Marketers, Don’t Worry About Losing 3P Cookies and Microtargeting. They Were Useless Anyway


For the last decade, marketers have been sold the idea that microtargeting would help them improve digital marketing. I realize that it will be hard for you to accept that it didn’t work — at. all. The intense focus on “the right ad to the right person at the right time” had led marketers astray from… marketing. Yeah, you heard that right. By over-targeting, advertisers forgot to do basic branding and awareness, which harms long-term growth; if no one knows about your brand or simply forgot about it, they certainly won’t be buying from you. And you can’t squeeze more blood from a turnip — i.e. no matter how many more super relevant ads you throw at someone, they simply can’t consume or use more of your product beyond a certain point. By focusing all your ads and digital budgets on the 1% that have purchased from you before, you are neglecting the other 99% that have not purchased from you yet, and therefore missing out on the vast majority of potential sales.

I’m sure you still don’t believe me, so let me dig in some more.

Microtargeting made adtech companies rich, cost you more money, and reduced your business outcomes

Where do you think ad tech companies got the targeting parameters they sold to you to use in programmatic campaigns? Right. They made it all up. When users use Gmail, Facebook, Snapchat, etc. they are logged in to those services. But when they visit websites, they are usually not logged in. (Despite the growth in digital subscriptions, most users are still not logged into sites when they visit to read content.) So how do ad tech companies track them? Right, they put lots of third party ad tech companies’ code on content sites that users visit. Each of those third party ad tech companies set third party cookies in the users’ browsers in an attempt to track their website visitation across sites. Why do they do “cross-site tracking?” They do that because they believe they can derive “who they are” and “what they like” based on website visitation patterns. In the early days, this may have been fine and for some parameters it may even have worked. For example, if a non-logged in user visited Sports Illustrated, Maxim, ESPN, beard trimming product site, etc. you could possibly deduce the user was male. If a non-logged in user visited Victoria’s Secret, Cosmopolitan, and a feminine hygiene product site, you could possibly deduce the user was female. But what can you deduce if the user visited a news site like NYTimes or an ecommerce site like Walmart? Right, it’s not clear what you can infer.

So ad tech companies are literally making up the targeting parameters and adding anonymous cookies into audience segments and selling those targeting parameters and audience segments to advertisers like you for use in your programmatic campaigns. And it’s been ridiculous for the last ten years how much (crap) they sold you and how much (crap) you bought from them. Do you remember the example of data broker Lotame purging 400 million profiles because even they could see the profiles were entirely fake — e.g. created by bots visiting various websites to build up desirable profiles. Advertisers were paying more to target those bots, based on their website visitation patterns. Further, academic studies [PDF] have shown that even for one targeting parameter, gender, this kind of estimation was less accurate than random. When derived from website visitation patterns, gender was only accurate 42% of the time, which is a worse outcome than if advertisers just did spray and pray.” At least they would have gotten gender right about 50% of the time. When two parameters were considered — gender + age — the accuracy drops to 24%.

Microtargeting with dozens of parameters or paying for audience segments is useless because they are all inferred and therefore highly inaccurate.

The illusion of success – higher click through rates

The next question you might be thinking is how are we getting higher click through rates, then? Doesn’t that mean the campaigns were performing better because the targeting was better? Nope, it was just bots clicking on your ads. If bots didn’t click anything, those campaigns would appear to suck and campaign optimization would have shifted budget away from them. But, due to bot activity, programmatic campaigns always appear to be performing better because advertisers and algorithms were using clicks and click through rates as the indicator of performance. Fraudsters using bot traffic simply tuned the click through rates to be just higher than real click rates on real publishers’ sites; and that was enough to fool the optimization algorithms to keep allocating more budget to programmatic channels or bidding higher on fake sites. It had literally nothing to do with better targeted ads or humans clicking on ads because they were more relevant. It was all bot activity.

Finally, consider your own experience as a consumer. How many relevant ads do you recall seeing over the course of a day, a week, or a month? And how many programmatic ads did you deliberately click on (not accidental clicks due to “fat thumb”)? Right, not many, if any. Experimental results also corroborate that despite all the targeting, the number of relevant ads that users see remains low to zero [see: How many relevant ads do we see each day?]. So let me re-iterate — you were sold snake-oil by the ad tech companies. They got rich; their investors got rich. You made them rich by buying lots of snake oil. Time to stop wasting your money and get back to real digital marketing. How? Simple.

Think of targeting as the avoidance of waste

Instead of buying hundreds of targeting parameters or audience segments from adtech companies, go back to the basics of marketing. If you are selling a beard trimming product, you can avoid wasted ad spend by targeting away from women — i.e. targeting males. You can also avoid targeting under 13, because young boys don’t need beard trimmers. With just a few basic targeting parameters, you can avoid the most obvious wasted ad spend. Also, in doing so, you will be getting ads in front of users who have not purchased from you before. That segment is far larger than the one of users that have purchased from you before. The corollary to this is to NOT overdo retargeting and remarketing. Of course you can do some retargeting (people who visited your site) and remarketing (people who purchased from you), but those people already know about your site or have already purchased from you. Don’t forget about showing ads to new potential customers who don’t know about you yet.

If you stuck with me this far, does it now make sense why I said in the title to not worry about the loss of third party (3P) cookies and microtargeting? Third party cookies were used by ad tech companies for cross-site tracking. Cross-site tracking was used to harvest website visitation patterns of not-logged-in users to infer who they were and what they liked. Targeting parameters and audience segments were created from these inferences and sold to advertisers for use in programmatic campaigns. This cost advertisers like you more money and made the ad tech companies rich. But the targeting parameters and audience segments were not accurate or were outright faked due to bot activity. The higher click rates were not due to more relevant ads; but were also due to bot activity. Academic studies and everyday experience tells you that despite all the money paid for microtargeting, the ads are not any more relevant.

Time to move on, and get back to good, basic digital marketing, that drives real business outcomes. You’ve been overdoing it — microtargeting customers that have already visited your site or bought your product — at the expense of showing ads to potential new customers who don’t know about you yet. Give that a try and let me know what you observe in terms of business outcomes.


Please enter your comment!
Please enter your name here