As advertising budgets get tight, remain tight, and come under increased scrutiny from CFOs and CEOs, marketers must not only increase the effectiveness of their marketing, but also look for ways to cut unnecessary expenses. The savings drop straight to the bottom line. A dollar saved is a dollar earned on your P&L, the “P” line, literally. The dollars saved have significantly more impact on profitability than the many dollars spent on getting more sales, especially in low-margin categories and industries.
Take for instance P&G cutting $200 million in digital ad spending, and seeing no decrease in business activity. In fact, during the same timeframe, “Jon Moeller, P&G’s chief financial officer, said the reduction in ad spending obviously didn’t hurt the company’s performance in the recent fiscal year because sales increased, up 1% from the previous year [and] organic sales grew 5%.” P&G saved another $750 million by 1) slashing its agency roster by 50% (CNBC), 2) reducing the complexity and number of adtech middlemen in the programmatic supply chain through in-housing (FT), and 3) increasing price transparency and accountability by buying from vendors directly and avoiding undisclosed mark-ups by agencies. Of course not every marketer can save a billion dollars like P&G did, but marketers of all sizes can find and cut unnecessary costs. Here are a few to start with.
Cheap Ad Impressions on Long Tail Sites
Many marketers, especially the biggest spenders, were lured by the large numbers of ad impressions and low prices offered by programmatic ad exchanges. These ads were supposedly shown on millions of “long tail sites.” But common sense will tell you there aren’t enough humans visiting millions of long tail sites enough times to create the trillions of ad impressions bought and sold through programmatic exchanges; fake sites and fake users (bots) are creating those impressions out of thin air. Those sites can sell ad inventory at low CPM prices because they pirated or plagiarized all the content and had little to no cost. There’s a reason mainstream publishers can’t sell you ads for much lower costs; their content is expensive to make. Good publishers have real journalists, real editors, and real content and therefore real human audiences.
The first thing to cut is unnecessary ad spending on large quantities of low cost ads. If it seems too good to be true, it is (too good to be true). You can do this aggressively by turning off Google Display Network and search partners and by turning off Facebook Audience Network. You can also start to use an allow-list or include-list of domains, instead of a block-list; there are so many fake and fraudulent websites you can’t block them fast enough. Don’t take my word for it; run your own experiment cutting low cost ads on long tail websites from your media buys. See if there is any change to your business activity and outcomes. I’ll bet you that you won’t see any difference. When Chase cut the number of long tail sites showing its ads from 400,000 to 5,000 (a 99% decrease) they saw no change in business outcomes, either (NYTimes).
Buying Your Own Branded Keywords in Paid Search
Many marketers, especially the biggest spenders, were duped by their agencies into buying their own branded keywords in paid search marketing. Why would agencies do such a thing? It is possible they didn’t know better. But it is also possible that those keywords are by far the highest volume ones (people knew the brand already) and had the highest click through rates, by far (people were going to click on the search result anyway). Those brand advertisers whose branded keywords already appear as the first organic search result should not be paying for search ads on those same brand keywords. Otherwise, a search ad is placed directly above the first organic result and users accidentally click the paid ad instead of the organic search result they would have clicked anyway. This costs the advertiser unnecessary expenses (see chart below). Agencies won’t tell you that because without the high volume brand keywords and high click through rates, their excel spreadsheets wouldn’t look as good, and they’d have to do more work to drive results for your paid search dollars.
Google search results (screen shot)
Fraud Verification Tech That Doesn’t Work, and is Black Box
Many marketers, especially the biggest spenders, were tricked into thinking they can detect their way out of trouble when buying low cost ads in programmatic channels. They thought that if some other party checked the ads for bot activity and fraud they would be protected from wasting money. But what they didn’t realize is that the bots can easily trick the detection too, and not get marked as “invalid” — as in IVT “invalid traffic.” Not being marked as invalid, does not mean that it is valid. It could be that the detection simply failed to detect the bots. To put it simply, bots can either alter the measurement with malicious code to defeat it; or bots could simply block the detection tags of the fraud verification companies to avoid detection altogether. Yes, read that one more time — the bots simply block the detection tags of these vendors, and get away with it. “Not measured” is also not the same as “valid.”
See: Examples Of Incorrect Measurements By “Black Box” Fraud Detection
Don’t take my word for it; turn off fraud detection for a period of time and see if you notice any changes in your digital campaigns. I’ll bet you that you won’t see any change. And you’ll never be able to verify if they measured correctly or not, because these detection technologies are “black box.” This means they give you a number — e.g. 9% IVT — but do not explain how they measured it. Even the MRC-accredited vendors may not be measuring correctly. MRC accreditation simply means they paid a fee to MRC, MRC had an accounting firm — E&Y — interview the vendor and verified they are measuring what they said they would measure. That is entirely different from measuring bots vs humans correctly.
Cut the fraud detection tech that does not work; you can’t verify if they work anyway. Oh, and definitely stop paying those fraud detection tech companies that extort their own customers, by threatening to mark them as fraudulent or remove them from indexes if they don’t pay up. That’s “protection money.” Don’t pay it. And you wouldn’t need fraud detection in the first place if you bought ads from real publishers, instead of millions of fake sites with fake traffic.
Fake Certifications
Many marketers, especially the biggest spenders, were lulled by fake certifications offered by industry trade associations into a false sense of security. They thought that if they bought from TAG Certified Against Fraud vendors that there would be less fraud. But what they didn’t realize is that the “certifications” are pay-to-play and self-attested. This means vendors pay the fees, complete paperwork, and promise they will do no fraud to get their certifications. Do you think bad guys intent on committing fraud will actually follow the rules? Even if they are caught three times, TAG’s own documentation shows consequences that amount to less than a “slap on the wrist” and “all consequences for non-compliance will be held in abeyance during the pendency of an appeal before TAG.” If ever they are caught, fraudsters will just appeal and continue business as usual “during the pendency.”
