RCT21, which stands for Randomized Control Testing for the 21st Century, is a collaborative research initiative designed to demonstrate methods for applying randomized control testing (RCT) to cross-platform advertising impact analysis.
The initiative is a proof-of-concept study that will apply experimentation methods to measure the incremental ROI of large ad campaigns run across multiple media channels at once, including addressable linear TV and multiple major digital media platforms.
The research, overseen by the ARF and peer-reviewed by a Steering Committee of industry and academic experts, is being executed by three collaborating companies: 605, Central Control and Bill Harvey Consulting.
What is the objective of the initiative?
The initiative looks to establish best practices for applying randomized control testing (RCT) methods to a variety of different cross-platform advertising campaigns.
What is randomized control testing?
Randomized control testing is the “gold standard” scientific method for establishing the link between cause and effect. RCT seeks to eliminate sources of bias in analysis by comparing outcomes for a strictly-comparable exposed group relative to its identical non-exposed group.
Why is this important?
While RCT techniques are already used within the “walled gardens” of large digital platforms, this study will demonstrate their application to the more complex environment of cross-platform campaigns involving both digital and TV.
The study aims to demonstrate how such methods can complement and enhance current observational approaches such as marketing mix modeling and multi-touch attribution. In the process, the ARF hopes to make it possible for advertisers to better pinpoint which marketing elements across all platforms are best contributing to ROI.
Why did the ARF choose these three firms for this proof of concept study?
605 has been conducting cross-platform RCTs since 2012 when it was the Advanced Advertising Department of Cablevision. That work continues today with ongoing large-scale cross-platform RCTs. 605 also brings access to a large population of addressable TV households to this initiative.
Central Control’s Rick Bruner has been conducting digital RCTs for years at Google, Viant, and Marketing Evolution.
Bill Harvey has conducted TV and radio RCTs, and matched market trials, for decades, and conducted some of the first RCTs in digital in the late 90s and into the 21st Century.
Will advertisers be participating in the study?
Yes. A select number of national advertisers are also being recruited to join in the project.
How can an advertiser participate and is there a cost?
Interested advertisers can contact team leader Rick Bruner at rick@centralcontrol.com for information about participation.
The cost of participation will vary depending on the specifics of a marketer’s campaign scope and on the number and type of marketing elements being tested in a randomized control test. For interested marketers, the research team will review planned campaigns and key priorities for testing and develop a relevant plan. In most cases, test costs will range from $75K to $125K depending on these variables. This will be the cost just for the RCT, not for the media buy or for any special data overlays required.
What is the benefit to marketers in participating?
Marketers will learn what worked and what didn’t work in the campaign under study. They also will learn the details of how RCT can more easily be applied to their other cross-platform campaigns. Depending on the details of their test campaign and research spec, they may also be able to use the results to establish a few empirically sound pillars that can make their MMM and MTA models more accurate.
What kinds of campaigns are best suited for RCT21? What kinds of campaigns wouldn’t be a good fit? For what elements of a marketing campaign do you expect to build experiments in RCT21?
Campaigns where sales can be measured at the individual household level are the most ideal for RCT21, although store visits, website visits, brand surveys can also be used. Long-running campaigns running across radio, outdoor, print moreso than TV and digital would be least appropriate for RCT21. The primary elements for these experiments are combinations of TV and digital media types, to better understand the incremental sales contribution of each type and each combination of types.
What type of commitment would a marketer need to make?
To pay the research and matching costs of the design they desire, and use at least one million dollars of the US media budget within RCT21, and supply IDs of target prospects/customers in quantity of at least one million IDs.
When are the results of the pilot test expected to be completed?
The study, to commence in Q3 of 2020, is projected to be completed by Q1 2021.
