Thomas Balling Rasmussen, founder of Effectiveness Advisory and former commercial director global marketing at Betsson, explains why operators may be dramatically overestimating how much of their digital advertising reaches real humans.
If your landlord walled off part of your office space yet kept charging you for it every month, I bet your CFO would act on this obvious waste. Something very similar is happening to digital media budgets across this industry, yet very little is done about it.
Any business relies on a range of inputs that enable it to create outputs. And it is a small but critical task to check that you receive the items you are billed for. If you pay for a box with 100 screws, you check that you received those screws. You cannot compete otherwise.
How do you open the advertising box? How do you know you took delivery of what you ordered? As a CFO friend of mine told me recently when I pressed the issue: “well… you simply assume this happens, right?” However, when credible efforts to quantify the issue of global digital ad fraud estimate the annual exposure to be at $70bn-$100bn (e.g. Juniper Research estimated 22 per cent of global online ad spend to have been lost in 2023), clearly, we should question that assumption.
With those numbers that takes the ad fraud “industry” to revenue levels in the same league as all of online gaming. And since they’re selling nothing for something, their profits might be even bigger.
For competitive marketing effectiveness, first your ads need to be seen by humans
Having worked on marketing investments across 100+ brand/market combinations spanning the cumulative equivalent of 250 years of quarterly spend and outcome data, I get asked one question constantly: how do ads drive sales? It is a fair question, and the honest answer gets complex. But at the level below this, there is a much simpler dynamic at play: ads that are not delivered cannot work. This question is less commonly asked at executive levels: do all our ads actually run?
Buying an advertising space is generally a matter of buying an intangible service. Sure, an out of home (OOH) physical billboard is easy to check on. But most ads cannot be touched. That is why e.g. TV, radio and print have a decades-old established ecosystem around this, which can mostly be trusted to deliver (still sensible to verify though).
Digital is different. Ad fraud precedes digital, but digital supercharged the issue. As attention moved online, media owners became heavily fragmented, down to each individual website or app. The digital world is characterised by vast complexity and indirect metrics. Ad buyers look at dashboards with billions of impressions over thousands of sites and mostly trust those numbers. But anyone can buy vast volumes of fake traffic to their sites. And many do.
Some of your own ads are likely to have been loaded in 1×1 pixel ad wrappers, entirely invisible but counted on the delivery sheets. Some into an alarm app on smartphones at 4am. Or perhaps on one of countless accidental-click-based webpages that load hundreds of ads before the visitor hurriedly closes the page.
And that is at the lowest levels of the digital jungle. Even at the highest levels we encounter serious issues. To put it plainly, advertisers have paid for premium ad placements, been told they received those premium placements, but received anything but.
The Wall Street Journal reported in 2022 that Gannett, owner of USA Today, had for nine months been misrepresenting inventory in its programmatic ad auctions. Premium advertisers paying for placements on USA Today found their ads had in many cases run on smaller regional titles in Gannett’s network, where the audience and the value were materially lower. The verification firms sitting in the supply chain did not raise the alarm. A journalist did. Gannett described the issue as a technical error, but the fact that no one noticed for nine months should give you pause. How do you actually know where your ads ran?
In 2024 the Wall Street Journal reported that brands paying for premium placements on Forbes.com had instead seen their ads served on www3.forbes.com, a subdomain operating outside the main Forbes editorial brand, where articles were reformatted into slideshows and listicles. In one example, a 700-word Forbes.com article was stretched into a 34-slide slideshow exposing readers to around 150 ads compared with seven on the original. Forbes shut it down and cited partner issues and technical errors. Whatever the origin story, many advertisers had paid Forbes prices for inventory that ran somewhere else entirely. Affected brands included household names spanning consumer tech, finance and media.
In neither case did the verification industry catch the issue, despite both situations being well within the scope of what these vendors claim to detect.
Surely we have Quality Assurance (QA) in our media buying?
Some will expect performance metrics and attribution modelling to manage the issue. These can be useful inputs, but they are not QA for media. They measure outcomes, not delivery, and many of them are gamed routinely by contemporary advertising fraud techniques.
The exposure is compounded by the legacy vendors tasked with managing ad fraud. They add some value, but independent auditing routinely uncovers large amounts of invalid traffic (IVT) those vendors miss. Part of the problem is that the digital ad supply chain tends to mark its own homework: vendors typically sit inside the ecosystem they are meant to police, and untracked impressions are commonly assumed clean. Relying on legacy vendor reporting can give teams and agencies a false sense of control. If they consistently report IVT of 1-2%, it is time to ask more questions.
This may matter even more in online gaming
You might think the online gaming sector is less exposed, since we can see whether customer accounts actually deposit money. But there are plenty of ways to get low or no value ad exposures stuffed into campaigns. Incrementality tests help but operate at a different altitude. Nothing uncovers the full story better than post-render ad analytics. And it is critical to not rely on sample-based post-render data; you need impression level data for every impression.
Another aspect of particular importance to online gaming is the geolocation of ads. Independent audits find geolocation ad targeting to be off by 1-15 per cent. With several European regulators having shown clear appetite for enforcement where gambling advertising is served into jurisdictions in which the operator is unlicensed, that is a meaningful regulatory exposure. Independent geolocation delivery evidence is a constructive way to minimize this risk.
Add to this that it is very easy to load cookies on human devices for ads that are never seen. This exposes affiliate efforts to attribution distortion, where ads are credited for purchases that would have happened anyway. Think promotional flyer handed to someone walking into the store taking credit for the sale. Sometimes this results in double paying for FTDs (first time depositors).
Open the box
I suggest you ask the following three questions of your teams:
- How much of our digital ad spend is independently verified, rather than verified by the platforms running the ads?
- What proportion of our ads are actually measured, rather than just assumed to have run?
- Can we evidence that our ads are not being served in jurisdictions where we hold no license?
This may all seem like a media buying effectiveness exercise. But I see it as a fiduciary challenge at the board level: if you simply don’t get what you pay for, real action is warranted. For many operators at scale in this sector, the non-working spend runs into the millions annually.
The good news is that most are surprised how easy the fix is. It is important to understand that like weeds in a garden, the issue can never be entirely removed, but it can certainly be managed. Post-render impression-level measurement is not exotic technology, and a small number of operators in this sector have joined some of the largest advertisers in the world in using it to recover material amounts of misallocated spend.
Stop funding bad actors. Redirect the spend to your growth. Count what is in the box.
About the Author
Thomas Balling Rasmussen, MSc, is an independent advisor and former commercial director of global marketing for Betsson. Since entering the iGaming sector in 2006 with PartyGaming he has held a range of leadership roles across marketing, commercial, analytics and finance departments. He has managed and advised on cumulatively +$2bn of marketing investments across 36 brands and 25 markets. Thomas is also an alumnus of The Wharton School at University of Pennsylvania and Copenhagen Business School. Connect with Thomas at www.effectivenessadvisory.com.