Like most digital solutions in the marketing space, personalization promises a lot. By gathering data, we can create an identity graph that allows for one-on-one marketing. Sure, but it comes up short often.

Not all personalization is suspect. Personalizing when marketing to existing customers is a no brainer. If an insurance customer bought a 30 year old house, they will need to buy an HVAC unit soon. If a individual bought baby formula, he or she is likely interested in strollers and carriers. There is a clear journey in what the customer needs will be soon. For retailers with much first party data, personalized recommendations result in huge ROIs. Personalizing for a potential buyer who is in the funnel already is table stakes. Featuring products from abandoned carts, targeted promotions and discounts, and audience-specific content is all doable today.

If we are not Amazon, what should our expectations be for personalized marketing? Let’s look at the biggest personalization engine and data management platform: Google. Google collects data from search, platforms like Google Analytics, Google Maps, and YouTube. Chrome allows Google to track everything a user sees and does.

Using clickthrough rates as a proxy, ads in search pages result between 1.35% and 2.65% CTR. Banner ads are even worse: between 0.14% and 0.84%. This is with Google knowing every search we do, every website we visit, every piece of content we view. Google knows our intent and moments of truth.

So it’s not surprising when the WSJ reports

What is the impact of Apple requiring user opt-in for tracking? How apps requiring privacy labels? Google is following in Apple’s footsteps with app privacy labels in 2022. How much will personalization hurt without user locations, Bluetooth connections, and device identifiers?1

Google’s new FLoC technology is a replacement for tracking with cookies. Targeted ads will have 33,872 cohorts. Of course marketers will target smaller number of cohorts. But the industry underestimates difficulties in communicating with large number of small groups. DMPs like Oracle’s BlueKai and DAM platforms like monday.com will do some of the leg work. Still, the practice of creating many buyer personas for personalization becomes overwhelming.

So far we have been talking about personalization in ads for direct marketing. Personalization engines like Demand Yield will become less expensive and used more. This saturation will lead to less effective campaigns.

The point of personalized marketing is to deliver individualized content and products. At the end of the day, data is only going to take you so far. The true differentiator will be finding more creative ways to personalize and engage. For example, IKEA helps consumers visualize furniture in their homes. Coke has been offering personalized labels on bottles for years now. All these are buyers are providing the personalization data directly. This is zero party data.


  1. Buyers don’t feel sympathy when there is no transparency in how the industry makes money. If users are opting out when given the choice up-front, the revenue model needs revisiting. But that’s a much bigger question for another day. [return]