yeah. Are less fraudulent than others just based on the way they're set up and way their way they attract their traffic.
So I think what you'll see is start seeing dynamic shifts from one channel to others as Ford become to bear in those markets or they license a fraud solution and they identify where the fraud's coming from and they're able to optimize that channel. So yeah, I'd like to see the, the second because that's the, the space I'm in and I hear that there's a huge market cap and total addressable market for our space over the next four or five years. So which I'm excited for the growth. But yeah, at the end of the day I don't ever seeing it getting to that number where it shuts everything down because I think the fraudsters are smart enough to know that that's how they get fed and people are smart enough to realize okay, if these channels are not performing, I'm going to do use other channels.
[00:23:51] Speaker B: Yeah, maybe at some point like you'll go to like a stripe payment page, try to do like a purchase and ask like do like a selfie or something like that before he can pay for it. Maybe we'll get to that point. I hope not.
[00:24:02] Speaker A: But anything you could think of is ultimately going to be spoof. But again yeah, it's spoofable. But it comes down to what are you willing to give up in order to get security. Right.
So right now there used to be a study and I'll go back a few years when you put a captcha up the percentage of clients that would drop off.
[00:24:23] Speaker B: Yeah.
[00:24:23] Speaker A: Because they would get to that form and they were like ah, I failed. I just go somewhere else. Right. Because it becomes too much friction for the buyer.
So again if you had it to grant access to a third party Website to look at your camera to confirm, confirm it's you for facial recognition, which facial recognition is very beatable still, especially with AI.
How many people will only give up that. Right. That a third party, unknown site can have access to their camera?
[00:24:50] Speaker B: Like, yeah, you know, that's an extreme example for sure. Yeah.
[00:24:54] Speaker A: But I'm trying to say is like there's that fine line of being able to be frictionless for the buyer while at the same time protecting your marketing budget. And I think I, of course I'm going to be a biased opinion, but I think we've come up with a solution that does just that. It's a frictionless component that can tell you if the person's real or fake in real time with extreme accuracy and it solves that problem.
[00:25:18] Speaker B: Love that. I think like the, the overall like theme is make sure that you have a good solution, whether it's something else or like a neural as a platform to ensure that you have accuracy is going to be like the main component here because you want to make sure that you don't get false positive, but you also to make sure that you're able to detect the fraud and increase your oas and you're not spending money on like ads that are obviously just going to organize crime as like you mentioned.
[00:25:41] Speaker A: Yeah.
[00:25:41] Speaker B: So a pretty important one here. I love it. Rich, I've definitely learned a couple things based on this conversation, but anything, any, like, parting thoughts, like any other thing that comes to mind, like, hey, if you're running a brand or you're maybe spending on ads, this is what you should be looking out for.
Anything that comes to mind there.
[00:26:01] Speaker A: Sure. This is my shameless plug on your, on your podcast.
[00:26:04] Speaker B: Go ahead.
[00:26:05] Speaker A: But it's a useful information. So. Okay, if you guys are spending money on digital marketing, the first question you need to ask yourself is how much fraud am I dealing with? So go to Anora IO. We offer a free traffic quality audit and there's no cost or obligation, no credit card needed. We're going to give you two things. One, we're going to tell you exactly how much fraud you're getting and where it's coming from.
And then if it makes sense, we'll sit down and have a discussion on how to mitigate that fraud. But at least, you know, hey, I don't have a problem and you'd be the first.
Or I have a problem and this is how much of my budget's being lost to fraud. Yeah, you know, do I continue to spend that $12,500 on fraud? Or I spend a few dollars and solve that problem then focus on other problems that I have in my business.
[00:26:49] Speaker B: Yeah, I don't speak to like a lot of people that like you. You hear like their offer and you're like I don't want to be biased. Again like like you said you, you don't want to be biased but like a no brainer where it's like hey you do this like you get this out of doing that. So yeah we're going to put that in the show notes description so like anyone can check that out. I if you're spending on ads, highly recommend that you go to that little Luck quality audit as Rich mentioned here. And then yeah to your listeners if you've enjoyed this episode like subscribe, review, comment, all that fun stuff. Reach out to Rich here. I know he's pretty active on LinkedIn as well here and I'm sure he'll be happy to get back to you and help you in any way can. And yeah until then keep pushing and
[00:27:31] Speaker A: see you in the next one.
