Analytics & Data

Cookieless Browsers Will Upend Ad Tracking

Trends toward greater consumer privacy may force ecommerce marketers to change methods of measuring performance, including turning to “marketing mix modeling,” a 70-year-old technique.

Most digital marketers depend on some form of multi-touch attribution (MTA) modeling to measure the effectiveness of their various promotions, such as pay-per-click search ads, display ads, ads in streaming video, or digital out-of-home promotions that combine outdoor advertising with modern technology.

In each of these cases, MTA requires some means of identifying a person across various devices, web browsers, and ad networks (e.g., Google, Facebook).

“Digital advertising provides the ability to reach people wherever they are with timely and relevant messages. To understand the effectiveness of these ads, advertisers have come to expect a direct and complete view of the customer journey, from awareness through conversion,” wrote Philip McDonnell, director of product management at Google, in an August 2020 Think with Google article.

“Successful online measurement has been heavily reliant on cookies that log useful information about what happens after a person has clicked an ad,” McDonnell said. “However, whether due to cookie restrictions in browsers or blind spots from cross-device shopping, there are increasing scenarios where it’s no longer possible to observe whether a conversion has taken place. Increased privacy regulation has also imposed strict guidelines for data collection by region.”


In 2020, some popular web browsers — Microsoft Edge, Apple Safari, Brave — have eliminated or curtailed third-party cookies.

Furthermore, all modern web browsers will likely be cookieless — they will not allow tracking cookies at all — in the next few years.

The European Union’s General Data Protection Regulation and the California Consumer Protection Act are examples of the restrictive privacy regulation McDonnell mentioned, but they are just one factor driving the move to ban cookies.

Apple Inc., which announced changes to its Identifier for Advertisers tracking feature on Apple devices, has been running commercials promoting privacy on iOS. In the ads, folks who use other mobile devices are seen calling out personal information in public, while Apple users’ info is kept quiet.

Screenshot of a YouTube video frame on an Apple commercial

In the Apple privacy commercial, a woman is seen shouting personal information through a bullhorn.

Thus the drive toward non-tracking browsers and devices comes from a combination of consumer demand, regulatory restriction, and competitive positioning.

In short, the future of digital advertising is cookieless. Measuring ad performance will soon be significantly different.

Marketing Mix Modeling

Advertising’s cookieless future does not, however, mean that all performance-based ads will be a guessing game. What will change is how marketers measure performance, especially for large and complex organizations.

One technique that is garnering attention is, again, marketing mix modeling.

MMM is an analytical, top-down approach to understanding how a company’s promotional efforts, the general economy, and even competitors impact sales.

Neil H. Borden, who was a professor of advertising at the Harvard Business School, first used the term “marketing mix” in 1949. So the approach has proven its usefulness over several decades.

Even before it became clear that privacy trends would restrict some forms of tracking, savvy marketers had already been using MMM, often combined with multi-touch attribution, to obtain a much better picture of a promotion’s impact and effectiveness, as well as improved data-driven insights.

For example, MMM can be good at identifying optimal ad frequency, also known as frequency capping. An advertiser might find his company could spend, say, 35 percent less on YouTube ads while getting 95 percent of the sales.

MMM Takes Time

However, MMM takes time. The models depend on historical sales and ad-spend info. Thus, a business dependent on multi-touch attribution should start accumulating MMM data now before browsers eliminate tracking cookies altogether.

Moreover, MMM tends to be custom. There are many good MMM providers, but building the models is not as simple as signing up for Google Analytics and learning to read the reports.

Finally, marketers who are not familiar with MMM will presumably need time to understand the technique and recognize the insights it provides.

Other Models Will Emerge

Marketing mix modeling and multi-touch attribution are by no means the only options (after tracking cookies go away) for analyzing advertising and marketing performance. For example, McDonnell’s Think with Google article, mentioned above, describes “conversion modeling.” And some marketers believe they can develop their own customer identity graphs that will ultimately provide detailed attribution.

Armando Roggio
Armando Roggio
Bio   •   RSS Feed