Crumbling cookies broke marketing measurement. But good science and common sense are fixing it.

Recent IAB research suggests that marketers know they need to update their measurement strategy, but 66% are not taking any meaningful action to do so. A new generation of vendors are using solid science to solve this problem. It’s time to have a look at what they have in common and where they need to grow. Most importantly, the new era of marketing measurement is here and it’s time for brands to shake off the wounds of MTA and get back in the game.

After leading multiple IAB Attribution bootcamps, talking to a thousand media leaders and measurement pros, and reading or hearing the consensus of most everyone we’ve met, the message is clear: marketing measurement is broken.

Let’s be clear, we’re not talking about counting clicks or measuring loyalty. We’re talking about the most critical decisions marketers make. Where should the media budget be spent to generate the best possible return? More specifically, what should be the weekly optimizations, the periodic rebalancing, and the annual planning that every brand must do? This is the stuff that drives revenue and profit.

What specifically have we heard?

First, most brands don’t have trusted data about spending performance across their entire media plan. We’ll get into “Why” later in this post, but trust is the key word here, since cross-channel measurement has always been about “best estimate” of performance, not 100% precision.

Second, on those occasions when reasonable performance data is in hand, it’s still nearly impossible to systematically use that data to optimize future periods. Since you can’t re-spend last week’s budget, many believe the big win will be using AI to optimize for the future. The industry wide opportunity here is massive.

The year is 2023, and we are still talking about attribution. However, that should not come as a surprise. Understanding the ROI and optimizing for future opportunities in even a modestly sophisticated media plan has always been a complex problem.

And in 2016, many good companies were many years into solving this problem when privacy jumped to the front of the line and became, almost universally and instantly, the top concern across the technology landscape. The end of reliable identity data was a meteor strike to the marketing measurement industry. Hundreds of “cookies are dead” stories were written. From a measurement and optimization perspective, what many considered to be progress simply vaporized. Vendors needed to start over.

The return of marketing science.

Starting over is always hard. When the MTA vision of holistic measurement of digital media plans fell apart, a new race began.

Leading customers, big platform providers and service providers turned to forms of testing for BOTH point comparisons of media tactics AND cross-checking the deteriorating quality of their MTA models. Platform vendors offered testing solutions inside their walled gardens. Plenty of media professionals gained exposure to tests and learnings. Solid science was applied to many point-comparisons inside media plans.

All good progress! Nevertheless, a holistic view of media performance was clearly missing.

Around 2020, a number of innovators began looking at digital versions of Media Mix Modeling. The sound economic principles of MMM have always been a trusted source of truth for brands with the budgets and patience to adopt them. Vendors began delivering streamlined versions of this proven technique. They promised to deliver with the agility necessary to better align with the needs of digital media buyers for weekly campaign level information.

More good progress!

There is solid science behind both testing and media mix modeling. They are big steps forward and should enable a level of analytic rigor that MTA never delivered. A number of new providers are working to varying degrees to apply science to the problem. Early user feedback says their success will ultimately lie in their ability to:

Deliver Trusted Information. 
In geek speak, brands need to know incrementality and elasticity. In plain English, brands need to know what happened and what might have happened. These are the basic metrics. Trust is key. Everything is an estimate. Trust gives you the confidence to act.
Tackle the Predictive Opportunity. Marketers want positive returns on their investments.  They cannot change past spending levels. Opportunity only exists in the future.  So, while projecting a holistic view of a media plan into the future is a tall task, the power of AI can make a dramatic difference in the quality of your predictions (i.e., your plans) and business outcome.  Executives should pay close attention here.

The path forward

While giving marketing science a bear hug is an important reset for the industry, more innovation is needed for holistic, cross-platform measurement to achieve ubiquity. We challenge industry leaders to deliver in three key areas:

  1. End the “Methodology Wars.”
    Marketing measurement is a complex problem. We can all agree that a single approach to estimating performance will never be the ideal answer for every scenario. A better future lies in integrating the best available techniques and data, not defending a particular methodology. More information leads to better answers. This shift will be difficult for vendors deeply invested in a single technique. However, it is foundational to establish trust.
  2. SaaS Please.
    The big MMM industry got us off on the wrong foot. Cross-platform measurement providers all carry some managed services legacy; every AI vendor does. Vendors should shift focus toward enabling a new generation of data-savvy growth marketers in brands and agencies. It’s time to replace PowerPoints and spreadsheets with software. The rest of the world adopted user-driven SaaS solutions long ago. We can too!
  3. Embrace the data scientists.
    A growing number of marketing teams now have internal data science capabilities, and this will continue to grow. Keeping them in the dark will only serve to inhibit overall marketing excellence. Data Scientists need transparency into their company’s marketing models to do their best work; they need input to control when and how new data and models are deployed. Some may even leverage their unique understanding of their own business to build new models as inputs to a comprehensive measurement system.

The embrace of science is a foundation upon which to build trusted relationships among business leaders about marketing performance, ROI and investment decision making. This path forward has the potential to make sophisticated marketing measurement attainable for brands in a way our industry has yet to do. Taken together, we have a real opportunity to retire “last click/last touch” measurement and move the industry to a framework for marketing investment decisions that will stand up to CFO-scrutiny.

And that would truly be a better place.

Posted in: