How to Make Money from Ad Fraud in Seven Simple Steps
Common sense can often be the best practice you employ.
In a nutshell, optimise your campaigns by checking the websites your ads are appearing on as you may be the victim of ad fraud. Fraudsters may use a simple seven-step process to set up a website you think is legitimate and start taking your money. Using a bit of common sense rather than a full reliance on technology can often be the best practice you can employ.
Fraudsters know that optimisations, mechanical or otherwise, are often done by looking at the performance of media metrics and the website attributes. This is something sites like mortgageafterlife.com appear to be taking advantage of.
I'm seeing the symptoms of corner cutting optimisations emerging at the intersection of traders being inundated with data to review and the myopic obsession with "time efficiency" or as I like to call it "do more for less".
----------
If you take the time to be selective with the context that fits your audience and brand, you can avoid many common brand safety problems and advertising on websites that are not what they make themselves out to be.
If you subscribe to the ‘Audience First’ ideology, it is important to remember that context is still an important component that needs to be considered. Neglecting the action of reviewing the context, as you’ll soon find out, creates an opportunity for fraudsters to take advantage of your buying and optimisation technique, while they take your money.
Above: Time poor programmatic traders, circa 2018-19
I don’t expect this to ever have been the intention of the optimizer, the optimisation process, or any optimisation mechanisms set up to make the most of the campaign results. The intention of the actions is most likely to make the best decision possible, as fast as possible. As I’ve written in the past, time efficiencies are continually being sought after, with the high-level vision being for traders to have a greater reliance on automation technology to make the call of whether or not the inventory is worth bidding for. It turns out though that robots can be fooled, and sometimes rather easily.
When numbers are taken on face value and there is no human digging beneath the surface, every algorithm can be taken advantage of in one way or another. This is the same concept as to how an influenza virus continually works its way around the prevention mechanisms our bodies set up every year. This process of finding holes in armour has been around since time immemorial and even had a place in Star Wars with the Death Star.
No matter how much tech you slap on things, having a human with a bit of common sense review decisions made by machines can often be the best practice you can embrace.
This story starts a few months ago when I was reviewing a client’s past advertising initiatives and performance. This is part of the process I use to design and define their programmatic advertising best practices. In going through the top thousand or so sites of all their historical performance, I noticed two peculiar sites that had appeared. Both seemed legitimate from the domain and top-level structure of the page, but when I applied my secret algorithm, the fraudulent sites lurking in this whitelist were exposed.
The secret to my algorithm is my judgement and common sense, both of which have come from experience in setting expectations and standards for teams, as well as designing best practices for agencies and brands.
Those that know me well know that I like to go through things in detail and check as much as possible when conducting an audit. When reviewing websites, I open tens of URLs at a time to review each one individually. My process is that each website gets checked, assigned a manual category about what I believe the context to be, and then judged based on whether I believe it fits in with my client’s brand and campaign objectives. It can be more work, but it is the right thing to do in the short, mid, and long-term.
When I found these websites, I spent a bit of time thinking about the countermeasures these fraudsters had put in place to get them added to the whitelist (they were smart about this). I’ve attempted to reverse engineer the approach they used to start collecting money from unsuspecting advertisers. I arrived at seven steps which I present to you below. I’ll be focusing on one particular website in this example, MortgageAfterLife.Com.
Seven Simple Steps to Making Money from Ad Fraud.
1. Register a domain.
When you do, make it seem like a legitimate website, one that if someone were to just read it in Excel when reviewing performance would think, ‘Hmm, that seems legit. I’ll add it to the whitelist WITHOUT CHECKING IT FIRST’. As mentioned above, I’ll use mortgageafterlife.com. If you were to read this at a glance, you would probably think at first it was a website about home loans or mortgages, right?
2. Use a website template.
Base the site structure on a template with a top-level menu that makes it clear what you want people to think you do. That way when it gets crawled and listed on Google, the site structure contributes to demonstrating validity. Don’t forget to do the meta descriptions for the homepage as well as the subpages. The hoop being jumped through here may be one where a whitelisting bot reads the domain, meta description, top-level menu and URL category.
Seems legit, it is categorised as being in finance, the meta text tells me it is about legal advice, retirement, loans & credit and refinancing.
3. Send heaps of traffic to the site – Fake or from social.
If you’ve been wearing your detective hats and monocles, you may have noticed the rankings in the SERP screenshot in step 1. The key to getting high volumes of traffic to your site from social, as it seems for this shady player at least, is to use an endless stream of content in the form of clickbait tabloids, trending themes, and general trash.
Ah yes, I see here. The Rock and Mortgages, a match I can definitely see my brand appearing in as valuable. When people think AirPods / The Rock, I want them to think ‘Mortgage’.
4. Get listed on ad exchanges, Google AdX, AdSpot, Taboola to name a few.
If you’ve done your detective work again, AdSpot lists the details and prices of inventory on this garbage website here https://adspot.co/website/mortgage-after-life/466
Good thing that they have this snapshot page. You can see that the description on the left, in fantastic grammar: Learn everything latest about finance and business, legal advices, loans and credits and retirement.
