Google marketing researchers challenge decades of frequency assumptions, arguing first advertising impact delivers greatest value and S-curves reflect measurement limitations.
Pablo Perez and Jesús Martín Calvo from Google's Marketing Research & Insights team today published an extensive analysis challenging conventional wisdom about advertising frequency. The publication dissects contradictions in how marketers discuss optimal exposure levels and argues that fundamental consumer behavior supports maximizing reach over repeated contacts.
The 20th edition of Quantified Nation, published on quantifiednation.com, represents what the authors describe as their "most honest opinion" on a subject generating apparently contradicting insights across the marketing measurement community. Perez currently serves as Google Marketing Insights for EMEA, based in Madrid, while Martín Calvo works alongside him in marketing measurement research.
"The debate on 'optimal frequency' has been constant in media circles over the decades," the researchers wrote. "Everyone has some kind of opinion, some companies have well-established traditions and rules of thumb, often with obscure or untraceable origins."
The analysis identifies imprecise terminology as a primary source of confusion. Reach and frequency calculations depend on explicit conventions regarding time dimensions and channel definitions. Weekly calculations produce different balance points than monthly or campaign-level measurements. Cross-channel measurement remains difficult, leading to isolated or aggregated reporting that obscures true deduplicated reach.
"Making those explicit and clear is our first recommendation," the researchers stated.
The term "optimal" requires equal precision. Marketers use the concept to mean maximizing return on investment for given spend, maximizing topline revenue through spend changes, achieving measurable campaign impact, or channel-specific goals. Business conversations rarely specify which optimization objective applies.
"A good analyst is the one that helps the business ask the right questions," the publication notes.
Reports advocating higher frequencies often adopt effectiveness stances rather than efficiency perspectives. When advertising channels can absorb significant investment at marginal returns exceeding average cross-channel returns, higher frequency becomes a consequence of proper investment levels rather than an optimization target.
The researchers distinguish between macro observations from testing or marketing mix modeling and individual consumer behavior. Any macro-level insight inconsistent with consumer behavior should face scrutiny unless network effects or emergent behaviors provide explanations.
The publication employs a thought experiment assuming homogeneous consumers with equal brand predisposition and media consumption patterns. After delivering a first advertising contact, the fundamental question becomes whether the second contact should reach the same person or a new individual.
"From the data we have seen in our careers, data collected doing proper brand lift meta-analysis while controlling all the variables, the optimal is to deliver the message to a new person," the researchers wrote.
The analysis challenges readers to consider extreme scenarios. Universal agreement exists that 100 contacts to one person produces worse outcomes than one contact to 100 people. The grey area involves comparing distributions like 20 contacts to five people, five contacts to 20 people, or one contact to 100 people.
Any answer favoring concentrated exposure implies tipping points in consumer decisions. The researchers reject this model. "What every advertising impact does is to increase the probabilities of a purchase. There's no tipping point, but rather an accumulation of higher probability as we increment the frequency."
This leads to a fundamental conclusion: the first impact delivers the best value, and ideal scenarios maximize first-time impacts over any alternative.
Three phenomena could invalidate this model. Consumers possess different brand predispositions, making some people more valuable to reach than others through segmentation-affinity effects and heterogeneous product distribution. Media consumption patterns introduce cost efficiency considerations. Advertising delivery involves frictions including simultaneous impacts, imperfect information about screen-side individuals, and gaps between planning and execution.
"All these elements are true and that's what justifies the need for sophistication in planning," the researchers acknowledged. "But they can't be above the fundamental truth that the 1st impact is the best."
Google's YouTube Brand Lift database provides what the researchers describe as the world's best evidence for frequency questions. The database contains massive samples per experiment, proper test-control designs, consistent key performance indicators across thousands of cases, and metadata on execution levels.
Google published data from this database in 2017, which the researchers state remains valid. The analysis shows indexes of 1.5 to 1.8 for second impacts, consistently below the 2x threshold that would indicate equal value to first impacts.
"The first impact is the best, the second delivers always less. Same goes with the subsequent ones," the publication states.
This high-quality data proves the consumer behavior model described earlier. Frequency of three produces bigger results than frequency of one, but optimization requires comparing alternatives rather than observing absolute performance.
The apparent paradox resolves through media planning realities. Building reach and frequency occurs simultaneously rather than as discrete choices. Planning capabilities including frequency capping, smart television insertion scheduling, and platform combinations determine the balance achieved.
Businesses perform better with frequency of three in underinvested channels with strong return on investment compared to placing funds in channels with lower returns. This reflects proper channel investment levels rather than frequency optimization. "Do not let the technical limitations of aggregated models (such as MMM) be inconsistent with real consumer behaviour," the researchers cautioned.
The analysis argues that S-shaped response curves contradict both consumer behavior models and Brand Lift data. Simple explanations often prove correct. Aggregated models like marketing mix modeling struggle to measure low execution levels properly rather than requiring complex consumer behavior assumptions.
"Most often the S-curve is explained as 'first you don't see a lot of impact because there's not enough consumers impacted by the investment, then you see it ramp up, eventually there's diminishing returns,'" the publication states. "Easy to understand but not true to reality: the earlier stages of the media investment do deliver value too - actually it's likely that they deliver the most value. It's just that we can't properly measure because of the low signal-to-noise ratio."