TAG documentation
TAG Certified Against Fraud companies were caught committing fraud. For example, Newsweek Publishing Group was caught committing ad fraud while they were TAG Certified. Executives later plead guilty to committing the fraud, so there was no question about whether the fraud happened or not. And Newsweek remains TAG Certified to this day. All of this makes you wonder about the efficacy of pay-to-play, self-attested “certifications” offered by industry trade bodies that have an incentive to reassure everyone that they can keep spending.
TAG Registry for Certified Against Fraud Vendors
Judge for yourself whether this expense is necessary, or useful at all. You can buy ad inventory that is already “filtered for IVT” and achieve the same “low, low fraud rate of 1.05%,” without paying for TAG certification. Just note that it may be that low because the fraud detection tech couldn’t detect the bots. So better yet, buy more from real publishers and less from programmatic long tail sites. You’ll have reduced your own exposure to ad fraud, certifications or not.
Targeting Data and Look-Alike Audiences
Many marketers, especially the biggest spenders, were suckered into paying for more targeting parameters thinking that would improve “relevancy” of their ads. While some targeting is better than no targeting at all, too much targeting is a complete waste of money. Beyond about 3 – 5 targeting parameters, the extra expense of hyper-targeting outweighs the incremental business outcomes that it drives. In fact, academic research has shown that more targeting, beyond a point, results in less business outcomes than less targeting. This is because the data used for targeting was crappy. The data was collected in privacy-invasive ways (observing what sites users visited, tracking them across sites, setting cookies, etc. all without their knowledge or consent). The insights and audience segments were entirely derived, because none of the users were logged in and none of them volunteered information themselves. (This is the opposite for so-called “walled gardens” where users are logged in to one or more Google services at all times, and voluntarily provided demographic and other info in exchange for the free services they were using).
See: Despite Claims That More Targeting Means More Relevant Ads, Nope. Here’s Proof
On top of the general crappiness of the data, there’s also the bot problem. Bots know that ad tech companies look at website visitation patterns to deduce what a user likes and which audience segments they belong to. So bots conveniently trick the ad tech targeting companies by visiting medical journal websites, so they look like doctors; or by looking at swing sets in the spring, and backpacks in the fall, so they look like “swing set intenders” or “back-to-school intenders,” respectively. These are bots, not humans who really want to buy from you. By pretending to be in high-value audience segments, like doctors, bots can trick pharmaceutical advertisers to pay higher CPMs in their desperation to get ads in front of physicians. Bots make more money; advertisers waste more money on said bots.
Don’t take my word for it. Cut the targeting data you pay more for in your digital campaigns and see if you notice any impact on your campaign outcomes. I bet you there’ll be none that you can notice.
Remarketing to Your Own Customers
If you thought black box fraud detection, fake certifications, and targeting audience segments of bots were ridiculous, wait till you hear this next unnecessary expense. Ad tech companies are charging clients to “remarket” to their own customers. They even require ecommerce merchants to upload email lists of customers who already bought from their online store, so they can remarket to them. Think about that for just one second longer. These are customers who already know the brand, and like it enough to have already purchased from their ecommerce stores or physical stores. What’s the likelihood of this customer buying more from the same advertiser? Pretty likely, if not 100%.
Oh, the remarketing company told you they have some special sauce that gets these customers to buy more from you? And they tell you not to worry because you only pay when you get the click anyway (performance)? They tell you, trust us, you don’t need to know where we run the ads since you’re only paying when you get the click. Did they run your ads on porn sites? hate speech sites? fake news sites? popunders? And they don’t provide you with placement reports of where ads ran? Are you sure that ads ran in the first place? Remember Uber’s lawsuit against 100 mobile ad exchanges for falsifying placement reports or fabricating them entirely when no ads were ever run.
Cut your remarketing expense and see if the sales continue. If they do, then those sales would have happened anyway – because your customers know you and will buy from you again — not because of some magic done by the remarketing company. The only magic they are doing is tricking you into paying for sales that would have happened anyway. See: Fraudsters Cheat By Tricking The Reporting to Look Awesome
Cut, Cut, Cut More
While you’re cutting, note the 50% “ad tech tax” when buying through programmatic supply chains. Fifty percent of every dollar you spend goes into ad tech middlemen’s pockets, instead of towards showing your ads. What if you bought direct from a good publisher? Every cent of your dollar will go towards showing your ads; that’s the whole idea behind digital advertising anyway, right? Save yourself 50% or more by “buying direct” from good publishers. They will still use programmatic technologies to place the ads, but you are shortening the supply path and cutting out the middlemen who are trying to maximize their own profits, on your dime.
If you were to look even more closely you may find that between crappy targeting due to crappy data and other “drop-offs” due to the limitations of technology, you’re left with 6 cents on the dollar going towards showing ads. It’s almost like a sale at Macy’s — save 94%! By cutting unnecessary expenses in digital, you’re saving money. The above examples are not a case of “save more when you buy more” (like at Costco, or UNinformed programmatic media buying for that matter); it’s “save more when you cut more.”
See: The Cost-Performance Paradox of Modern Digital Marketing
Cut costs; increase savings, and get better outcomes all at the same time. Who doesn’t love that. Go check your own spending. And let me know what unnecessary costs you decide to cut.