A Sequel Better Than the Original: How a Return to Media Measurement Roots will be Different

Joe Foran
.
April 28, 2022
A Sequel Better Than the Original: How a Return to Media Measurement Roots will be Different

Client driven, in-housed, home-grown

“What has been will be again, what has been done will be done again; there is nothing new under the sun.”  Ecclesiastes 1:9

There have been a lot of articles published over the years on how media measurement has changed.  When people realized that they could string together touchpoints using Google’s DCM ID, they were amazed at the detailed insights they could gather and how frequently they could generate them using Multitouch Attribution (MTA).  “Legacy” techniques like Marketing Mix Modeling (MMM) were eschewed, especially by those who felt that MMM underappreciated the Digital Revolution and overvalued other “legacy” channels like Linear TV.  Pragmatists continued to use MMM (“guys, I still spend a lot on channels that MTA can’t measure”) in conjunction with MTA in Unified approaches, but many were more enamored with the cool new machine learning approaches.  This helped CMOs start to embrace machine learning  and how it could help solve other thorny customer analytics questions, leading them to start hiring their own Data Scientists. Then, growing privacy concerns drove the call to deprecate cookies; suddenly, media agency experts told us that MTA was dead and led marketers back to the venerable MMM.  “What has been will be again”.

While many now ponder “The Return of the MMM”, they miss how this sequel will be different.  Four shifts that occurred during “The Rise of MTA” will forever change how marketers approach media measurement, and so also change how agency partners need to support them.

1. In Housing of Media Buys

During the early days of MMM, clients counted on analytics agencies to be a media-savvy check on their agency partners to ensure they were getting the most for their media dollars.  Since then, there has been a major shift in media buying, with marketers bringing some or all their buying activities in-house. In the 2021 Garner Annual CMO Spend Survey, we see in-housing continue even while budgets have been cut in half on average; this is because clients feel they can execute some media buying better and cheaper.  In housing of media buying means clients no longer need to count on an outside partner to help them with the intricacies of the media landscape; they live with these complexities every day and know them better than outside providers.

2. Expanded Analytics Capabilities

In-housing drove companies to own more of their data; even those who didn’t in-house wanted more positive control over their marketing data and so they stood up cloud environments to democratize marketing data throughout their companies.  These environments enabled machine learning at a scale previously unheard of, to the delight of those Data Scientists mentioned earlier.  The rise of R and Python freed organizations from large investments in proprietary statistical software, but Finance teams started to ask “why should we continue to use outside consultants to run MMM for us if we’re paying all these Data Scientists?  Can’t they do the work?”

3. Disillusionment with Market Solutions

When Unified Measurement became the hot topic, it led vendors to offer large, productized solutions to deliver a box of UM to every company.  These vendors did a great job selling these into the C-Suite; many months and millions of dollars later, organizations were often disappointed with the results. Many abandoned these mammoth solutions; they know they need measurement, but their trust in analytics agencies to consistently deliver on their specific needs is broken.

4. Identity

One way that marketers are trying to maintain MTA is by switching to click-based, identity-driven MTA.  While not as compelling as the original instances of MTA (due to the lack of Impression data), it is powerful in its ability to examine Customer Journeys at an individual level in a privacy-safe way and enable superior activation capabilities.  One key difference in how marketers are implementing this compared to early MTA solutions is that they don’t want to rely on somebody else’s ID ever again; they want to own their private Identity Graph, in just another sign of how capabilities and control have shifted from vendors to marketers.  

Where does this leave vendor/agency partners in this brave new world?  While the phrase “evolve or die” gets used a lot, this is a clear instance where partners need to understand that things have changed, and to continue to deliver the value they need to change their approaches.  No longer can vendors have a single operating model; they must be ready to act as a “player-coach” in some engagements to help their clients stand up their own capabilities while maintaining a “Build and Deliver” capability for clients whose Data Scientists are decisively engaged in problems only they can solve. In other cases, the Co-Creation of custom solutions can drive value for both marketers and their partners.  Only by moving beyond the model of “this is our proprietary, multi-zillion dollar product solution that you just NEED to buy!” will Data Science agencies be able to meet the needs of evolving customer needs. And really, isn’t that the definition of “customer service”?  

There really is nothing new under the sun.

