Measurement Changes Extend Beyond The Cookie

I recently shared some of my thoughts around what paid media professionals and their teams can expect to change over the next two years. While I covered four different topics at a high level, I’d now like to zoom in on one area in particular; measurement.

If you keep up with any industry news sources, you’ve undoubtedly seen multiple articles discussing the death of the third party cookie. I won’t harp on alternative solutions too much here, but instead focus on how the ripple effects of this change will impact the day-to-day life of anyone managing paid media campaigns. The days of creating GTM tags will soon be behind us, and instead we’ll be focusing on measuring business impact and designing lift tests. I’ve broken my predictions out into four distinct behavior changes.

Change One: Understanding Test Designs

Testing isn’t a foreign concept to anyone working in digital marketing, it’s a staple in the vast majority of roles. However, testing takes on many different forms. Tools like Google Ads’s experiments or Facebook’s A/B testing capabilities will be familiar to many, but the loss of third-party cookies on Google Chrome will limit (or even prevent) the use of some of today’s in-platform testing features. This means that setting up an experiment won’t be as simple as making a few clicks and letting the ad platform do the rest.

While the streamlined process of creating a test might be disappearing, I believe that advertisers are actually gaining much more by moving away from this approach. When implementing solutions like incrementality tests, organizations can get a better read on the true business value of their paid media program. This eliminates concerns over different attribution models across platforms, as well as advertisers reporting on “conversions” instead of revenue.

Given all of the benefits that incrementality testing has to offer, it will be important for paid media professionals to understand how to design these experiments. Each test needs to be thoughtfully set up, and everyone will have to lean on the good old fashioned scientific method from fourth grade.

What might this look like?

1. Define your hypothesis

   1. Ex: “I believe that TikTok ads will generate incremental revenue for my business.”

2. Identify your test variables

   1. Ex: Users exposed to TikTok ads vs those that are not.

3. Set your test parameters

   1. Ex: Philadelphia will serve as the control market where no TikTok ads will be run, and that location will be compared to Dallas where users will be exposed to the campaign.

   2. Pro Tip: Become more comfortable having conversations involving data analysis and statistics. It can be helpful to partner with members of an organization’s analytics team to determine how much budget we need to be secured for a test, and how long it should run for in order to generate statistically significant results.

4. Execute the test and make minimal changes to any of the campaigns involved

5. Analyze the results of the test, and iterate based on the insights that are captured

The process and design outlined above is about as simple as it gets, so understanding it thoroughly is helpful when creating multi-cell tests for more complex programs. This exercise will become routine for paid media professionals everywhere over the next two years as organizations prepare to update processes to account for the death of the third party cookie.

Change Two: A Mindset Shift in Time Horizons

If you’ve ever been a part of business development discussions on the agency side, you’ve probably heard a statement like:

“We’re looking to move on from our current agency because they’re not making changes frequently enough. We want someone to make daily optimizations to our campaigns.”

Back in a time where manual bidding and extreme segmentation was still standard practice, daily optimizations made sense. Now, a cadence of that frequency would be considered borderline reckless. Automation aside, making changes to a campaign too frequently will limit any learnings (and therefore progress) when attempting to analyze performance. Taking a look at three different perspectives will showcase why less frequent, more thoughtful, optimizations will provide a greater benefit to a paid media program.

• The first perspective circles back to touch on incrementality testing. If you’ve set up a test to run for four weeks, you don’t want to be making any changes to a campaign’s bidding strategy on week two. That would prevent you from understanding if bid strategy A or B was responsible for driving the results you saw at the end of the four week test.

• Another measurement framework that would benefit from less frequent optimizations would be media mix modeling. Even some of the most advanced models can’t report on results more frequently than once a week. Naturally, report cadence will also be dependent on the volume of data the model has to work with, so only some of the biggest brands will have access to data at that speed. Therefore, even some of the world’s biggest advertisers shouldn’t be making changes to campaigns more than once a week.

• Lastly, even if you’ve determined that in-platform performance directionally aligns with business performance, you’ll still want to shy away from daily optimizations. Between the time it takes for algorithms to learn from those adjustments, and any sort of data delay associated with conversion APIs, platform data takes at least a few days to settle. That means there’s a good chance you’re making decisions based on incomplete data if you make optimizations too frequently.

Paid media used to be a world of instant gratification where you’d be able to watch the clicks and conversions stack up within a few hours of launching your campaign. However, that world is changing. Running campaigns on a platform like Google Ads isn’t a short-term game anymore. It should be considered an iterative process where data is evaluated from multiple weeks of performance before making an optimization.

Change Three: Focus on Business Impact Over “Conversions”

This is something more advanced programs have already been doing, but there are still teams who report and optimize on platform conversions instead of revenue (or any other bottom line metric important to a business). Losing the ability to track conversions through a cookie-based tag will give any remaining advertisers a helpful nudge in the right direction.

The loss of cookie-based events will push advertisers in one of two directions by default:

1. Regressing to minimal measurement on performance with little-to-no understanding of the impact paid media has on business goals

2. Breaking down silos between analytics and paid media to understand how changes in ad campaigns influence business KPIs like revenue

While this concept is fairly straightforward, I think this change will be a manufactured inflection point for any program that isn’t already headed in one of these two directions. This concept also provides a nice segue into my next and final measurement related change…

Change Four: Breaking Down Departmental Silos

In order to measure the business impact of paid media, the marketing team will need access to the relevant data. This means they’ll need to work more closely with teams that handle data engineering as well as finance/accounting.

In order for the marketing team to report on the performance of their efforts, they’ll need access to relevant business data. This data is often structured and owned by data engineers, so developing workstreams in which these two teams collaborate is critical. It’s also important to create an environment where the collaboration and communication between these teams is structured and ongoing. They should have regular check-ins and both be working to the same goal. Validating the effectiveness of any marketing channel can be difficult without this level of alignment.

In addition to working with any team members handling business data, advertisers should also become more comfortable speaking the language of their counterparts in finance. Once marketers can relay the value of paid media in terms of ROI, it allows finance teams to more easily understand that marketing is an investment and not a cost. From here marketers can more easily model out a business case explaining why an investment in that next great martech tool will help to take the business to the next level.

There are undoubtedly other teams and departments that marketing will have to work more closely with, but those will all be more specific to individual organizations.

Wrapping Up

There is a lot of change on the horizon in the paid media world, and if you’ve made it to this point I appreciate you taking the time to read my perspective on what this change looks like from a measurement perspective! To quickly recap, the way advertisers will operate on a day-to-day basis will start shifting in the areas related to:

• Designing tests

• Setting timeline expectations

• Communicating business value

• Working with other departments

Digital marketing is a dynamic field that is constantly evolving, but the next two years will shift many of the core principles that anchor the responsibilities of these professionals. This post is one in a series on my analysis of the future (we’ll see how accurate my predictions are in the end…), and next up I’ll be focusing on changes related to automation.

Have questions, considerations, or critiques? I’d love to hear them! If you’re reading this via email, just hit respond. Otherwise, you can find me on LinkedIn and X (Twitter).