Marketing mix models typically incorporate Hill functions on total impacts or impact subsets. The Meridian framework from Google uses linear assumptions on reach with Hill functions for frequency. Hill functions contain two parameters, with S parameters above one indicating S-curve behavior and values below one producing diminishing returns curves.
"It is our view that values <1 are the only ones that are consistent with true human behaviour," the researchers stated.
Modern measurement combines data sources and methods to produce optimal answers. Calibration of marketing mix models using A/B test data represents established practice.
The publication proposes incorporating Brand Lift results of frequency by level as priors to Meridian model parameters. "Seems like a good way forward," the researchers noted.
This approach would align model outputs with real-world consumer response patterns documented in controlled experiments. The Google Meridian framework received significant updates in September 2025, including channel-level contribution priors allowing users to guide models with business knowledge.
The researchers reference several measurement developments. The IPA Effectiveness Awards 2026 opened for entries with early registration closing January 9. Charlie Ebdy from Omnicom Media Group UK noted that "the advertising industry's shift toward digital media has outpaced its ability to prove effectiveness."
Cross-media measurement initiatives continue expanding. Google Ads introduced Cross-Media Reach measurement in June 2024, providing deduplicated reach and frequency insights across YouTube and traditional television. Amazon launched frequency capping features with cross-campaign reach measurement on December 23, 2024.
Industry analyst forecasts highlight incrementality and attention as metrics to watch. The shift away from impression-based measurement toward proving business impact resonates with the frequency discussion. "As the industry matures, expect more scrutiny on whether those extra impressions actually drive incremental value," Ad Age reported.
The publication notes increasing Google Trends interest in Panettone, the Italian sweet bread export appearing increasingly on shelves. "Have we reached 'peak Panettone' yet? Apparently not: yet another year with record interest in this product."
The researchers discuss Erwin Ephron, whose 1995 Recency Planning concept fundamentally challenged frequency threshold obsessions before his 2014 death. Conventional wisdom at the time, rooted in Herbert Krugman's three-hit theory from the 1970s, held that consumers needed multiple exposures before advertising could work.
Ephron argued that advertising works by reminding people already in the market. Since knowing when consumers will be ready to buy proves impossible, the best strategy maintains continuous presence maximizing the probability of reaching someone at the right moment. Reach, not frequency, should be the priority.
"One exposure is not a waste. It's all you need to remind a person who is ready to hear," Ephron stated in his 1997 paper.
The analysis notes the irony that three decades later, with vastly more data and computing power, the industry continues fighting the same battle Ephron won on paper.
The publication includes a footnote explaining their use of the term "tradition" to describe ways of doing embedded in organizations, accepted without challenge, and hard to track in origin. "They are a social consensus and a way of operating."
A second footnote acknowledges one potential emergent behavior: multiplier benefits from seeing the same campaign with iconic assets and consistent ideas across different media. However, the researchers note this effect lacks proper individual-level measurement beyond typical marketing mix model statistics on synergies. These inherently represent cross-channel conversations, while frequency already presents sufficient complexity in single-channel contexts.
The publication appears amid significant measurement infrastructure evolution. AudienceProject appointed Bruno Furnari as Chief Product & Technology Officer on October 27, 2025, as the company advances independent cross-media measurement capabilities.
Media fragmentation demands new budget allocation strategies as traditional channel loyalty diminishes. Effective brand building requires utilizing multiple channels including linear television, connected television, online video, social media, and open web platforms.
Comscore expanded its Cross-Platform Campaign Results to include streaming audio measurement and enhanced Meta reporting on December 15, 2025. The platform aims to provide advertisers with comprehensive campaign measurement across platforms representing modern consumer media engagement.
The advertising measurement ecosystem continues consolidating through partnerships addressing fragmentation. Médiamétrie partnered with AudienceProject in September 2025 to develop a Cross-Media Video advertising measurement solution targeting first quarter 2026 launch.
Technical specifications for the solution include accurate assessment of deduplicated campaign reach and frequency across multiple channels, encompassing major platforms, social media, connected television, and online video.
Who: Pablo Perez, Google Marketing Insights for EMEA, and Jesús Martín Calvo, both from Google's Marketing Research & Insights team based in Madrid.
What: Published comprehensive analysis of advertising frequency in Quantified Nation edition 20, challenging industry assumptions about optimal exposure levels and arguing that first advertising impacts deliver greatest value. The researchers propose calibrating marketing mix models with YouTube Brand Lift data as priors to align with actual consumer behavior.
When: Published on February 3, 2026, representing three decades after Erwin Ephron's Recency Planning concept and nine years after Google published YouTube Brand Lift database findings.
Where: Published on quantifiednation.com and promoted through LinkedIn by Pablo Perez, reaching marketing measurement professionals globally. The analysis draws on Google's YouTube Brand Lift database containing thousands of experiments with proper test-control designs.
Why: Addresses persistent confusion in marketing circles about optimal frequency, where apparently contradicting claims and unclear terminology create difficulty seeing a clear picture. The researchers argue that imprecise language, mixing of concepts, and technical limitations of measurement tools obscure fundamental truths about consumer behavior. The publication matters because many aggregated models produce S-curves inconsistent with real consumer response patterns, potentially leading to suboptimal media investment decisions. The analysis provides measurement practitioners with a framework for thinking about frequency that aligns macro-level optimization with individual consumer behavior, proposing practical calibration approaches for modern marketing mix modeling tools.