Joe Foran is Vice President, Decision Sciences for Blend360, a next-generation data science & professional services firm that focuses on helping clients develop their own in-house capabilities.

Client driven, in-housed, home-grown

“What has been will be again, what has been done will be done again; there is nothing new under the sun.”  Ecclesiastes 1:9

There have been a lot of articles published over the years on how media measurement has changed.  When people realized that they could string together touchpoints using Google’s DCM ID, they were amazed at the detailed insights they could gather and how frequently they could generate them using Multitouch Attribution (MTA).  “Legacy” techniques like Marketing Mix Modeling (MMM) were eschewed, especially by those who felt that MMM underappreciated the Digital Revolution and overvalued other “legacy” channels like Linear TV.  Pragmatists continued to use MMM (“guys, I still spend a lot on channels that MTA can’t measure”) in conjunction with MTA in Unified approaches, but many were more enamored with the cool new machine learning approaches.  This helped CMOs start to embrace machine learning  and how it could help solve other thorny customer analytics questions, leading them to start hiring their own Data Scientists. Then, growing privacy concerns drove the call to deprecate cookies; suddenly, media agency experts told us that MTA was dead and led marketers back to the venerable MMM.  “What has been will be again”.

While many now ponder “The Return of the MMM”, they miss how this sequel will be different.  Four shifts that occurred during “The Rise of MTA” will forever change how marketers approach media measurement, and so also change how agency partners need to support them.

1. In Housing of Media Buys

During the early days of MMM, clients counted on analytics agencies to be a media-savvy check on their agency partners to ensure they were getting the most for their media dollars.  Since then, there has been a major shift in media buying, with marketers bringing some or all their buying activities in-house. In the 2021 Garner Annual CMO Spend Survey, we see in-housing continue even while budgets have been cut in half on average; this is because clients feel they can execute some media buying better and cheaper.  In housing of media buying means clients no longer need to count on an outside partner to help them with the intricacies of the media landscape; they live with these complexities every day and know them better than outside providers.

2. Expanded Analytics Capabilities

In-housing drove companies to own more of their data; even those who didn’t in-house wanted more positive control over their marketing data and so they stood up cloud environments to democratize marketing data throughout their companies.  These environments enabled machine learning at a scale previously unheard of, to the delight of those Data Scientists mentioned earlier.  The rise of R and Python freed organizations from large investments in proprietary statistical software, but Finance teams started to ask “why should we continue to use outside consultants to run MMM for us if we’re paying all these Data Scientists?  Can’t they do the work?”

3. Disillusionment with Market Solutions

When Unified Measurement became the hot topic, it led vendors to offer large, productized solutions to deliver a box of UM to every company.  These vendors did a great job selling these into the C-Suite; many months and millions of dollars later, organizations were often disappointed with the results. Many abandoned these mammoth solutions; they know they need measurement, but their trust in analytics agencies to consistently deliver on their specific needs is broken.

4. Identity

One way that marketers are trying to maintain MTA is by switching to click-based, identity-driven MTA.  While not as compelling as the original instances of MTA (due to the lack of Impression data), it is powerful in its ability to examine Customer Journeys at an individual level in a privacy-safe way and enable superior activation capabilities.  One key difference in how marketers are implementing this compared to early MTA solutions is that they don’t want to rely on somebody else’s ID ever again; they want to own their private Identity Graph, in just another sign of how capabilities and control have shifted from vendors to marketers.  

Where does this leave vendor/agency partners in this brave new world?  While the phrase “evolve or die” gets used a lot, this is a clear instance where partners need to understand that things have changed, and to continue to deliver the value they need to change their approaches.  No longer can vendors have a single operating model; they must be ready to act as a “player-coach” in some engagements to help their clients stand up their own capabilities while maintaining a “Build and Deliver” capability for clients whose Data Scientists are decisively engaged in problems only they can solve. In other cases, the Co-Creation of custom solutions can drive value for both marketers and their partners.  Only by moving beyond the model of “this is our proprietary, multi-zillion dollar product solution that you just NEED to buy!” will Data Science agencies be able to meet the needs of evolving customer needs. And really, isn’t that the definition of “customer service”?  

There really is nothing new under the sun.

Joe Foran is Vice President, Decision Sciences for Blend360, a next-generation data science & professional services firm that focuses on helping clients develop their own in-house capabilities.

Download your e-book today!