Getting Serious About Measuring Media Impact

Marketers, driven by demands to demonstrate true incremental impact, will increasingly look to more statistically rigorous techniques. These include media mix modeling and randomized controlled trials, which can help determine the channels truly delivering value and the channels simply coming into incidental contact with customers as they move along their customer journey.

This post is part of our 2023 Data Transformation Outlook series

Hot Take on Marketers’ Hot Seat

The economy will be bumpy in 2023, so the screws will tighten on the marketing team to demonstrate how their media spend impacts the top line. Here’s what will no longer cut it in marketing: Media team slide decks reporting “below benchmark CPMs” and “above benchmark click-through rates” delivered at the same time that everyone hand-wrings the impending death of third-party cookies that’s causing their attribution models to undercount the value they’re delivering.

The Complete View of Customer’s Interactions: A Siren Song

When digital advertising emerged in the early 2000s, companies rushed to embrace it. That was the right call: marketers need to put their messages in front of their prospects’ eyeballs, and their prospects’ eyeballs were (and still are) spending more and more time looking at screens. The bonus of that shift—or, so it was thought—was user-level visibility into their customers’ journeys and what that might mean: A complete, rich dataset would finally deliver “the truth” about the value of each channel and each hyper-detailed tactic within each channel!

Unfortunately, this misunderstanding became Accepted Wisdom: Maybe a last click model wasn’t “accurate,” but then, surely, a linear attribution model…or a time decay model…or a data-driven model would yield “the truth.” Of course, the dataset was never particularly complete—tracking a user across different interactions, across different devices, over time, and from online to offline (and vice versa) has always resulted in broad, deep, unknowable holes in the data.

And, those holes have gotten exponentially larger. Bots are on the rise (nefariously arbitraging brands’ media dollars into their bank accounts), privacy regulations are drastically curtailing when and how brands can track user behavior, and privacy changes in browsers and operating systems are further cutting off that visibility.

Today’s Attribution Problem

But, that’s not the main challenge! Consider the branded paid search conundrum: “How many of the users who are coming to our site from clicks on our search ads when someone searches for our brand would have come anyway by clicking on an organic search result if our paid search ad wasn’t there?”

This problem exists across all channels. Media targets “users who would be most likely to buy their product or service.” Unfortunately, that media will invariably target some users who would have bought anyway. This is simply a reality, but the more marketers (and their finance counterparts) realize this, the more they will realize that those “would have bought anyway” interactions are not delivering incremental value to the brand, and, therefore, should not get credit for delivering value.

The Antidote to the 360-Degree Fallacy

Media mix modeling and randomized controlled trials are two techniques grounded in decades of research and refinement in fields extending well beyond marketing. They both estimate the incremental impact of individual channels in very different ways.

The Importance of Showing Incremental Impact

In a scenario where branded paid search was 100% cannibalizing (unpaid) organic search, these techniques (if properly designed and executed) would show that the incremental impact—the actual ROI—of branded paid search was effectively…zero. (Plus or minus—both techniques include measures of uncertainty of their estimates, which is a good thing! Marketing and humans are messy, and statistics provides greater clarity—not omniscience.)
Measuring Media Impact 1

The Importance of Online and Offline Tracking

As a bonus, both media mix models and randomized controlled trials have no reliance on user-level tracking and can easily factor in both online and offline media and results!

Better Measurement of Media Investments is a Must

Increasingly, marketers are seeing so much junk and noise in their ecosystems. As they face economic pressures, they will start to demand better measurement of their key media investments. This means accepting that pixels and cookies and “traditional” multi-touch attribution are all increasingly not what they were (never) going to be in the first place!

Addressing the Analytics Fraud Problem

Composable CDPs vs. Suites

Search Discovery experts love digging into media measurement challenges. Let us help you sort out yours.

Related Posts

Join the Conversation

Check out Kelly Wortham’s Optimization based YouTube channel: Test & Learn Community.

Search Discovery
Education Community

Join Search Discovery’s new education community and keep up with the latest tools, technologies, and trends in analytics.




Scroll to Top


Catch the latest industry trends we’re watching and get new insights from our thought leaders delivered directly to your inbox each month.