top of page
  • Writer's pictureTPA Digital

Will digital marketing measurement die alongside third-party cookies?

by Lloyd Greenfield, Strategy Director at TPA Digital

TLDR: Digital marketing measurement will not die but marketers cannot expect the same level of granularity that they have been used to yet.

The restriction of user tracking caused by the blocking of third-party cookies will have a significant impact on the ability to measure marketing performance in real-time. However, all is not lost, not all measurement techniques require the mass tracking of users. In this blog, we explore the measurement techniques available to marketers and some of the solutions proposed that address privacy-first measurement.

How do marketers measure the impact of their marketing today?

The impact of digital marketing on marketing objectives is measured using several tools and techniques. The technique chosen is specific to the outcome being measured and the data available. Some measurement techniques rely on tracking user actions to attribute conversions where others measure the macro effect on KPIs such as sales.

Below I have outlined the four primary marketing measurement techniques and their reliance on user tracking.

Source: IAB measurement toolkit

Brand studies are a collection of tools used to measure brand metrics that cover awareness, familiarity, favourability, consideration and intent. They can also cover claimed behaviours and attitudes. While brand studies can be used for a wide range of use cases, they are most effective when used longitudinally to provide quantitative evidence of the impact of longer-term brand activities.

Brand studies typically survey a sample of consumers across the lifecycle of a campaign so do not need to track user actions on mass. They do not rely on third-party cookies for this reason.

Econometrics is a set of statistical tools that aim to quantify the relationship between cause and effect in economic data. In marketing, this takes the form of Marketing Mix Modelling (MMM) which predicts how all advertising activity (e.g. TV, print, out of home, online video, social media, and search) translates into incremental sales.

It measures macro factors such as channel spend, base sales, weather, and interest rates rather than measuring the specific actions taken by users. For this reason, it also doesn’t rely on third-party cookies.

Controlled experiments randomly assign a group of people to a test or control group to observe and quantify the impact of a change in media over a defined period. The test group is exposed to a change in media (e.g. your new display advert) whilst the control group sees no change (ideally users are shown a ‘ghost’ ad that presents a relevant competitive baseline ad).

Controlled experiments can be conducted in several ways. Some advertising platforms split the delivery of ads to two groups as part of an A/B test or run ghost ads to one group. With this method, the ability to track the user is essential ad heavily relies on user IDs such as third-party cookies.

Advertisers can also use location to split their audience without the need to identify specific users. For example, measuring the uplift in regional sales on both control and exposed locations.

Attribution modelling is a technique that evaluates how different touchpoints contribute to a sale or action by assigning credit based on their level of involvement. It measures the micro growth drivers such as strategies or audiences that are contributing towards a goal. While statical models can be used to analyse user touchpoints across advertising, most advertisers use the last touch model as their model of choice.

It heavily relies on user tracking, in most cases third-party cookies, to analyse the converting and non-converting paths taken by users before assigning value to each touchpoint. Without access to user data, attribution models cannot attribute value at a user level.

What is the best marketing measuring technique if advertisers cannot track users?

There is not one single technique that measures marketing effectively. Marketers typically use a combination of them to get a well-rounded view of macro and micro growth drivers. Brand Studies and Econometrics are used for macro, high-level measurement while controlled experiments and attribution are used to understand micro growth drivers, often at a tactical level.

It is evident that the techniques used to measure micro growth drivers, controlled experiments, and attribution modelling, are the most impacted by the availability of third-party cookies due to their need to identify specific users.

What are the solutions to micro measurement without third-party cookies?

Some new solutions in a privacy-first ecosystem aim to replace the use of third-party cookies, namely for attribution and controlled experiment measurement solutions. Third-party cookies were not purpose-built for digital marketing measurement, so solutions built on them have never provided a very clear understanding of the growth drivers within marketing.

The new solutions outlined below do not rely on third-party cookies but also will not provide a crystal-clear understanding of marketing performance. To achieve this, solutions require user-level tracking on mass which has become an unrealistic expectation in a privacy-first marketing ecosystem.

Firstly, browsers are developing initiatives that enable privacy-safe attribution. For example, the Event Conversion Measurement API, from Chrome’s Privacy Sandbox, enables the correlation of an event on a publisher’s website with a subsequent conversion on an advertiser site without involving mechanisms that can be used to recognise a user across sites.

These solutions are purpose-built for advertising measurement and make it difficult to identify users across domains. The browser becomes the solution that attributes advertising performance and reports back to advertising technologies instead of the technologies tracking the users directly.

Source: A more private way to measure ad conversions, the Event Conversion Measurement API (

It is important to note that these solutions only support click-based attribution in their current iterations. Some API solutions such as SKAdNetwork have been criticised for the delay in reporting and for not providing the depth of insight that marketers have become accustomed to. These solutions are in continuous development so while we may not see post-impression attribution for some time, I believe we can expect more depth in reporting to be added soon.

Questions remain on the specific model that is being used to attribute conversions and how open-web marketing can be measured against marketing within closed-ecosystems.

Additionally, unified ID solutions aim to replace third-party cookies with a shared, persistent ID. They typically have access to persistent user data, such as email addresses, and are integrated with several technology providers to solve targeting use cases. While they do not offer attribution and controlled experiment solutions themselves, they help facilitate them by allowing existing technologies to identify users more persistently. By identifying more users persistently, advertisers can begin to analyse user behaviour after the exposure of ads and segment users into test and control groups.

While these solutions have the potential to replace third-party cookies, these solutions will not be able to identify all users across the internet. The technologies using these solutions will also still require consent from users to collect and process their data. As a result, advertisers cannot expect a 100% accurate view of digital marketing measurement.

What should advertisers do next?

The first thing advertisers should do is build a list of requirements for digital measurement if they have not already. These could be the ability to measure performance in specific channels, measure against specific KPIs and the speed of reporting.

Following this, they should look inward to measure their reliance on user-level marketing measurement. What attribution model is being used today? How much activity does it track? Is it fit-for-purpose in a privacy-first world?

Finally, speak to your technology providers to assess how they are adapting to privacy challenges and their involvement with industry initiatives such as Privacy Sandbox. Ensure your business is futureproofed for measurement in a privacy-first ecosystem, and whilst there may be some unknowns, there are plenty marketers can do to ready themselves for changes as highlighted above. Alternatively, you can contact us to learn more!

bottom of page