More Marketers Are Running ‘Turn-Off’ Experiments. Should You?

Following the lead of P&G [1], Chase [2], Uber [3], eBay [4], and Airbnb [5], more and more marketers are running their own “turn-off” experiments. Some of these experiments started during the pandemic in 2020, when marketers paused their digital spend, but found that their sales continued, or even increased. Now, as marketers are turning digital marketing programs back on and facing tighter budgets and more scrutiny going forward, they are looking more closely at what worked and what didn’t work.

In the past, many marketers didn’t want to “rock the boat.” They preferred the relative safety of continuing to do what they’ve always done. A little voice told them “no one got fired for buying more ads, especially at lower prices.” And that worked nicely with their need to “spend it all” anyway. But that’s not going to fly any more going forward when the CFO and CEO are looking more closely at digital spending and asking if it were actually generating incremental outcomes. Even with this in mind, most marketers are not prepared to turn off $200 million like P&G or $800 million like Airbnb. What follows are examples of smaller steps that marketers are taking right now to truly test the effectiveness of their digital marketing, and to improve it in the process. 

Turn off domains and apps

Marketers, do you get detailed placement reports that show you where your ads ran? More specifically do those reports list the domains and app names and the quantities of ads that ran on each? If you’re not getting that right now, you’re getting ripped off — your digital ad budgets are flowing to fake or non-existent sites and apps, and definitely now being shown to humans. Get yourself those reports. Next, do you look at those reports? If you did, you’d probably notice strange domains and apps, like the ones listed in the following slide. Even if you wanted to visit one of those sites, how long would it take for you to type that domain name correctly? Right, those are domains and apps that no humans would ever use. But countless such sites and apps are being used to commit digital ad fraud at scale. When you see such sites and apps in your placement reports, you can add them to a block list to stop wasting money on them. 

Turn off (b)adtech targeting

Marketers were led to believe that more targeting is better — i.e. more relevant ads for consumers. But sadly, the only people who think ads are better targeted are the marketers themselves and the ad tech companies that sold them the targeting services. If you asked any consumer, they don’t recall seeing the ad, they don’t recall the contents of the ad, even if they saw the ad they don’t think it’s relevant, and all of the above apply only if they didn’t block the ad in the first place. 

Furthermore, did you know that most of the targeting and audience segments sold by ad tech companies are inferred? This is because users are not logged in when visiting content websites. So ad tech companies plant hidden trackers on publishers’ sites to track users. They use third party (“3P”) cookies to identify anonymous users and re-identify them as they visit other sites. Over time they build up a site visitation history for that cookie/user and infer characteristics about them. But even for the most basic characteristics like gender and age, the inferences are comically bad. 

For 1 parameter – gender – the average accuracy was 42%, worse than if no targeting was used (“spray and pray” would have hit about 50%). For 2 parameters, accuracy drops to 24%. Note that this is entirely different for Google and Facebook, where users are voluntarily logged into various services, and likely provided accurate information when they created the accounts. Those are far more accurate than gender, age, and other targeting parameters inferred from website visitation patterns. So marketers should turn off targeting parameters and see if anything changes. Astute marketers will set up new campaign lines with no targeting, to compare to existing campaign lines that have targeting, and see if there are noticeable differences. If there are no differences, then why are you paying extra for ad tech targeting parameters?

Turn off programmatic ad buying

If you’ve stuck with me so far, and perhaps have run some of the experiments above, then you may be ready to take the next step — and run the experiment of turning off programmatic media buying entirely, for a week, a month, or longer. This is likely to save you significant money because 50% of your dollar goes to ad tech middlemen instead of towards showing ads, when you buy through programmatic channels, compared to buying direct from good publishers.

Further, there are significantly greater risks of fraud, waste, and hidden arbitrage when buying through programmatic channels, all of which reduce the productive part of your dollar down to 1 cent. Don’t believe me? Have a look at some of the things that eat away at your dollar: 2021 End-to-End Supply Chain Transparency Study. Even if you assumed there was no risk of ad fraud, no risk of bad targeting, no risk of hidden waste and arbitrage, you’re still only getting 50 cents of working media. If you cut out the ad tech intermediaries and bought directly from good publishers (yeah, the large ones that humans have heard of, and do visit), you could save yourself half of the cost. In other words, you eliminate the 50% ad tech tax. Who pays 50% tax on anything? Well, you did, and will, if you continue buying through programmatic channels. And don’t let them trick you by saying you’re getting much better prices — like lower CPMs. 

Even if you are paying $3 CPMs in programmatic, versus $30 CPMs in direct buying, you’re buying 10X the quantity in programmatic. So you’re still spending $30 total, except that you are in a much higher risk environment, and you’ve just paid the 50% ad tech tax. You’d stil be better off buying $30 CPM ads from real publishers with real human audiences, and buying less quantity. Your ads will be more likely shown to humans, less likely eaten away by fraud and bots, and might help you drive more business outcomes. 

So What?

If you are turning off bad domains and apps, you will be improving your digital marketing programs, because less of your dollars are flowing to criminals. If you are turning off (b)adtech targeting that was useless anyway, you are saving money and improving your digital marketing programs. You were paying extra for targeting parameters inferred from website visitation patterns, and getting worse outcomes because of it; and consumers didn’t think any of the ads were better targeted or relevant to them anyway. Cut it out. And if you turn off programmatic ad buying and turn on direct ad buying from real publishers with real human audiences, you might surprise yourself with surprising business outcomes. Hint: it was because more of your dollar went to showing ads (instead of into snake oil salesmens’ pockets) and more of your ads were shown to humans. That’s what you call real digital marketing. 

Are you ready to take the next step?