Marketecture Podcast: Episode 123 Summary
Title: Olivia Kory from Haus tells us why YouTube is underrated. Plus saying goodbye to Xandr Invest
Host: Marketecture Media, Inc. (Ari Paparo and Eric Franchi)
Release Date: May 16, 2025
Guest Interview: Olivia Corey, Chief Strategy Officer at House
Introduction to House
Olivia Corey, the Chief Strategy Officer at House, a research company focused on measuring ad effectiveness, joins Ari Paparo and Eric Franchi to discuss her work and insights into the advertising landscape.
Background and Mission of House
Olivia shares the origins of House, highlighting their roots from a former Netflix team dedicated to rigorous experimentation in marketing effectiveness.
Olivia Corey [04:22]: "We were able to bring our CAC down significantly in the time I was there through kind of continuous experimentation."
House's mission is to democratize high-quality experimentation in marketing by offering it as a SaaS platform, thereby reducing costs and enabling broader access to incrementality testing.
GeoLift: House’s Experimentation Platform
Olivia introduces House’s flagship product, GeoLift, an experimentation incrementality platform designed to allow brands to configure experiments across various channels using geographic segmentation.
Olivia Corey [06:15]: "We're going to build on Geo. This will persist through the future."
GeoLift facilitates the setup of geo-based experiments, measuring sales lift through backend data integration, thereby providing a credit-score-like assessment of media quality.
The Underrated Power of YouTube in Advertising
A significant portion of the discussion centers on the underutilization of YouTube as an advertising platform. Olivia presents data indicating that YouTube's effectiveness is vastly underreported when measured through traditional click-based metrics.
Olivia Corey [07:21]: "It's way underrated. In Google Ads, the effect in your CPA is underreported by 3.4x."
She explains that YouTube excels in driving brand awareness and sales lift, benefits that are not captured when solely focusing on click-based attribution models. The platform's true value lies in non-click engagements, akin to traditional TV advertising.
Transition from Video Action to Demand Gen on YouTube
The conversation delves into YouTube’s strategic shift from Video Action campaigns to Demand Gen, a broader approach similar to Google's Performance Max (PMAX).
Olivia Corey [10:12]: "Demand gen is analogous to PMAX... We're delivering across all of Google's O and O, but primarily video."
Despite initial skepticism from marketers who felt a loss of control, Olivia notes that Demand Gen campaigns have been outperforming traditional Video Action campaigns, indicating improved outcomes and efficiency.
Applovin Studies and Market Reactions
Olivia discusses House's research on Applovin, a prominent player in digital advertising. Their studies revealed significant performance gains at lower spend levels, though scalability remains a concern.
Olivia Corey [13:22]: "Brands are seeing very strong lift at lower spend levels, but efficiency falls as spend increases."
Her findings led to backlash from some segments who misconstrued the research, accusing her of biased reporting despite her clear stance on the study's objective nature.
Future of Measurement: Experiments vs. Media Mix Modeling (MMM)
Olivia emphasizes the superiority of experimental methods in establishing causal relationships over traditional Media Mix Modeling.
Olivia Corey [16:20]: "Experiments are the best way to establish ground truth... MMM makes it easier to do things like budget planning."
House plans to integrate MMM with their experimental data to enhance scenario planning and budget allocation, maintaining experiments as the foundational approach.
Challenges with Ghost Bidding and Conversion Lift Studies
The discussion touches upon the decline of ghost bidding and conversion lift studies post-iOS 14 due to privacy changes, which have complicated user-level attribution and measurement.
Olivia Corey [18:24]: "Privacy changes have really challenged user-level lift studies... it's tough."
Despite these challenges, House continues to advocate for geo-based testing as a reliable alternative.
Industry News and Insights
Microsoft's DSP Xandr Invest Sunset
Ari and Eric discuss the recent announcement of Microsoft shutting down their DSP, Xandr Invest, effective February 2026. They analyze the likely reasons behind this decision, including business efficiency and a strategic pivot towards a closed-loop advertising system leveraging proprietary data.
Ari Paparo [21:29]: "Microsoft’s advertising business is killing it… shutting down Xandr DSP probably has very little effect on their core metrics."
The shutdown reflects a broader industry trend of major players focusing on their proprietary ad platforms to enhance margins and streamline operations.
Luma DMS Conference Highlights
The hosts recap key takeaways from the Luma DMS conference, particularly the rise of AI-first startups and the booming creator economy.
-
AI-First Startups: Emphasis on startups launching with AI at their core, achieving rapid traction with lower capital investments.
Eric Franchi [28:49]: "AI-first startups are launching today with less capital, getting traction faster…"
-
Creator Economy Growth: Unilever's commitment to 20x their spend on creators signals a significant shift towards influencer-driven marketing strategies.
Eric Franchi [28:49]: "Unilever said that they are going to 20x their spend on creators this year."
Trade Desk’s Open Sincera Initiative
Trade Desk announced the open release of Sincera's data set, aiming to standardize advertising quality metrics across the ecosystem.
Eric Franchi [33:15]: "Trade Desk's Open Sincera data set lets competitors participate in the same data…"
This move is seen as a strategy to undercut third-party verification companies by providing free, standardized data for quality assessment.
Perion’s Acquisition of GreenBiz
Perion expanded its capabilities by acquiring GreenBiz, a custom algorithm company from France, for approximately $65 million. This acquisition aligns with Perion’s strategy to build a robust real-time bidding (RTB) stack.
Eric Franchi [36:29]: "Perion has acquired GreenBiz, which broadens its vision to include custom algorithms…"
Mountain’s IPO and Market Position
Mountain seeks a $1.24 billion valuation in its IPO, positioning itself within the competitive Connected TV (CTV) and online video advertising space.
Eric Franchi [37:33]: "Mountain is seeking a $1.24 billion valuation on the IPO…"
House anticipates sharing more data on platforms like Mountain, highlighting their performance in upcoming episodes.
Netflix’s Ad-Supported Tier Growth
Netflix continues to expand its ad-supported subscription tier, reaching 170 million people through 94 million subscribers. This growth positions Netflix as a major player, surpassing traditional broadcast and cable networks in reaching the 18 to 34 demographic.
Ari Paparo [42:10]: "Netflix added 94 million, up to 170 million people. They reach more 18 to 34 year olds than any broadcast or cable network."
Agency Layoffs and AI’s Impact
The episode concludes with a discussion on significant layoffs within major agencies, such as Group M, where up to 45% of US staff may be impacted. The hosts reflect on AI’s role in increasing efficiency and the ongoing challenges organizations face in integrating AI effectively.
Olivia Corey [43:56]: "AI is going to make folks more efficient and more productive, but it hasn't happened yet."
Conclusion
Episode 123 of the Marketecture Podcast offers a comprehensive exploration of the evolving landscape of digital advertising measurement, the undervalued potential of YouTube, and the transformative impact of AI on the industry. Olivia Corey’s insights provide a nuanced understanding of current challenges and future directions, while the hosts’ analysis of recent industry developments underscores the dynamic nature of ad tech.
For detailed discussions and further insights, listeners are encouraged to subscribe to the Marketecture Podcast available on Marketecture TV.