There’s even a screenshot to boot so you can see exactly what you’re buying and how much for. $5USD for 1,000 sometimes viewable impressions? Yeah, I’d rather invest in the Opal tower thanks.
5. Recycle content and repost, repost, repost. Maybe even Spin between.
I’m not entirely sure where this content comes from as there are no authors on any pieces. Seems like all contact with these fraudsters is meaningless. Probably a network with a central distribution hub. Just nick other people’s crap and repurpose it—I mean, there was an article about Ariana Grande’s Ponytail Endorsement Deal, so you can just make anything up next time you’re at the pub.
6. Get a high click-through rate or cheap CPC so you seem like good value.
Don’t be fooled though, these clicks might have a good result on the media metric front but as soon as you look at what happens on the site, the results are usually garbage. Excessively high bounce rate (based on the index of the page), extremely low time on page if there is no bounce, single interaction then exit, no real business goals being achieved…you get the idea.
7. Make money. From these guys specifically.
Here’s a list of advertisers I found on this site. If you’re on the list, you might want to do a review of your programmatic initiatives.
Advertisers I found in an article titled, 'Dwayne ‘The Rock’ Johnson is giving AirPods a run for its money with the launch of this new product'
- Pandora Jewellery.
- NSW Gov, Workplace Safety.
- LinkedIn – Outstream video. Auto-playing in the bottom right-hand corner of the page. Questionably ‘Great’ completion rate and other vanity metrics. Horrible for just about everything else including user experience.
These guys on an article titled, ‘Chefs, Get In Here! Here’s an opportunity to live in the Buckingham Palace’.
- Tableau (No frequency cap with these guys, they were on every page)
- Dan Murphys
- Zoho One
- Kia
- IC Markets
- American Express
- SSangYong
These guys in an article titled: 'ARIANA GRANDE’S PONYTAIL GETS NEW ENDORSEMENT DEAL!'
- Prospa – Just about every variation of the ad, one after another.
- Qlik – Tableau was feeling lonely
- Shutterstock
- The Motley Fool – Text ad on GDN.
- Blackmores
By this point, I couldn’t be bothered taking screenshots and refreshing the browser anymore, but I think you get the idea. Want to see your advertisers there? Just visit the programmatic money drain that is mortgageafterlife.com
So to recap, here is the checklist that this website seems to have used.
- Have a legit-seeming domain relevant to Finance
- Have a legit meta description, relevant to Finance.
- Have a top-level menu that is relevant (Finance & Business, Legal Advice, Loans & Credit, Refinance, Star Advisor)
- Be categorised as Finance
- Ensure descriptions are about finance: Learn everything latest about finance and business, legal advices, loans and credits and retirement.
- Make the about us section seem legit. Every sentence is a lie starting with ‘At Mortgage After Life we understand that the most valuable investment you’ll ever make is your house, so we dedicated our entire site to homes and houses only.’ Yet every article is garbage clickbait.
A final note on Common Sense
It won’t take much for you to at least think about reviewing your optimisation methods here and bring it back to context and basics. Think about where you’re advertising not just about to which cesspool of cookies you’re misguidedly pegging efficiency on.
If you’re an agency and are seeing these sites in your list, remove them if they don’t work for you. If you think this site works for you, keep it on. If you’ve got any shadow of a doubt about how your client/s would feel if you were to open the website in front of them at your next WIP and for them to see their ad there, you probably shouldn’t be there in the first place.
Make it happen.
9 个月Now even easier! The Wall Street Journal recently published an article by Jack Brewster who had such a site built by paying $US105 to a freelancer on Fiverr.com. “Purchasing an AI content farm on Fiverr.com is as easy as ordering on Uber Eats,” wrote Brewer, who is also enterprise editor for NewsGuard. “I searched ‘AI generated news website’ on the home page and up came dozens of developers offering to build my site.” The developer said ChatGPT was ordered to rewrite articles which, he argued, meant the result was a new work “fresh, copyright-free content with no plagiarism”. Ongoing costs: Based on OpenAI’s listed API prices, $3 or less a month to publish 50-100 articles a day, plus $5 a month for a web-hosting company. https://www.adnews.com.au/news/ai-generated-news-sites-masquerading-as-the-real-thing-now-in-australia
Join our 10th Anniversary at B2B Global Conference on 25th of October at Parramatta | Up to 50 exibitors | 10 plus sponsor | 200+ Attendees
1 年Victor, thanks for sharing!
I help B2B tech companies propel their GTM and scale revenue
5 年Great article. "No matter how much tech you slap on things, having a human with a bit of common sense review decisions made by machines can often be the best practice you can embrace." Completely agree with this and not only talking about ad fraud but it applies to everything we do. Nice work
Founder & CEO at Com Olho | Crowdsourced security that never sleeps
5 年Nice write up. It’s great to read and understand that more and more people are exposing the unethical practices existing within advertising sector which is leading to data theft and financial gains.