ARF Announces a Scientific Solution to Complete Advertising ROI Measurement

By In Terms of ROI Archives
Cover image for  article: ARF Announces a Scientific Solution to Complete Advertising ROI Measurement

RCT21 is a factor in the equation of the future. Random Control Trials for the 21st Century is an initiative of ARF, the Advertising Research Foundation. Problem to solve: lack of certainty implicit in ROI modeling methods.

There have been, in my view, two classes of ROI modeling methods, one typified by TRA, NC Solutions, and especially 605, where there is vast data at the household level and a high degree of representativeness, and the less expensive category of MMM/MTA. The massive passive household deterministic data make for the highest confidence levels – but none of these observational methods delivers certainty, none delivers causality. So better companies have always proposed that RCTs be used to confirm results before making major media or creative decisions. 605 and NC Solutions have been doing RCTs for years.

This is not to be dismissive of attribution. In the main, all ROI practitioners have raised the understanding of the sales effects of brand advertising from approximately zero when I joined the business, to maybe 50% today. We are all in their debt.

Yet Advertiser Perceptions reports that only one in nine advertisers is extremely satisfied with ad measurement, and ROI modeling is the ad measurement type most criticized.

We all know the reasons why, including:

"Black boxes which disagree with one another, so who knows which is right?"

"Attribution says I made the company a lot of money, but the company somehow can't seem to find a trace of it in their coffers."

"Were these incremental sales, or are media being credited with some sales that would have been made anyway?" Former Chief Data Officer at Havas Nathan Woodman writes in ADWEEK that some media purposely target known buyers so as to appear to have contributed incremental sales, in full knowledge of the fact that those sales come from existing brand buyers.

Lacking confidence in ROI findings leads marketers to make small shifts even when they feel larger shifts are needed.

ARF saw this as a priority problem needing real solution, and that solution is the use of scientific experimental design – Random Control Trials (RCT) across platforms, initially TV and digital, later across all marketing and media stimuli. By setting up an experiment, incrementality is intrinsic, there is only one variable different between the test and control group, and so causality can be determined with certainty.

RCT21 is the name of the ARF initiative which brings together 605, Central Control, and Bill Harvey Consulting along with the ARF itself, and participating advertisers, to demonstrate the Proof Of Concept. The aim: one more tool in the marketer's toolkit, when certainty is desired or essential, and a "truth set" that attribution researchers can use to calibrate their findings, making all of attribution more accurate and actionable.

The press release is here:

There will also be a special ARF webinar open to Members and non-Members on July 23 at 3PM Eastern Time, in which Andy Fisher, Head of Merkury for Advanced TV and Chief Analytics Officer at Dentsu Merkle, and Vadim Tsemekhman, Director of Product Management at Walmart Labs, join with RCT21 team members in a panel discussion delving more deeply into RCT21. For free registration: Click here

As we traverse COVID-19 into the new world, we now have bedrock ground truth to rely upon. Whenever we need certainty, we can have it. The $2 trillion in global marketing dollars needs it today more than ever before. (That number includes eCom, martech and data, user experience, promotion, advertising, sponsorship and brand content. Thanks to Chick Foxgrover at 4As for this number.)

What Makes Science Work?

Despite the current disregard for science in high places, we all must give science its due: it is the greatest thing the human race has created, even in light of the awe-inspiring art and all of our other achievements.

Science knows its Creator, whether you believe that Creator to be random chance or something else, that is an amazing accomplishment, even though as Arthur C. Clarke and others have pointed out, what we know is a trivial fraction of the totality of what is to be known. That just means it will take time, but does not denigrate the victory of having a process that delves into knowledge of who oneself is and where did I come from and what is this thing we are all a part of?

So, how did science get there? What did it do, to become so appropriately exalted as to supersede religion (although one's feelings bring one back to spirituality)?

Just two things. Very difficult things, especially when you consider the apish creatures we were when we first started to think this way.

  1. Inner observant articulation of hypotheses derived from outer observant inductive reasoning.
  2. Testing of hypotheses in experiments where everything is the same for two treatments, except the one difference, that would confirm with utter certainty that the cause of any difference in outcome was the one pre-identified suspected cause.

Those are the two quintessentially brilliant tricks of science.

In marketing, the first trick is used all the time by every practitioner.

Today, the second trick is used by the walled gardens for their own media, is desperately needed by the marketers to be able to use via disinterested and qualified third parties across all media, and ultimately across all marketing stimuli, but is only being used regularly by a very small percentage of marketers – typically DTCs, and it is part and parcel of their meteoric rise.

That is a long-winded way of saying that what ARF is doing with RCT21 is of major significance, and could be a watershed moment.

Click the social buttons to share this story with colleagues and friends.
The opinions expressed here are the author's views and do not necessarily represent the views of

Copyright ©2023 MediaVillage, Inc. All rights reserved. By using this site you agree to the Terms of Use and Privacy Policy.