Marketers are poised to spend more than $350 billion on digital marketing worldwide next year. But yet, they cannot tell you if it works or not. Don’t believe me? Ask a few marketers and CMOs yourself. And listen for “hand-wavey-ness” and “wishy-washy-ness” and general “bull.” They assume that digital marketing works because sales are happening. But they cannot say concretely. Saying that some “brand favorability or recall” metric went up is not concrete. Sales were happening anyway, whether digital ad spending was occurring or not. Don’t get me wrong, though; digital advertising works (think DTC “direct to consumer” brands). But the spray-and-pray digital advertising done by the biggest brands is not advertising at all; it’s just “ad spending.” Here’s why.
You don’t know if digital marketing works; but some DO know it didn’t work
Most marketers don’t know if their digital marketing drove business outcomes. But some marketers found out concretely that their digital marketing did not cause sales or business activity, when they paused or turned it off. See: When Big Brands Stopped Spending On Digital Ads, Nothing Happened. When P&G turned off $200 million in digital ad spending, nothing happened to sales. When Chase reduced programmatic ads shown on 400,000 sites to 5,000 sites, nothing happened to business activity. When Uber turned off $120 of $150 million in mobile app ad spending, nothing happened to app installs. When eBay paused all paid search spend in the western U.S., nothing happened to site visits and sales.
Not all marketers can run “turn off” experiments — where they turn off digital ad spending to see if anything changes. Most don’t have the courage to lead such an experiment; and some hide behind the excuse that “if I turn off spending, but my competitor doesn’t; I might lose market share.” But that logic doesn’t hold up if your digital ads were not working in the first place. Those ads were not what helped you keep market share. To put it more bluntly, if your ads were shown to bots, those ads were not helping you maintain market share. Ad fraud is not the only problem. Other things can be optimized too.
You don’t know WHERE your ads ran; do you know IF they even ran?
Let’s start with something simple. Where did your ads run — like which sites and which mobile apps carried your ads? How many sites and apps ran your ads (think, the Chase example above)? Most marketers are not getting detailed enough reports to know where their ads ran. Some marketers are not getting any reports at all, just excel spreadsheets that tell them how much they spent. So how would they know where their ads ran, let alone if the ads ran on porn, fake news, hate speech, or disinfo sites? The only thing they are sure about is that they paid for it. And independent research shows that mucho dollars are flowing to fake news sites ,  and funding disinformation , like the disinfo that fueled the capitol riots and people refusing to wear masks because they are convinced the virus is not real. By the way, whitelists are suggestions; blacklists are too. That means that even if you use whitelists — a list of domains to include in your programmatic buys — are you sure your ads actually went to those sites? Or if you use exclude lists, are you sure your ads didn’t go to those sites? Many brands thought they blocked Breitbart, but they are still making lots of money from lots of ads. See: Dark Pooling At-Scale With IAB’s Ads.txt Protocol.
Oh, one other problem, if you don’t know where your ads ran, or if they ran on fake news and disinformation sites, how do you know if your ads ran? Bad guys can easily make the spreadsheets that show you nice large numbers and low prices. They falsified and fabricated reports to trick Uber into thinking they got billions of ad impressions and millions of clicks. See: Uber’s Lawsuit Comes Full Circle, They Won. Bad guys have also been caught faking the analytics — e.g. insert false clicks and traffic into Google Analytics when there was none (demo video below); or faking the attribution platforms to claim credit for sales and app installs that had already happened . Be sure to get proof that your ads ran, and where specifically your ads were shown.
Don’t know IF the fraud detection worked, but DO know you paid for it
Given all the fraud and other shenanigans mentioned above, marketers have also all paid for fraud detection, possibly multiple times over. The agency paid for fraud detection; the ad exchange paid for fraud detection; and the publishers paid for fraud detection. Everyone paid for fraud detection and justified it as the “cost of doing business” or “they won’t buy from us unless we have it.” But can you tell if fraud detection worked? How would you tell if it worked? These vendors are all “black box” which means they tell you a number — e.g. 1% IVT (invalid traffic) — but don’t explain how they got that number or show any evidence. How do you know they measured it correctly, caught all the fraud, or prevented your ads from going to bad places? You don’t.
A “mystery shopping” experiment illustrates this perfectly . Using a fake site — ecelebnews[.]com, he “signs up as a publisher & then sends 100% robotic traffic to the site. IAS only identified 17% of that traffic as fake & did not kick the site off the platform (they just told him to get it (the fake traffic) down to around 5%). Moat did better—successfully blocking almost 40%. (but missing the other 60%).” Countless other examples over the years show the fraud is still getting through in broad daylight. The fact that these detection vendors don’t catch anything or catch something once in a while for self-PR purposes is well documented . The fraud detection reports that have been reporting 1% fraud for years have been used for CYA purposes, not for truly detecting and reducing fraud .
Don’t know how much goes to showing ads; but DO know half went to middlemen
Last one. How much of your digital ad dollar goes towards showing ads? All of it, three-fourths, or a half? Sometimes that is hard to determine due to the complexity of the programmatic supply chain. There are many middlemen taking their cut whether they add any value or not. What IS known is that at least half of your dollar goes to such adtech middlemen, instead of to the publisher for showing ads. Wouldn’t it be better if you bought directly from the publishers so you save yourself the “50% ad tech tax?” Three industry-wide studies since 2016 have all confirmed that at least 50% of every dollar goes into middlemen’s pockets and not to the publisher for showing ads. What’s worse is that on average 15% of the dollars “went missing.” When the accounting firm PwC attempted to trace the ad spend through the programmatic supply chain from the advertiser to the publisher, on average 15% could not be accounted for. In one campaign they analyzed 86% of the dollars “went missing.” Do you know how much of your digital ad spending can be traced entirely through the ad tech supply chain?
Given that so much is being spent in digital advertising, but so much remains unknown or unsure, should you keep spending and assume everything is fine? Or should you look more closely (i.e. get detailed reports), run turn-off experiments to see what is actually driving business outcomes for you, and cut the costs of things that don’t do anything for you? Are you spending money just because you need to spend it? If you saved the money it actually increases your company’s bottom line more dramatically than typical marketing outcomes.
If you really really wanted to continue buying digital ads in programmatic, please follow the sage advice of these two seasoned practitioners in the space. When reviewing data on large quantities of faked impressions and devices on certain ad exchanges, they said
“Buyers that buy on exchanges and through DSPs that have metrics like you cite get what they pay for. They deserve to be ripped off. They should buy in well-lit environments with strict controls. Pay more but get way better outcomes,” tweeted Paul Bannister.
“The most effective way to counter this is to only buy from trusted intermediaries. App/site/channel lists won’t work. Curate the sellers you transact with.” said Jud Spencer.
The key is to work with sellers and publishers that you can see and trust. If your ads are being sprayed out to millions of sites and apps, most of which humans have never heard of, you’re probably getting ripped off. That’s why you’ve been kept in the dark and left unsure why or if your digital advertising is working or not. Marketers can take charge now, reduce risk, and be more sure that their digital ad spending is driving real business outcomes.